In our final CES 2026 Q&A session for Tom's Hardware Premium, we had the opportunity to sit down with AMD in Las Vegas, Nevada, to discuss the evolving PC hardware landscape and how products like the Ryzen AI 400 might make a splash over the coming year.

If you've not checked out our previous rounds of transcripts, be sure to catch up with our FSR Redstone roundtable transcript, AMD ROCm transcript, Jensen Huang Q&A, and Intel Panther Lake Q&A.

As ever, elements of this transcript have been lightly edited for flow and clarity. In the weeks following this roundtable, we have also reviewed AMD's Ryzen 7 9850X3D CPU, so if you're wondering about how AMD's claims stack up to real-world tests, be sure to check it out.

AMD Representative: ... a little bit about the Ryzen AI 400 box, but that's not my primary area of expertise. I think from a component, CPU, a graphics standpoint, yes there is the 9850X3D that was part of the announcement (although not included in the keynote yesterday) that will be a new product that comes into the AM5 portfolio, that really will complement the existing 9800X3D and market. I think as we look at that product and its performance, there are certain games and workloads that really respond to the higher frequency, and so it was one of those things where, just kind of as we've been gaining more experience with our X3D products, and with the 9000 Series X3D products in particular, we took the opportunity to bump the clocks on that a little bit, and deliver a part that I think will be another choice for gamers who play more of those frequency-sensitive eSports titles and things like that, in addition to the games that our X3D products really excel at already.

You know, outside of that, on the graphics side, while we didn't announce anything new at CES, most of what we did on the graphics side really was in the run-up to CES, last year, with the introduction of FSR Redstone, and a host of new technologies that augment our FSR portfolio, bringing in, you know radiance caching, frame generation, and a handful of things like that to continue the progress on our FSR journey. I think we see that as a key part of the portfolio going forward; you'll continue to see us investing in and growing the capabilities, and part of that as well is utilizing things like GPUOpen to proliferate that throughout the gaming ecosystem and developers who may not have been on that first wave of games that those technologies get integrated into.

And then I think maybe the last topic to talk about a little bit before we just open it up for a little bit more free-form discussion is, you know, one of the products that we didn't make a lot of noise about last year that's starting to get a little bit more traction is the Radeon AI Pro R9700, which is a 32-gigabyte version of our Navi 4 products. It really comes in a couple of different flavors: one that is sort of a workstation form factor, dual-slot full-length card, actively cooled, and then late last year, we also introduced two versions of that card that are passively cooled cards designed to go into a server form factor platform, either as a dual-slot full-length card or as a single-slot card. And I think that's kind of complemented with the release of ROCm 7.2, which will come in January, supporting both the Ryzen AI and AI Pro 400 products, the Ryzen AI Max products, and our Radeon and Radeon Pro products out of the box. That includes both the standard 9000 series as well as the 9070 cards, both in the Windows ecosystem as well as the Linux ecosystem.

Maybe sort of last-but-not-least, just kind of recapping some of the things that we've been doing on the AI front, just to simplify that experience is, in our Adrenaline software stack we're now [including] the AI software bundle, just trying to make it easier for people to access the tools, to experiment with and use AI on either their notebooks with integrated graphics, or Strix Halo type products, or discrete graphics systems. That's really intended to be kind of an 'easy button' that pulls all of the tools and frameworks that are needed to just, out of the box, run AI workloads on our GPUs, and do it in a way where it lowers the bar for entry for more people to have that experience on their system.

To talk about the keynote a little bit, I think that one of the things that in more the client and gaming section of the keynote, one of the things that I think is really important and really interesting is this migration of models from the cloud down to local devices. I think that whether you're a creator or whether you're trying to run agents locally on private, proprietary data, we see an enormous amount of interest, of traction in that right now, we think that that is one of the trends that is really going to accelerate in 2026. And it spans from being able to do those capabilities on a notebook, like some of the things you saw last night, to dev platforms like the Halo box, all the way up to single or multiple R9700 parts in a system to tackle a much bigger workload. I think that whether it's individuals trying to do that, or developers, or even sort of departmental deployment of that as a workgroup AI solution within a business, we think that's going to be one of the significant trends that really shapes demand for those products in 2026.

So, with that, maybe I'll pause for a minute and see if you guys have any questions or things that are on your mind that you want to talk about.

Journalist 1: I have a couple of related questions about the impact of AI on [unintelligible] and the recent concern that RAM prices have gone way up, and my question is, what can you do about it? On the gaming side, AI demand is driving it up; do you have any insight into this problem? I guess the companion rumor is that GPU prices are going up, and that's more speculative, but [unintelligible]

AMD Representative: I think that the press in general has done a pretty thorough job of covering some of the dynamics around memory, supply, and demand, and I think that it is definitely true that some of these large deals to secure the RAM supply have affected the overall supply chain for memory. I think that when you look at it at a macro level right now, there is a finite supply of capacity at DRAM manufacturers that shared, whether you're talking about HBM memory, DDR memory, GDDR memory, it all comes from the same handful of companies around the planet, and I don't think there's any segment that's immune to the sort of unprecedented forces that are shaping demand in that part of the market right now.

I think on the CPU side, you know one of the phenomena that we have seen more recently that is really interesting is, as we talked to a number of our retail partners and E-tail partners, they are seeing quite a few more purchases of a CPU without a motherboard in the in the shopping basket, which indicates to us that the in-socket upgrade market and the longevity of our AM4 and AM5 platforms is actually turning out to be a really big benefit for gamers or creators who want to get more out of the platform and components that they already own without having to make that investment in a new RAM kit and a new motherboard and everything else. So I think that will be potentially a trend we see in 2026 with more component upgrades as opposed to full-system swap-outs and altogether rebuilds. It's interesting; as we look at our own telemetry data on users who are part of our Adrenalin software package that have opted-in to share some of their information with us, some of the most popular CPUs that are still running in gamers' platforms are like the [Ryzen 5] 2600, back to the Pinnacle Ridge days, or 3000 series Ryzen products. Stepping even from there into a little bit more modern 5000 series processors in an AM4 socketed motherboard, there's a pretty big boost there.

Journalist 2: I wanted to follow up on that right quick, because —

Journalist 1: Wait, just, the other half of the question —

AMD Representative: I'll get to GPUs in a second; I'm not going to forget. I had a bet, honestly, going into this, that this would be the first question, so thank you for that.

Journalist 1: So you're saying that Zen+ and Zen 2 parts are popular, from your own data; the upgrade path to Zen 3 — a lot of that's been blocked; the 5800X3D isn't available, the 5700X3D isn't available, and there's no option to scale up to AM5. I know you've done a couple of the Zen 3 refreshes, for the XT models, was it? In 2024? Are there any plans to refresh any of those, given how much demand there is for them now?

AMD Representative: I don't — so with those XT models, for instance, I think one of the things that we found is a small speed bump in a product doesn't materially affect the experience that a customer gets out of those CPUs, right? And I don't know that it's tremendously valuable to a user to do like a little tweak to an existing product definition. That being said, if you look globally at the sales and the demand that we see for our products, our AM4 processors and products still represent a very healthy, like, I'm going to say, between 30-40% of our total sales revolve around that AM4 platform. Some of that's new build, some of that's upgrade-in-socket, but it's still a platform which has very, very healthy demand that goes along with it. We're certainly looking at everything we can do to bring more supply and kind of reintroduce products back into the AM5 ecosystem to satisfy the demands of gamers that maybe want that significant upgrade in their AM4 platform without having to rebuild their entire system, so that's definitely something we're very actively working on right now.

I think on the GPU question, there is no product that has memory in it that's immune to some of these forces around DRAM pricing and what it's doing to the market, and I think the truth is the volatility that we've seen over the past two months or so has really been unprecedented. I think that there's a lot of behavior that we saw in the market with — um, once those deals were announced, significant acquisition of DRAM products to make sure that some of the larger customers have supply for those products, and I think that's what created a large volatility. What I expect to happen is that some of that volatility is going to settle out over the first half of this year, I think some of that volatility is going to settle out in the market over the next 3-6 months, and I think we're going to see pricing stabilize.

That being said, I think it is a given that the cost of memory is going to go up in every platform. That's going to affect prices in desktop builds, graphics cards, you name it, and we're doing everything we can to mitigate those cost increases, but our commitment is, just as we launched the product, to make sure that we're, from a price-performance and competitive positioning standpoint, maintaining the value that the Radeon 9000 Series products brought into the market when we launched them, and we're going to do everything we can to maintain that.

Journalist 3: You want to talk about X3D, and memory, and how it [unintelligible]?

AMD Representative: Okay, that's, that's a good point, and what is interesting —I think that there's maybe two aspects to the topic that Matt was bringing up. One is, for our X3D processors, often, kind of the rule of thumb with Ryzen products from the very beginning was you need the fastest memory kit so you can overclock the fabric on the processor to get the most gaming performance out of it, and that's, I think, a true statement for all of our products except for the X3D products. The X3D products, because of that really large cache next to the processor, are much, much less sensitive to memory latency, to fabric overclocking, and so I think one of the things that we've seen as we've peeled the onion and tried to dissect 'how do you build the best value for somebody that's trying to build a new system', I think one of the things the X3D products do really well is they mitigate that need to have the most expensive, lowest latency, fastest kit that you can buy. You can get almost all of that performance with a much less premium kit that's part of that. Now, I'm not saying that — of course, there's a premium across all memory kits right now — but I think it gives you an opportunity to open that search range and still get great performance without having to buy the best of the best memory kit to really push the envelope.

And I think the other thing that we've seen is, again, I'll say, in some parts of the world, not every market, every part of the world, there's a lot of our partners and customers that are looking at, 'okay, what does single memory channel versus dual memory channel performance look like?' And I think the thing that we've seen in some of the data is, of course it's title dependent, but a lot of games that are that are super popular, that people play all around the world, have relatively little sensitivity to dual channel versus single channel memory. So I'm not necessarily advocating that single memory channel systems are the right choice for everybody, but I think that one of the things that's going to be a theme throughout 2026 is really scrutinizing every component that's a part of the build to make sure that you're getting the best value for what you're putting together. And I think the second part of it is this phenomenon of a more significant trend toward component upgrades versus full system upgrades. I think it's something we're going to see, like upgrading the CPU, or the GPU, or the motherboard to get a little bit more out of that system without having to refresh everything. I think is going to be something that's really a trend throughout the year.

Journalist 4: The other big story [...] What will PCs, what will gaming desktops look like, between the Steam machines, between collaboration with Microsoft and whatever they're doing, it seems like the gaming desktop could look very different this year, and I'm just wondering how you guys are thinking about that as you're doing [...] because not everyone needs a big tower.

AMD Representative: That's true. So, I'll use the Halo box down there as a great example, right? I think that the Strix Halo product was a product where, when we first were going through the process of designing and building the part, there was a lot of conversation about what really is the value proposition of of this really unique architecture and its capabilities, and I think certainly we saw an opportunity as a single chip that could replace a mid range to lower end GPU really, really interesting, really compelling. As a mobile workstation product, super interesting as well, or even a small form factor like that, as a desktop workstation is a really interesting use case as well. And then I think, over the course of the past 18 months, the role of local AI has really changed so much, the memory architecture of that type of product has kind of given it a whole other dimension of spaces and use cases and workloads that it can serve very, very effectively with [a] kind of unprecedented level of capabilities. So I think that when we look at 'what does a desktop look like over the next couple of years,' I think that that's a great example of a unique system architecture that solves the problem in a different way, potentially leads to something that is more "console like," in terms of its size, capabilities, and dimensions. I think that you're already seeing the transformation that's happening, like with these FSR capabilities that we're introducing, and how much ML or AI-enhanced techniques for rendering are augmenting more traditional rasterization techniques to drive the gaming experience. I think there will always be a space for those mid-tower desktop systems, for people who want the absolute best, and the highest resolutions, the fastest refresh rates, the most detail in their games that they play. But I think there's also a pretty significant class of gamers who want something that is just easy, small, quiet, and fits about anywhere, and I think that you will see this — I guess 'resurgence' may be the wrong word, but I think the demand for the small form factor gaming desktop will be something that really picks up steam in 2026 and 2027, and I think it's just going to be something that continues forward as you look at how ML capabilities affect the way that GPUs render different games, and experimenting with different architectures to just get the most out of limited power, and maybe slightly different silicon capabilities and construction to go solve that gaming problem.

Journalist 3: Is graphics memory protected in some way from the surges in memory pricing, or are they in the same boat as main memory?

AMD Representative: I think that if you look at who builds graphics memory, it's the same DRAM manufacturers that build DDR memory, HBM memory; again, there's a relatively small set of companies on the planet that build those chips. There is a lot of overlap in the technology, and again, that's why I say I don't think that any memory type is immune to the effects of how AI is shaping the overall supply and demand picture for memory this coming year.

Journalist 3: It seems like the industry [could have] planned a little better, because you need a lot of AI PCs out there to meet that AI demand...

AMD Representative: I can't comment on planning cycles of the DRAM manufacturers, but maybe, just going back to it, the DRAM industry, if you look back in history, has been characterized by these mega cycles, right? And I think you're seeing another mega cycle right now that maybe is exacerbated a little bit by this very sudden demand shift to AI GPUs, which makes it probably a little bit more painful because of the speed at which it went from 'no problem' to, you know, 'DEFCON 1'. Yeah, I think that's very true, that just, the speed at which the change happened in the market is pretty unprecedented, and I think that everybody is reacting to the sudden demand shift that you're seeing, but it takes time, right? It takes time to bring new capacity online, to increase production, and all of those things are problems that could still have in past memory cycles, but I think we're definitely seeing this one hit very, very suddenly.

Journalist 4: Do you think we're working on a longer timeline in that cycle? We've heard anywhere from like —

AMD Representative: I am not the expert there to be totally — what I couldn't tell you — I don't know who you speak with, from some of the DRAM manufacturers or things like that, but I think those are good questions for some of the companies that make DRAM, to understand what they're doing, what their timelines are. If history is any indication, it follows the same pattern that we've seen in the past, right? Demand outstrips supply, more fab capacity gets put in place, and eventually you get to a point where you have oversupply compared to the demand that's in the market, and the cycle repeats itself. So, I don't know the answers to those questions. I think you'd have to talk to those guys to really get the get a get a better understanding of what they're planning on doing and how they're planning on reacting to the current situation.

Journalist 3: I think I read too that Samsung came out and said, 'expect your cell phone to go up,' right?

AMD Representative: Yeah. I think I saw Samsung said nothing's immune, TVs, dishwashers, and then everything else.

Journalist 4: I do have a question you can answer about the 9850X3D. So, during the pre-brief, we got a taste of — I think 7% was the number that was quoted, we got performance after the fact, and I think it was averaging 3% or somewhere around that?

AMD Representative: I remember hearing 7% in the pre-brief, but I think the 7% was mis-statement. There are some games — it depends on what your suite is, right? You look at e-Sports games, and it's going to be in that 5 to 7% range; you look at AAA games and it may be parity on an AAA game where the X3D effect dominates the influence on the game engine and how it plays. I know that from our internal data, and correct me if I'm wrong, but I think, like over a suite of 25 to 35 games, blended, it's in that 2-3% range overall, with the outliers on one end and the 'no change' on the other end.

Journalist 4: Well, my question wasn't so much the discrepancy, my question was — the point is to have these two live side-by-side, we don't know pricing yet; I just want to confirm once again that that is the plan, given the small gap in both performance and frequency? Given that you can just turn on PBO with the 9800X3D and easily get another 200 MHz or so out of it. The plan is to still have both of them?

AMD Representative: Absolutely, absolutely, we will definitely have both of them. I think that as our Ryzen business has gotten more diverse, I think what you'll see is — look, a lot of system builders like stability. They'll want to keep that 9800X3D in a lot of the prebuilt systems that they're manufacturing. I think even as an individual processor in a box sale, you'll see both available in the market. The 9800X3D has been a stellar product for us, tried and true, such a great reputation. These transitions take time, and we're certainly not going to do anything to artificially force this to be a faster transition that it needs to be. Both are really valuable products, and have have their own space where people are going to gravitate to one or the other.

Journalist 3: I also think it's going to be a fun value proposition; how you guys characterize it, right? Because there are some people that absolutely want the best of the best. Obviously, there'll be a price gap there, but they will go for the top-end one. The other thing that's going to be interesting is eSports professionals. There are people that, 'five frames per second? Here's my money.' So it will be the top, they will exist together, and I'm really excited to see how you guys — because I'm pretty sure when you do the math, the best dollar per frame is going to be the 9800, right? Because of that gap. But what is it worth to you? You know? And to David's point of in socket upgrades, do I want to go for that? So to have that choice is really cool. For me, I'll turn on PBO. Like, that's no problem for me. There are people that don't want to — have never entered the bios, right? So you can tell them — and PBO is 200 megahertz? You're not going to get 600, or 400, excuse me. So this came up; I don't know if it was on you guys' call or another one, but somebody asked about the 400 [MHz], and our overclocking team was like, "if you squint at like, you have to dial that in, to get anywhere close to that." You can! But a number of things have to fall in line to get there, including silicon lottery.

Journalist 4: I'm just curious to see how much it costs.

Journalist 5: I have a slightly different question. Since the market is shifting in terms of pricing, do you expect people to lean a bit more toward pre-builts as opposed to going out and building their own PCs?

AMD Representative: It's a good question. I think that the next couple of months are going to be hard to predict, how all these forces shape what people buy and how they buy it. I was saying a little bit earlier; I think one of the things that we've certainly seen through the holiday period is talking to many of our retail and e-tail partners; they see a significant difference today in the shopping cart that users check out with, the number of processors to motherboards. In other words, more people now are upgrading in-socket than ever before. I think that's an option for somebody who already has an existing system, but those pre-built customers, a lot of those kind of fall on two ends of the spectrum: you have people who want the best of the best — and look, if your system costs a little bit more, it's not going to affect their decision to go build that system.

I think on the other hand the customers that are more budget-conscious may be making different choices than they would have six months ago. In our own tracking data, we see a little bit of an uptick in the percentage of AM4 versus AM5 platforms, DDR4 memory has been a little bit more stable in terms of pricing, especially recently, and I think that's driving a little bit of that. So it may force some different behaviors than what we would have expected. My statement around DDR4 and AM4 is also very regional. In the US, or in North America in general, AM5 is the heavy, heavy majority — I'm talking 80-90% plus of North America has moved to AM5 — but if you look at it globally, it's a lot more balanced.

However, I think even in North America we are seeing a little bit of an uptick in AM4 that started to show up in December as things got so volatile and difficult to predict with memory pricing overall. So I think that's something that would be an interesting sort of trend to watch as we go into 2025 [sic]; as we're here at CES, that's the conversation that many of our partners are having with us, motherboard partners as well as system builders, just to make sure that they've got the right mix between AM4 platforms and AM5 platforms so that, whatever the price point is that they need to address, or customer that they're going after, that they've got a solution that fits that space.

Journalist 6: [mostly unintelligible question about AMD's Radeon lineup]

AMD Representative: I think that 2025 was a pretty special year for Radeon, and the response from gamers and the community has really been pretty awesome. I think that we, as we look at 2026, and beyond, our goal with Radeon is to make sure that we're following that playbook that was so successful for us with our Ryzen CPUs, and I think over multiple generations, what Ryzen really built was something where, if you were building a Ryzen system, you knew you were getting a great value, you were getting more for your money, and you're getting great features in a really highly competitive product. As we plan out the future of Radeon, it is the same thing. I think we're on that same curve; we're a few clicks earlier in the journey, but we want to do exactly the same thing, and as time goes on, continue to expand the coverage within the graphics market to make sure that we're addressing as broad of a range of gamers as we possibly can. I think that we've seen it this year; bringing choice into the gaming graphics ecosystem is a good thing for everybody. Having two strong competitors in that space is good for gamers, it's good for PC builders, and I think it's strategically very, very important to AMD. Whether you're talking about notebooks and integrated GPUs or discrete graphics, what we do with visualization and gaming is really core to AMD's DNA, and it's a part of how we build our product portfolio and our roadmaps.

It's kind of ironic; a lot of people look at gaming as a workload that really isn't challenging, like AI or computational fluid dynamics or something else, yet visualization and gaming workloads are really some of the most complex workloads there are. As we move from traditional rasterized gaming to ray tracing and path tracing — like, very, very complicated, both in how that works and how it works efficiently, and, drives high frame rates and low frame times and all of those things, and solving those problems is actually really important, not just for our gaming graphics business, but really kind of across the board. So it's an area that we are absolutely focused on, and absolutely focused on growing that portfolio. I think we've been very open; you've certainly seen and heard and read the conversations that Jack [Huynh] has had with with various members of the press about how our strategy, very deliberately, was to go after the spaces where gamers live and go after the, really big parts of the market, and I think we're going to start there and just continue to expand our coverage as time goes on.

Journalist 4: You said you can't speak on Ryzen AI 400 very much?

AMD Representative: You can ask me questions; I'll try my best. [chuckles]

Journalist 4: I just want to get a reaction to — you know, we have to get a refresh. AMD was the performance leader in that segment of low-power SoCs in this previous generation. We have this refresh now that looks like the same architecture, similar specs, with some higher memory speeds and frequencies. I want to get your reaction to Intel's announcement yesterday, with a very large GPU inside of a mobile chip, and where AMD sits now, given that they're just kind of going after gamers with this huge push for this entry-level gaming platform that might be competitive with AMD's current offerings.

AMD Representative: The announcements from Intel yesterday, betting on a bigger GPU, that has been a part of our portfolio strategy for quite some time; products like Strix Halo fit into that overall scheme. There are going to be places in the portfolio where betting on a really big GPU is important. There's going to be places in the portfolio where having the balance between CPU cores, GPU cores, NPU, and total cost is more of a driving factor in the construction of a particular chip. I think that what Intel has done, stepping up in their graphics configuration in such a substantial way, is very aligned with the approach that we've been taking for years across the portfolio, which is: big graphics in chips that go into notebooks matter. It is important; it enables a class of gaming systems, or mobile workstations, creator systems, whatever, that doesn't require a dGPU and offers really exceptional power, and performance, and gaming characteristics. So I think that what they've done is really quite well aligned with what we've been executing in our strategy over the past couple of years.

Journalist 3: Kind of a follow-up question to that is, we've seen NPU tops — short of Qualcomm, who is coming out and saying 80 TOPS on their parts — we've seen NPU TOPS level off around 45-50, and more focus is going on the GPU, from Panther Lake in particular. I'm wondering: do we see that kind of inference capabilities living at this level? Or is there going to continue to be that push over the next few years for faster and faster NPUs?

AMD Representative: I think that the NPU is something that fills a very specific need on the part, and you have to look at the characteristics of the NPU in multiple dimensions. It's not just TOPS, it's TOPS per watt, it's bandwidth, it's all of those things, right? Because, really, at the end of the day, the NPU was intended to be that engine that is handling high-duty-cycle always-on AI workloads in the background, and doing so at low power without materially affecting the battery life in your system. That's kind of the objective of having that accelerator in there. That being said, it's not the sort of 'one tool that solves all problems,' because I think if you look at CPU versus GPU versus NPU, an NPU is incredibly targeted at being very efficient, at running persistent AI workloads. However, its programmability is more challenging than a GPU, for instance.

A GPU is more flexible in terms of its programming model, but its power characteristics are not such that you would want that to run high-duty-cycle AI workloads all the time, right? And so, to me, it's kind of 'right tool for the right job', and so this I think this 'plateauing' of NPU tops that you've seen over the past couple of generations from us, from Intel, really is indicative of that there's a set of workloads that are pretty well-defined and characterized as those high duty cycle AI workloads, and that kind of fits the characteristics of those NPUs. As those workloads evolve, I think the NPU will evolve as well, and I think that those two things go hand in hand with each other to really get the value out of the NPU. I think a more powerful NPU without an evolution of those high-duty-cycle AI workloads that are running on it is probably less valuable than — in the specs itself, having higher specs on an NPU doesn't necessarily translate to value for the end-user.

Journalist 1: [unintelligible question about Microsoft Copilot]

AMD Representative: I'm not super close to the Microsoft features, but just as an end user, I do see those slowly evolving and getting more and more integrated into the way that Microsoft is trying to deliver them. So that, too, is probably its own evolution. I think that the early steps with Copilot created an expectation that made that a little bit more challenging, and now I think it's actually starting to come into its own a little bit more with some of the things that it's able to do.

[Session ends]

That wraps up our Tom's Hardware Premium CES coverage for the year. We hope you enjoyed this insider look at our conversations with the biggest companies around.

After AMD's keynote at CES 2026, we were invited to a roundtable discussion at the show with AMD in Las Vegas, Nevada, covering their latest graphics technologies, like FSR Redstone and more. Some highlights include insight into why AMD chose not to put RDNA 4 graphics in its AI400 mobile chips, the open sourcing of FSR, and much more.

Before getting stuck into the press Q&A, be sure to check out our previous coverage on AMD's announcements at the show, including Redstone on RDNA 3, hinting at open sourcing FSR 4, and the company's overall announcements they made at their keynote last week.

With all of that out of the way, please note that some elements of the transcript have been lightly edited for flow and clarity, and we have also identified certain speakers when they were made known. While recording in a noisy, ambient environment, some things can get lost in the audio mix we've denoted as such in the copy.

Josh Hort (AMD): So here, we're at the FSR Redstone Roundtable; my name is Josh Hort. I'm Senior Director here at AMD in the Computing and Graphics group, so I work underneath Jack Huynh. You've probably heard that name before. I lead ISV enablement. So what does that mean? It means everything from AI PC software enabling, to building that ecosystem up, to all the game engineering and integration of FSR technologies, to benchmark optimizations and benchmark engineering.

So my team is responsible for working with the ULs and Primate Labs of the world to make sure we get the scores that we do on the products that we deliver, and we're also responsible for developer programs for GPUOpen; well, that's the website that my team manages, as well as hardware seeding to developers and other key parties and partners. So that's me in a nutshell. I've been at AMD about seven and a half years; you can find this on LinkedIn. Before that, I worked at Intel for 17.

I don't know how much you guys are aware of Redstone. I can go really fast through this, or I can go really slow. If all you guys know all this stuff, we can just blow through it, because it's really just a recap of what happened on December 10.

Journalist 1: Yeah, I know it. I think we all do.

Other Journalists: [chorus of acknowledgment]

Josh Hort (AMD): If you know what was released; we have over 200 titles that we delivered in 2025, which is — I do have to pause though, because the amount of progress that we made in 2025 is phenomenal. I mean, from the outside looking in, I hope you guys see it the same way, because when we launched at the end of February, we had, what, thirty? Thirty-two titles, 33 or something. I'll tell you internally, I told my team: by Computex, let's have a stretch goal of 100 titles, and they blew that number way out of the park. So it's just fantastic, the amount of reception we've had?

And I think — combined with putting it on GPUOpen, which has really also increased the uptake. Because we can't be everything to everyone. [...] We can't cover the whole sphere of developers that are out there, right? So putting the code out on GPUOpen has also leveraged some uptake that we weren't even anticipating. So it's working, I guess, long story short, right? If you look at FSR 1 / 2 / 3 support, it's 500-plus [games]; Redstone, it's 200-plus. And you guys are aware that Redstone is for the 9000 series cards. Okay, so if there aren't any questions, I can just stop [...], and you guys can ask away. We don't need any slides. Sure.

Journalist 2: Multi-frame generation.

Josh Hort (AMD): Multi-frame? I can't comment on future plans. I can say absolutely we're looking at it. I think in general, we need to get the right things right first, and multi-frame gen, of course, introduces latency. And so how do you combat that? We have technologies like Anti-Lag, but we have to marry those two technologies together so that you can improve the latency the best as possible.

Journalist 3: So I think a more rounded question is, do you do you see a demand for a multi frame gen? I know when I go to play a game, I rarely find myself going above 2x frame gen.

Josh Hort (AMD): It's always in the eye of the beholder, right?

Journalist 3: Sure.

Josh Hort (AMD): Some people are more tolerant to latency. Some games like, Twitch games, FPS games; it's not really appropriate, but if you're playing something casual, usually on the casual side, it also has less demands on the GPU, so you don't need it; your frame rate is already high enough, right? And if you're a casual gamer, you're not trying to get to 240 FPS. So is there a place for it? [noise of uncertainty] Like I said, we're looking at it, but we don't have any product announcements to make at this time.

Josh Hort (AMD): How are you? I'm Josh. Dean? Pleasure.

"Dean": Will I get better at 240 FPS?

Josh Hort (AMD): Will you get better at 240? I dunno. If you've got the golden eye, maybe. eSports games, sure, right?

AMD Representative: For that, you've just gotta get really close to the monitor, and turn it to 240.

Josh Hort (AMD): I know eSports guys that can see to the pixel at 240 Hz. Some of these people, even if they see one pixel change — and they can see that. Okay. I dunno about you guys, but my vision's getting a lot worse. [chuckles]

Journalist 4: So what about support for RDNA 3.5? Because there was an announcement yesterday, the new mobile Ryzen AI 400 series [...] a new stack, but they don't get new [...] performance.

Josh Hort (AMD): I feel like the question's been answered already, so I don't have much — anything more to add. I'd say that we're always evaluating the roadmap, and we have to make the right priority call. [...] I will say in general that we have a lot of products out there, and it's a lot of support. It's not just hey, "I took some leaked source code, and I put it on the internet, and it works." That's not how you make a product.

Journalist 3: I don't want to monopolize this conversation, so if anyone else—

Josh Hort (AMD): No. I mean, you guys, feel free to riff off each other, right? Let's make it a natural conversation.

Journalist 3: I had a question on handhelds. Right now, AMD's hardware is dominant in the handheld space, but there's a—

Josh Hort (AMD): Another player that's out there that's pretty small but substantial.

Journalist 3: Yes.

Josh Hort (AMD): And I love 'em by the way.

Journalist 3: Right now, the feature support is there [...] on the Windows handhelds, but there's kind of this disparate integration; there's features like RSR and FMF to be brought to the forefront to that kind of handheld experience. I'm wondering how you're engaged with OEMs to ensure that those features are brought to the forefront so that's easy for players.

Josh Hort (AMD): It's a different team from mine that works with the OEMs, so I can't really speak from any first-hand knowledge, but nothing would prevent them from bringing the Adrenalin driver to the device, and then getting the analytical-based FMF or upscaling as wel.

Journalist 3: So would you be engaged with Valve on the SteamOS side of things, or Lenovo?

Josh Hort (AMD): Yeah, I can't comment on the specifics, right? But absolutely; our partnership with Valve is deeply important to us, and those kinds of features are important to them as well. But I don't have any announcements, [or] feature sets to talk about.

Journalist 3: Okay. So, without going into announcements, is the interest there from — SteamOS is becoming an increasingly popular destination for handhelds; is there interest in those types of features? Is it something that you're discussing?

Josh Hort (AMD): We're absolutely investigating it with them, yes. But again, I can't speak for Valve, or what their intentions are.

Journalist 3: And this is, I guess, a broader Linux question, because the Adrenalin support in Linux is not really there.

Josh Hort (AMD): Well, RADV has really taken off, and a lot of it is _because_ of Valve. Have you seen the amount of contributions they've made to the driver? It's phenomenal. And that's not to say that the [AMDGPU-PRO] driver has gone away, but there's just fewer releases, and they're very targeted releases. But RADV is meant to be the main open-source driver going forwards.

Journalist 3: I'm aware that OEMs can tap in to any feature within Adrenalin; I'm moreso curious if there's any engagement from AMD on that side, because — of course, I understand that, but when I'm going to pick up a handheld, and there's the Lenovo one, there's the MSI one, and there's the ASUS one, and I want the Ryzen Z2 Extreme, and it doesn't have the features that I'm expecting that are easily accessible, it doesn't create the best user experience.

Josh Hort (AMD): Well, FSR and frame gen are on the ROG Ally X, right? And the ROG Ally.

Journalist 3: ASUS has done a good job with it. Moreso than Lenovo and MSI.

Josh Hort (AMD): I see; so that's more what your question is targeted at. I mean, I can't speak on behalf of them, and what they chose for their products is their decision, but the technology exists and it works, right? The Xbox devices do support the analytical frame-gen and upscaling. And actually you can get really good framerates on some games that are very challenging.

Journalist 5: I hacked Lossless Scaling onto my Steam Deck for that exact reason. Hey man, if I had AFMF on my Steam Deck and I didn't need to do that, that would be great.

[some discussion of Samsung's new 1,040 Hz Odyssey 3D 6K monitor]

Josh Hort (AMD): Well, you gotta remember that means you have to make 1,040 frames per second on the GPU side. Which is going to eat into your performance. Your frametime is sub-millisecond. You're rendering not a lot of geometry at that point.

AMD Representative: If you had the choice, then, would you rather be at 240 or that brand-new 1040 Hz monitor?

Josh Hort (AMD): Personally — this is just my personal — this is not an AMD-based opinion — [...] Everybody's trying to push technology to the limits. I'd much rather get 4x the frametime to get 4x the amount of pixels and quality pixels into it, than running it at some ridiculous framerate. And the ML techniques can improve that, right, like MFG, multi-frame gen? Sure, absolutely. But if you're doing 6x, 8x, 32x frame gen? The latency falls apart. Your mind is very quickly going to realize that—

Journalist 3: Even 4x is pretty rough.

Josh Hort (AMD): Yes.

Journalist 3: So Redstone was kinda this big moment for, for the most part, feature parity with — short of MFG, which I think is a question of value to a lot of people — this level of feature parity and pushing toward what NVIDIA would call Neural Rendering. You're not announcing any products, not doing anything like that, but I'm wondering: are there more applications of AI and machine learning in rendering? And where do you see it having more applications outside of upscaling and frame generation?

Josh Hort (AMD): The easiest place — because when we talk about internally, some people just synonymize it with gaming, and I have to stop them and say "no, time out." In workstation, there's a lot of CAD/CAM applications, that whole segment, from Blender, to Autodesk applications, etc etc; you guys know them all. [Those applications] lend themselves very well to not only upscaling and frame gen, but also things like ray-trace denoising, like ray generation, and even neural radiance caching; the more advanced things we've been bringing towards the ray-tracing [and] path-tracing acceleration. So yeah, the first step is — and actually, we are working with ISVs in bringing FSR to their applications. NVIDIA already does this; the competition already does this, so it's just a natural progression. Did that address your question, or was there another part of it?

Journalist 3: I guess what I'm saying is, for me, when I hear about — obviously, frame generation, and upscaling, those make sense, but then I hear about ray regeneration, about neural radiance cache, and as someone consumes this content, and doesn't make it, these are applications of machine learning that I hadn't realized, and they can make a really significant difference. And so I'm saying, from your perspective, what are those other applications for machine learning in the rendering pipeline?

Josh Hort (AMD): I see; so like "what other things could we do?"

Journalist 3: Yeah.

Josh Hort (AMD): Well, there's certainly ReSTIR, which is reservoir sampling; I'm trying to think of what else there is. There's neural intersection functions, what we call NIF, with importance; basically, when you're doing ray tracing, you're looking for those intersections where the ray intersects with a certain volume, and if you can create a neural method to figure out which rays are important and where they're going to intersect, you can accelerate not having to go through the BVH tree traversal, which is a very expensive thing. So the less BVH traversals you have to do, the faster your ray-tracing and path-tracing is going to be. And really, your path tracing is what, just multi-bounce ray-tracing, right?

So we can actually get to the path-tracing era, because there really aren't hardly any titles out there that support path tracing, because you need a very expensive graphics card in order to run it. So that's what ML is going to bring is really democratize, bring it more to the masses, right? I think Lisa said it best, yesterday — AI everywhere, for everyone, right? We really believe the same thing, and with FSR Redstone, and what's coming next, and so on and so forth is — yeah, we want to bring this to the masses. And it's really the functions and the features that we think are going to be useful, versus just some candy or powerpoint where it doesn't have ISV uptake, developer uptake.

Journalist 5: Will there ever be support for Redstone on RDNA 3?

Josh Hort (AMD): [group laughing] Haha, he asked the same question. [Journalist 4 above]

Journalist 6: I got to play today with Panther Lake with multi-frame generation. It was pretty good for a very sleek and mobile device. Ryzen AI 400 is coming not only to the notebook and [unintelligible] and also for the desktop part, so there will be small desktop PC with Ryzen AI 400, and you get a pretty good experience on a big screen; you have a small tiny box over there, with good frame generation, so, it is ... the way to go, I guess.

AMD Representative: That's kind of a hardware question, right? It's kind of hard for Josh to answer.

Journalist 6: Yes, sure, but —

AMD Representative: Your real question is "why isn't it RDNA 4 in Ryzen AI 400," right?

Journalist 6: No, that's [laughing] no, that's really another question — that's a BIG question.

Josh Hort (AMD): That one was a product decision, right? And I'm not the product decision-maker.

Journalist 6: Ryzen AI 400 is also a refresh of the 300 series, I get that. Maybe in the next version we will see RDNA 4 or something like that. [...] I think the market is big, especially for the notebooks, and the tiny desktops.

Josh Hort (AMD): The demand is out there, is what you're saying.

Journalist 6: Yes.

Journalist 3: I'll tell you what I would like to see. [Journalist 5] mentioned Lossless Scaling ealier; when AFMF first came out, I was expecting there to be an approach by AMD to approach the developer of Lossless Scaling to bring that into the driver, and speaking of multi-frame gen, I think one of the applications for it that would make a lot of sense to me is in the driver, through AFMF, to scale up higher, which is obviously something Lossless Scaling can do to varying degrees of success. That is definitely something that I think could turn [points to small PCs] that, that, to — not a gaming machine necessarily, but —

Josh Hort (AMD): Capable of running something, yeah.

Journalist 3: And obviously you guys have made a lot of strides with ISV engagement, we mentioned FSR3 earlier, but you can't engage everyone, it's not possible, and that driver-level solution for me has been really great.

Josh Hort (AMD): It's definitely something we're investigating closely, how we can bring more of the technology to the driver so it doesn't have to be what I call "enlightened," which means like, in-engine, in-game, versus "unenlightened," where the driver takes care of it. And yes, we are definitely investigating — again, no product plans, no announcements today — but absolutely we're looking at how we can bring some more ML tech to the driver level. It makes it easier to distribute, it gives you backward compatibility with games that will never get updated ever again. The publisher might not even exist anymore.

Journalist 7: [Andrej Zdravkovic, AMD SVP of GPU Technologies] mentioned that the plan was to still open-source FSR4, and I wanted to confirm that with you after, obviously, the source code came out in August I think it was? That's still the plan, to open-source—

Josh Hort (AMD): Each of the technologies will be on their own trajectory for open-sourcing. I don't make the decisions or do the open-sourcing; my team executes it, like puts it on the website. But it's a strategic discussion on what gets open sourced and what doesn't. If Andrej told you we're gonna open source it, then — well —

Journalist 7: [laughing] You won't dispute that claim?

Josh Hort (AMD): Now, [as for] when? I can't say when. Obviously, we try to be as open as we possibly can without giving away the farm. Because we want proliferation of the technology, right? There's partners who come out of the woodwork that could be competitors of ours, who also want to be partners on things like this because it's just good for the ecosystem. FSR was picked up by Microsoft for their AutoSR implementation, for instance, and the main reason why they picked it is because it was open. Open source is good for everybody. [Note: this is incorrect as far as we know; AutoSR is ML-based and uses a custom CNN.]

Journalist 8: You made a comment that we've made a lot of progress with ISVs [...] if your team was in charge of this, was there like a change in strategy, or more of a focus? How did you get from point A to point B, where you got like 200 games in a couple months?

Josh Hort (AMD): I think I kind of hinted at it, right? I think putting it out on GPUOpen, we hit critical mass. I think it was a flywheel where the game publishers saw "OK, this tech actually works pretty well," and the Digital Foundry review came out, they did the pixel peeping and were like, "Wow! This is pretty fantastic." It was a glowing review, which was great for us, right? But it puts us on the map where like, OK, the technology's at a point where it's mature, it's ready to go, and my team put in a lot of effort in the ISV enabling portion, but also like I said, putting it on GPUOpen so that anybody can pick it up I think also really helped propel us way faster than we could do with just the people power that we have.

Journalist 8: Following up on the change in strategies, when you're approaching ISVs for integration, particularly in games, outside of budget and time constraints, do you hear anything else from developers concerning their hesitation for adding features, especially the latest and greatest features?

Josh Hort (AMD): It depends on what piece of technology is getting inserted into the pipeline and where. If you look at things like super resolution, and frame generation, those are both post-processing steps at the end of the pipeline, so it's very easy to integrate them. Now, denoising, the ray regeneration portion, that can be more tricky, because some game engines have a fused denoiser where they want the denoise and the super-res step to happen at the same time; others want them completely separated because they're at different parts of the pipeline.

So when you have a fused denoiser, it provides performance, but it can't be used by all game engines, just because of the way they are written. On something like the neural radiance caching, that one is even more complex, because it needs a lot of different input information into the model to get the output, and the information might not be readily available in the format that's required by the model in order for it operate. It's also kind of intrusive into the pipeline. Long story short, it's harder to integrate.

Journalist 8: So, especially for those features, you're looking to engage developers before release as much as you can?

Josh Hort (AMD): Yeah, so like we did a lot of work with Fatshark, because they can move really fast, and they were embracing the technology; not necessarily "building the plane as we're flying it," but super cutting-edge, rapid iteration, super close, deep technical partnership. The amount of work we did in a small amount of time is phenomenal. But we have more work to do, and that's going to be a focus in 2026, is getting that Neural Radiance Caching feature — not only getting it into more titles, but also improving on the integration, the API and whatnot, so as we get more feedback from our ISV partners, we'll be improving it across the way in 2026 and beyond.

Journalist 1: How's Redstone been with VR? Any issues with ghosting, or?

Josh Hort (AMD): With VR? I haven't tried it personally; I don't have a VR headset — I used to. But in my copious spare time, it was collecting dust, unfortunately. I haven't heard it first-hand of any issues. I would say frame gen in general might be a problem? Because I know from past experience that latency is so critical to avoiding nausea. If your brain can perceive one frame that's behind it immediately gets sick, you're done. Some people don't notice it, but a lot of people do, and that's why they get sick. And if you think about it, frame gen today is, you're taking a past frame and the current frame and inserting another one, so you have to hold a frame. You're naturally introducing that latency?

By default, you've already just voided the benefit there, because you're going to get sick. Now super res, sure! It's a fantastic technology. It's almost a requirement, because you want to get to 240 Hz. Like, he was talking about 1040 Hz, in a VR headset 1040 Hz actually might be pretty cool. High-speed, super realistic — but again, you're cutting your frame time down, your rende time, to one millisecond or something, depending on what your resolution is and whatnot, so you're not going to be able to do a lot of high-quality features.

[Session ends with off-topic VR game discussion.]

This year, CES was all about AI-powered gadgets and home assistant robots, which meant 3D printing and laser technologies fit right in. Creality, Snapmaker, AtomForm, and Protopasta each showed off the latest trends in the world of consumer 3D printing.

Creality

Creality emerged as the show’s clear winner, earning Tom’s Hardware’s Best 3D Printer of CES 2026 award for its new Creality SPARKX i7. The SPARKX i7 is an easy-to-use multicolor printer with AI-assisted features aimed at reducing friction for new users. While reviewing the SPARKX i7, we were impressed with its super-simple CFS Lite and quick-swap hotend. The programmable RGB lights along the bottom aren’t just for bling; they’re also a cheerful status indicator. When paired with the Creality Cloud app, users can pre-select files from the Creality library and send them directly to the printer, making the SPARKX truly plug and play.

Creality’s booth also showcased an updated Ender-3 V4, the Falcon T1 Laser Engraver with five swappable laser modules, and the Sermoon 3D scanner, all reinforcing the company’s push toward a consumer home fabrication ecosystem.

AtomForm

Newcomer AtomForm debuted the Palette 300 3D printer at parent company MOVA Group’s booth. MOVA is known for making robot vacuum cleaners, and they themselves are a sub-brand of Dreame, maker of premium household appliances, which is part of the Xiaomi Ecosystem, who are known for their smartphones and consumer electronics. TL’DR is AtomForm is not just a couple of dudes in a garage making cool stuff; they have some serious backers in the world of Chinese consumer electronics.

The Palette 300 is a fairly large, fully enclosed Core XY 3D printer with 12 nozzles stored on a wheel that it can swap like a mini tool changer. It can handle up to 36 colors and has a 300x300x300mm build volume. We’ve been hearing about this machine since last August, but this is the first time we’ve been in the same room with one. Unfortunately, this was a static display, so we can only assume they’re using a nozzle swapping method similar to Bambu Lab’s H2C or Bondtech’s INDX that will soon be featured on the Prusa CORE One.

The OmniElement automatic nozzle-swapping system is intended to dramatically reduce purge waste while maintaining print speed and quality through the use of four internal cameras and 50 sensors. The company claims the printer can hit speeds of 800mm/s with a 25,000 mm/s² acceleration speed. We’ll report back once we get a test unit if the AtomForm lives up to the hype.

AtomForm plans to launch the printer through Kickstarter later this year. Prices will start at $999 for a standalone printer and $1299 for an RFD-6 Combo, which will have a six-spool filament management system. A 12-spool system (with 2 RFD-6 units) will have an early bird price of $1808, nearly half off the full retail price.

xTool

xTool had a large display area to show off a wide variety of lasers, plus new technology like this prototype UV printer coming in Q2. This machine will be similar to the EufyMake E1 and will print a 3D texture onto objects using UV ink that it can cure as it prints, building up layers of texture. xTool promises a true prosumer-level machine that is more than a hobby printer, but not as expensive as an industrial unit. This is more of a lateral move for xTool, as they already market to small business users and have a deep catalog of blank merchandise you can buy to engrave and now UV print.

The unnamed UV printer does not yet have a price. xTool said the prototype brought to CES is used for design validation, and that while the core features are developed, they are still refining hardware and software.

xTool also showed off their new flagship P3 CO2 laser (starting at $6499), which we reviewed in November, as well as the xTool F2 Ultra fiber laser, which we just finished reviewing. The F2 Ultra is a MOPA laser that can color engrave metal with interesting results, and is currently on sale starting at $4699.

Protopasta

Protopasta showcased a materials innovation with its Quantum Dot filament. The PLA-based filament is filled with tiny semiconductor crystals, just a few nanometers across, that hyper fluoresce when excited by black light. This produces material that has a significantly brighter and more vivid glow than anything currently on the market. The filament was developed in collaboration with artist Olga Alexapoulou of Quantum Light, who pioneered Quantum pigment.

Quantum Dot filament is part of Protopasta’s experimental line of limited edition filaments. If you want to secure a sample or a few spools, you can subscribe to Endless Exploration before Feb. 28, 2026.

Snapmaker

Snapmaker was invited to showcase its affordable, four-color tool changer at the Kickstarter booth. The U1 had a highly successful Kickstarter launch that raised $20 million. The tool changer is our current pick for best multicolor 3D printer, as it offers high-quality speed with practically no wasted filament. Rather than announcing new hardware, Snapmaker focused on community engagement, allowing CES attendees to see the machine in person and marking the project’s transition from crowdfunding success to production reality.

Ohsung System Co., Ltd., was at CES 2026 with its next-generation metal paste 3D printer dubbed the Gauss MT90. This device sets itself apart in the metal 3D printing niche with its compact desktop form, by working at room temperature, and by leveraging office‑safe paste-based (patent pending) metal extrusion (PME) tech, which eschews powders. Ohsung System Co. claims the Gauss MT90 is “a game-changer that is reshaping the landscape of the metal 3D printing industry.”

Metal 3D printing has until now been the reserve of companies that are set up to work with large, expensive, and potentially hazardous machines. They typically use high temperatures and metallic dust materials – which heighten fire and explosion risks.

With the Gauss MT90, using PME tech, the size and hazards involved in metal 3D printing are greatly reduced. As you can see in the images, MetalPrinting’s new 3D printer looks a lot like many other desktop 3D printing devices on the market, and it even has user-friendly ‘printer cartridges.’ Its operating noise, of at about 40dB, is also reasonable.

The site shows that MetalPrinting has prepared ‘GaussInk’ cartridges with stainless steel, copper, titanium, nickel, iron, and even tungsten materials, and says “custom and alloy ink materials are available upon request.” These cartridges contain the PME paste-like inks.

In brief, the Gauss MT90 extrudes its inks at room temperature. An integrated HEPA filter minimizes any potential emissions into the surrounding space during this process. You can watch through the window, control and monitor it via a 7-inch touch display, and an AI-supported camera is on guard to make sure everything proceeds as it should.

To be clear, the Gauss MT90 output doesn’t simply dry and harden and give you usable output straight off the bed. The 3D printed part has to also be heated in a sintering oven, to strongly fuse the metal paste particles, before it is ready. There isn’t much information available regarding the sintering oven process, apparatus, or specifications.

We have previously covered metal 3D printing in space. That Airbus system, deployed in the ISS, was a relatively hazardous device to use in such a constrained environment, with its stainless steel filament melted at 2,192°F (1,200°C) to additively manufacture parts. At least it didn’t introduce powdered metal into the ISS’s cabin atmosphere…

The Gauss MT90 3D printer won a CES 2026 Innovation Award. For availability and pricing details, you are encouraged to contact the company directly.

Edit 2/5/2026 2:50am PT: Clarified company naming.

At CES 2026, AMD used an interview with PCWorld to restate its official position on FSR Redstone while softening the surrounding edges. During the interview, AMD Chief Software Officer and head of graphics software Andrej Zdravkovic said the full Redstone feature set remains tied to RDNA 4 hardware. At the same time, he stopped short of shutting down the idea of an experimental Redstone build for RDNA 3 users who are trying to make it work on their own.

Redstone bundles several ML-driven graphics features, including upscaling and frame generation, under a single umbrella. AMD has been clear that these features are designed around the performance characteristics of RDNA 4. Zdravkovic reiterated that position, saying the decision is not about artificial product segmentation, but whether the hardware can deliver a consistent experience across a wide range of games and systems. If enabling a feature degrades performance or image quality, AMD does not see value in shipping it.

Zdravkovic also explained that frame generation and similar ML workloads must complete within a single frame budget. If the GPU cannot do that fast enough, the result can be counterproductive. “If you don’t have enough time to do… the machine learning operations required, then you have to reduce the frame rate… to double it,” he said, describing a scenario where the feature undermines its own purpose.

Inevitably, the conversation switched to community experimentation. Asked about users who are already forcing parts of Redstone to run on RDNA 3 GPUs, Zdravkovic said, “all the power to them,” noting that such hacks may work on a specific machine or in a specific title. “I’m a geek myself, so I would do that for any technology,” he added.

Could AMD offer a beta or prototype Redstone build for RDNA 3? Not right now, according to Zdravkovic — such a release is “currently not in the plan,” but he did thank PC World “for the hint” and expressed interest in thinking through how such a prototype might work. AMD also emphasized that it continues to improve older architectures where it makes sense. Zdravkovic said the company is “definitely not going to withhold anything that really makes a difference,” but only when the net result is a clear improvement to gameplay quality or responsiveness.

So, there you have it. FSR Redstone is an exclusive RDNA 4 feature set for now, and RDNA 3 users will have to fall back to earlier FSR implementations. However, it’s encouraging to know that AMD is acknowledging the demand and recognizing the modding community’s efforts and not completely ruling out official support with a hard no.

Superheat was at CES 2026 to showcase what it describes as “a water heater that pays for itself.” The new Superheat H1 is a $2,000 water heater that warms up your H2O using waste-heat from its built-in Bitcoin mining ASIC hardware, rather than an immersed resistive heating element.

It is claimed that owners of a Superheat H1 can “offset up to 80% of electricity and water costs” with earnings from cryptomining. Superheat extrapolates these numbers to suggest that a 700-apartment community could raise “up to $980,000 yearly earnings.”

The infographic above shows that a Superheat H1 can earn $1,000 in passive income per year. Therefore, your initial hardware investment should be entirely paid-off in two years, while it reduces your water heater energy spend by 80%. The H1 has a predicted service life of 10 years, very similar to a regular domestic boiler.

Waste not, want not

Cooling data centers that run demanding workloads like cryptocurrency mining is a notoriously expensive business. It is estimated that cooling is the second most costly activity after actually powering the systems to chew through the tasks central to their existence.

But what if the ‘waste’ heat became a benefit? Indeed, some businesses, and most households, spend a lot of their energy budget heating water, for washing and heating systems, and so on, and get nothing else back except a hefty utility bill.

Enter the Superheat H1 with its “dual-value operation.” It requires roughly the same amount of energy as a regular electric water heater, according to the maker. Moreover, it is claimed to be scalable beyond homes, for apartment blocks, hotels, and so on.

“Heat is one of the world’s most overlooked resources,” said Andrew Geng, Co-Founder and CTO of Superheat. “The H1 proves that home appliances can create real economic and environmental value. As we expand into distributed AI and cloud computing, Superheat will redefine how buildings produce, reuse, and monetize heat.”

Bitcoin pricing is rather volatile. In Q3 last year, it nudged over $125,000, but 1BTC is currently valued at around $91,000. Superheat H1 owners should benefit from a higher Bitcoin valuation, but no one really knows which direction, nor how high or low BTC could go in 2026. CNBC recently measured investor and analyst predictions and estimated that Bitcoin valuations of between $75,000 and $225,000 would be seen this year.

Owning a Superheat H1 might be yet another reason for not being able to tune out from watching information feeds and stressing about trends, charts, and valuations. Can I heat water with a heatsink on my forehead?

CES may have started out as the Consumer Electronics Show, but it has never been solely about consumers. That was true before the AI boom, and it’s even more true today. Jaded tech journalists who’ve prowled the halls for a decade (and usually more) don’t want to hear the AI talk, but the topic is so pervasive at the event that even hopeful skeptics can turn to cynicism. It’s not that AI is a worthless advancement. Quite the opposite; AI is such a monumental shift in computing that it begs for bandwagon products aimed at easy investor dollars.

Nvidia and its monumental rise into the mainstream of speculative assets has led that charge, but AMD has been taking notes. And at CES 2026, AMD looked a lot like Nvidia. CEO Lisa Su took the stage for a two-hour keynote, joined by an ensemble of partners that don’t come anywhere near consumer technology, all driven by a simple thesis: “AI everywhere.” Although it won’t comfort PC enthusiasts, from a business perspective, it’s a fair focus for the opening CES keynote.

Still, it’s a stark contrast from the AMD of even 12 months ago, and an even starker contrast to Intel. Under the leadership of Pat Gelsinger, the public-facing Intel quickly jumped on the AI boom. Presentations became winding events focused on road maps and geopolitics, as Intel tried to play a game it was struggling to be a player in. The Intel at CES 2026 was far different. It presented a tight, 45-minute keynote that was laser-focused on its latest, most exciting consumer announcement: Panther Lake.

AMD briefly touched on its new Ryzen AI 400 ‘Gorgon Point’ chips during the keynote, but it felt like little more than a footnote in the two-hour presentation. And the Ryzen 7 9850X3D, which is probably the most important announcement for Tom’s Hardware readers, didn't even get a callout. That’s not to mention the fact that Gorgon Point is little more than a refresh of the Strix Point chips that are already available. I wonder, if AMD hadn’t changed its mobile branding to include “AI,” if Gorgon Point would’ve been cut from the keynote, as well.

AI is everywhere, and for AMD to ignore that fact would be detrimental to its business. But we should approach this topic realistically, lest it devolves into complaining that AMD didn’t announce or focus on the products I (or you) specifically want. The problem isn’t AI. The problem isn’t that AMD isn’t in an opportune time in its product cycle to announce new CPUs or GPUs. And the problem isn't that AMD has bold data center ambitions.

The problem is that AMD spoke little to consumers and gave them even less to actually care about. Intel did the opposite.

There’s still room for consumers in AI

We need some context before continuing. Last year at CES, AMD’s keynote included several consumer announcements. The company revealed the Ryzen 9 9950X3D and Ryzen 9 9900X3D, expanded Strix Point and lined up Kraken Point, teased RDNA 4 and FSR 4, and introduced Strix Halo. The year prior, AMD introduced the RX 7600 XT, refreshed select Zen 3 chips, introduced Hawk Point as a refresh of Phoenix, and introduced Zen 4 APUs under the Ryzen 8000G brand. Both of these keynotes focused heavily on AI, make no mistake. But they also spoke to consumers.

This year, for its CES launches, AMD announced refreshed Strix Point processors called Gorgon Point, introduced new Strix Halo SKUs, and revealed the Ryzen 7 9850X3D. Of those three consumer announcements, only one was mentioned in the keynote. Further, even with a section of the keynote dedicated to personal computing and a subsection within that talking specifically about Strix Halo, AMD didn’t even discuss its new Strix Halo SKUs. Instead, it announced a DGX Spark competitor called Ryzen AI Halo, undercutting the last 12 months of effort in engaging OEMs to build out their Strix Halo offerings. It also ignores the clear issue AMD has with accelerators compared to Nvidia, with AMD still using two entirely different architectures across the consumer and data center segments.

Perhaps the clearest example that AMD lost its consumer focus is the new AI Bundle coming to Radeon Software. This bundle includes some AI software like PyTorch that’s configured to run on Radeon GPUs. The idea, as AMD’s president of GPU Technologies and Chief Software Officer Andrej Zdravković later shared with me, is to remove the barrier to entry in installing and configuring all of this software. Its goal is to get consumers with consumer-grade AMD hardware playing with these tools to leverage the AI capabilities that the broader industry insists are important for consumers. It, too, was cut from the keynote.

Instead, much of the two-hour run time was filled by AMD’s partners. OpenAI’s Greg Brockman presented a story about ChatGPT users sharing their health information with AI and suggesting it was better than a real doctor. John Couluris of Blue Origin — a company perhaps best known to consumers for sending Katy Perry on a mission and creating a suggestively shaped rocket — talked about AMD hardware going into space. And Michael Kratsios, scientific advisor to the White House, ended the keynote talking about supercomputing and President Trump’s plans for AI. Regardless of your political views, stance on generative AI, or interest in space exploration, none of this is even remotely focused on consumers.

A two-hour keynote doesn’t need to pander to one specific audience. AMD is a large business with many customers across different fields, and it carries a certain responsibility in opening the largest tech trade show in the US to set the tone of the event across multiple categories. It served only to reinforce the idea that consumers don’t have a place in the AI future. Even with announcements focused on consumer AI, AMD chose to cut them.

Even with partners that create consumer products, AMD focused on the amount of compute it was able to ship to them. With 120 minutes to tell consumers about why they should care about “AI everywhere,” especially with RAM prices soaring and data centers putting unreasonable demands on electrical infrastructure, AMD mostly reinforced what many consumers currently feel, regardless of if it’s true of not — AI is a trend for Wall Street speculators that has infiltrated enthusiasts circles to rip apart pricing and supply chains, not dissimilar to what this group has experienced with cryptocurrency multiple times in the past.

The consumer market is still very important for AMD. In its most year earnings report, AMD's client and gaming business accounted for roughly half of its revenue — $4 billion compared to $4.3 billion in the data center — and saw nearly three times the rate of growth as its data center business year-over-year. Nvidia, which has largely abandoned any consumer focus during its keynotes, is seeing less than a tenth of its total revenue from its gaming business.

Intel went back to basics

Intel needs context, as well. Last year, Intel’s hour-long CES keynote included a victory lap with Lunar Lake, the announcement of the lower-end Arc B570 discrete GPU, Arrow Lake-H for mobile, additional Arrow Lake desktop SKUs, and a focus on Copilot+ and Intel vPro. The year prior, Intel carried the same thesis of “AI everywhere” as AMD shared this year, contextualizing its announcements of U-series Meteor Lake processors, 14th-gen HX processors for laptops, and additional 14th-gen desktop SKUs around their use in AI applications.

Again, there was a stark change this year. Intel presented a 45-minute keynote that felt like you’d blink and miss it. It was laser-focused on Panther Lake, and Intel still touched on the geopolitics of its position as the largest semiconductor manufacturer in the US, gaming, the applications of Panther Lake in AI, and even some retrospectives on its history with Arc graphics. It seemed like a self-aware Intel that we haven’t seen in at least a decade, probably longer. There’s no greater evidence of that than Dan Rogers, VP of PC products at Intel, stating that Intel’s graphics drivers have been “Ubiquitously deployed, with a massive install base, but previously, lacking in key features, driver support, and frankly, performance.”

Intel has broader news to talk about. It showed off a humanoid robot powered by Panther Lake, showcasing its embedded validation for the generation. Over the past six months, Intel has seen major investments from Nvidia and Softbank. And just last month, Intel announced major leadership changes, including a new executive leading Intel’s relationship with the government and politicians.

Intel certainly had fewer announcements than AMD. Although the consumer front is roughly identical, with both companies introducing new ranges of mobile chips, AMD also revealed the Helios Rack it's been teasing for half a year, expanded its Instinct roadmap, and brought forth its Ryzen AI Halo development box. Intel didn’t talk about new Xeon chips, rack-scale solutions, or engagements in robotics or AI. Intel hasn't clearly defined its AI strategy for the enterprise yet, and perhaps if Intel were in the position AMD is currently in, there would be a greater data center focus.

The greatest move Intel made at CES 2026 this year was to tell consumers why they should care about their latest product; in this case, Panther Lake. A clear example of that was Perplexity CEO Aravind Srinivas appearing during Intel's keynote. Srinivas is an AI executive, but he joined Intel to make an argument for on-device AI capabilities, touching on the security, economic, and speed benefits of having local AI compute. Indirectly, it spoke to the concerns of rising RAM prices and massive AI data center expansion. It made an argument to consumers about why they should care about AI without ignoring the broader implications about what AI expansion means for all of us.

There is no doubt that Intel would show off rack-scale solutions and hammer on a road map for AI accelerators if it had anything to share on that front; it doesn't. AMD does, so it makes sense that AMD touched on its future Instinct MI500X accelerators, Helios Rack, and Pensando NIC. It's the surrounding narrative around these product announcements where there's a significant divergence between Intel and AMD.

The table is set

Regardless of what you think about AMD or Intel, the dynamics between the two companies are shifting. A year ago, it wasn’t clear what the future of Intel looked like, and just six months ago, there were legitimate questions about whether the brand would even survive. From a market share perspective, Intel is still competitive with AMD. When looking at financials, Intel brings in more than double what AMD does in the consumer segment. The writing has already faded on the walls, though. AMD has seen massive momentum quarter-over-quarter, while Intel's financials look caught in stasis.

That dynamic showed up at CES 2026. AMD’s presence suggested its consumer announcements weren’t important, with releases like the Ryzen 7 9850X3D feeling like table scraps for a company going after the multi-trillion-dollar market cap of Nvidia. Intel, delivering its keynote to just a fourth of the audience of AMD's — likely even less, but I’m being generous — couldn’t afford to wax poetic about AI. It had to focus, and focus it did.

Where we go from here remains uncertain. Neither Intel nor AMD had major desktop announcements at CES, which isn’t surprising. Intel confirmed we’ll see an Arrow Lake Refresh on desktop at some point, while the heavily-rumored Ryzen 9 9950X3D2 is presumably in AMD’s back pocket. The real battle happens later this year. Zen 6 is on track for the back half of 2026, as is Nova Lake, where Intel is expected to finally introduce its bLLC packaging to combat AMD’s wildly successful 3D V-Cache.

The dynamics at CES this year provide some hints on how AMD and Intel will approach the next generation of desktop CPUs. Intel has clearly recognized its new position in the market, and it seems AMD — at least based on my conversations at CES and AMD’s focus on AI market expansion during its keynote — may struggle to reckon with the potential of stagnation as a market leader.

For Intel and AMD, focus isn’t a zero-sum game. An emphasis on the data center doesn’t have to come at the cost of another product category. At this point, however, it feels like the AI focus does come at a cost. AMD will have to balance a business that has grown even wider than it already was in the past 12 months, especially if Intel continues to double down on the consumer market, where it has lost so much ground.

Following Intel's keynote at CES 2026, where it formally launched consumer Panther Lake products, our resident CPU reporter Jake Roach attended a press Q&A at CES in Las Vegas, featuring some of the senior staff behind Intel's Panther Lake architecture. The following article is a transcript of that Q&A session. We've edited some elements for flow and clarity, and have made notes where we couldn't quite make out what was being said. As speakers did not introduce themselves at the beginning of our transcript, we have denoted them as Journalists, denoted in the sequence that they spoke in.

Some highlights include Intel being particularly proud of making the lineup significantly easier to understand, combining the benefits of Arrow Lake and Lunar Lake, while also making fewer SKU denotations, making the lineup easier to parse as a whole. Additionally, while Intel did not announce any EVO processors at CES 2026, the company is still keeping the brand alive, with 'stricter' qualifications in place for the future.

Ahead of reading this press Q&A session, be sure to familiarize yourself with all the elements of Intel's new Panther Lake architecture by checking out some of our coverage, or watch the keynote yourself below.

So, now that you've learnt a little bit more about what Panther Lake offers, here's our full, slightly edited press Q&A session. This session was around 25 minutes long, and while recording in a noisy, ambient environment, some things can get lost in the audio mix, and have been denoted as such in the copy.

Shalini Singh: Hi. My name is Shalini, Technical Marketing, Intel CCG (Client Computing Group, with me, Damien, who leads technical marketing as well. [laughs]

Damien Triloet: Yep, Technical Marketing, a lot of time spent on graphics and [inaudible]

Mitch Lum: I'm Mitch Lum, I'm the Product Manager for both Core Ultra Series 3 and Core Series 4.

Client Product Team Lead: I'm [inaudible], I lead the client product team.

Journalist 1: So, my first question is, you guys are using a lot of panthers in the presentation. It was nice. And we are used to [having], like, the codename. And when we have the product, the codename disappears. You guys don't use the code name anymore just presenting the unique architecture [...] But you guys are using a lot of panthers in design. [...] Are you guys using [Panthers] to represent this generational product for clients? The guys actually buying the product?

Client Product Team Lead: I was just gonna say, you did see the panther walk away. We will focus on [the naming convention of] Core Ultra Series 3.

Shalini Singh: Like previously, [you] will see many places where we'll continue to use the codename. Then there's just no going behind it. But the brand police will come to us and tell us "you always have to use the [inaudible] Core Ultra 3 [laughs]."

Mitch Lum: What's your advice?

Journalist 1: I know it's easier to understand names when it's just a numeral sequence, you know? But you can see how it's easy to identify this. Oh, this is a new thing. This is the Panther thing, and whatever the next thing is, the new codename [...] and you can [...] see the difference in generations this way as well. It's a good visual representation of the product.

Damien Triolet: It's a good visual representation of the product.

Journalist 1: Yeah, especially when [it's] a nice thing, a panther is a cool thing. So if you're using cool things as code names, maybe it works.

Shalini Singh: And also, for you guys, after you've seen so many panthers, it's difficult to unsee it.

Journalist 1: I mean, I mean, 300 series, 400 series. I can see where this is going, but it's a number. It's not as cool, I guess.

Client Product Team Lead: Oh yeah, we'll have it in the [press pack] that will look different, so that there'll be all the retail differentiators, and in any tech related areas, you'll see it's called call out as Panther Lake, but in end user facing [materials and assets] is going to be Core Ultra Series 3, Series 4. So, that's the least offensive way of describing the products, and I could keep going. But they'll have codenames.

Journalist 2: Do you guys see any sort of ripple effects from the ongoing price hikes, with RAM and storage as an example, like [...] Copilot+ PCs, I believe, have a minimum spec of 32 gigabytes [this is not correct, and was quickly clarified as a 16GB requirement]. So if that's true, does that affect the sales of AI PCs or Copilot PCs, going into the new year?

Client Product Team Lead: So in general, a lot of the OEMs, and you can talk to them in detail, they plan for nine months to 12 months in terms of allotment of their [products] so in the beginning through towards the end of the year, I think a lot of them are going to be reasonably okay in terms of memory. But, in general, we have to wait and see. It's like, if we can predict the memory prices, we would all also invest in the stock market, that kind of thing. So it is going to be a challenge, which we are carefully monitoring and trying to work as much as we can closely [presumably meaning with OEMs].

Journalist 3: Are we really going to be seeing developers using [inaudible, but the question pertains to AI TOPS and NPU performance]

Damien Triolet: One interesting detail about NPU 5 is that our first [inaudible, potentially Blackwell?] compatible, and PYP. So it will help developers move the roles they've done on NPU 4, and then keep investing into NPU so we see increasing traction.

Client Product Team Lead: A lot of the time, we want to make sure they're using the right engine for the right workload. So from an Intel perspective, it's not just about the NPU. AI means it's the GPU, NPU, and CPU. So as long as they're using the right engine, we're good. And from an NPU 5 perspective, our biggest thing was we shrunk it down in size so that we can fit a lot more inside Panther Lake. So from Lunar Lake to Panther Lake, the peak TOPS, it's fairly similar.

Mitch Lum: It's 40% more die area, if you think about Lunar Lake and the kinds of systems that it went into. Now we're going into many more kinds of laptops, and we wanted to make sure that we had that scalability.

Shalini Singh: But, it's probably the only x86 product to have fully featured XPU engines to run a performance GPU. You will not see that from any of our competitors.

Damien Triolet: And you will see that in the mlperf data that we have in the project, you should have a look later. When you look at the way our different engine, GPU and NPU, behaves in approach like LLM are very consistent. Both engines have, like, full access, memory bandwidth, a ton of compute for AI. That's not something you will see on the company's source product. They will either have the GPU enable something, or the NPU enable something. We fully enable both of them. And this gives a niche set [of tools] to developers [offering] the flexibility to target the right engine, for the right competitors.

Shalini Singh: In most cases, when they don't have either the fully labeled GPU or the NPU. They even have hybrid solutions, CPU plus NPU. You probably know that, right. But our GPU and NPU solutions in most cases [inaudible].

Journalist 4: I also expect this will require all four of you to answer. Why did you add X to the names?

Client Product Team Lead: In terms of X's, there was [...] strong retail feedback. There are 12 [Xe3 cores]. They needed an easy way for customers walking to be able to identify that. So those [products with X in the name] will have the 12 Xe [cores], the bigger graphics, Arc graphics. So that is an easy way to identify that.

Shalini Singh: It's also the highest-end config. So most of the data that you see in the press deck will be based on that config.

Tom's Hardware: I have two questions focused on the Ultra 5 range. First, we know the core split for the eight-core configurations, and the 16-core configurations. I'm just wondering what the core split is for the 12-core configurations. And then there is a core Ultra 5 that has 10 Xe3 cores in it, but it didn't get that Core Ultra X5 branding, and I'm just wondering why that's the case.

Mitch Lum: Yeah, so, on that one, the 10 xe3 [cores], it [has] our B370. There's B390, it's 12 Xe, 370 is 10 Xe. We wanted to make sure it was really clear with the X designation that we were really referring to our top configuration, which is the B390. So that's the first piece. And then I think your second question was that [Core] Ultra 5, we've got a 12-core configuration. That's 4P, 4E and 4LP-E.

Journalist 4: I had a question on the Xe cores and the GPU. Now, are you finding most of your improvements from just increasing the number of cores? Or, were the cores redesigned? And what sort of improvements are you seeing internally today?

Damien Triolet: So the bigger part of the improvement is coming from a larger configuration. But as you've seen, some of the general numbers that we share is over 70% so there's obviously a little bit more coming to that. There's a clock increase, that's part of it, and then the IP improvements, as we've detailed, at ITT if you were there.

Mitch Lum: And, also the bigger cache, so the GPU has a bigger cache.

Client Product Team Lead: 50% more cores, double the amount of cache.

Journalist 5: Yes, you guys told us that on stage, that Panther Lake [has] 15% better performance per watt compared to last year's chips. How is that translated in real use cases, like how many more hours [of battery life] people can have on their laptops?

Shalini Singh: We expect to deliver Lunar Lake level efficiency. The exact battery life translation will obviously depend on the overall design. It depends on the display, the battery capacity, [and] other components that they will have on the PC. But in most cases, at the most when we when we measure power at the SOC level, Panther Lake is more efficient than Lunar Lake.

So we are seeing power savings compared to Lunar Lake itself. If you have one, or if you tested it, you probably know it's really good on battery life. So Panther Lake is very similar, and so it's basically a combination; Lunar Lake efficiency and Arrow Lake performance, right?

So, you'll see about a 60% increase in multithreaded performance, battery life, and all the SOC, power savings are going to translate to a similar to better battery life [when] compared to what you had in Lunar Lake, because it also has a [higher] number of cores. So there's going to be a little bit of a balance there.

Mitch Lum: There was the 15% number from the very beginning of the keynote, when we were talking about process technology. [...] I think that was specifically comparing at a transistor level, 18A transistor performance per hour, versus [the previous generation].

Client Product Team Lead: The Lunar [to] Arrow numbers would be much higher. It's in the press deck [inaudible].

Shalini Singh: Yeah, you'll see that in the press deck. Basically, we're delivering similar performance [to] both Lunar and Arrow at 40% lower power. So [in] Arrow Lake, obviously, there is a big efficiency [gain] to be had, versus Arrow Lake, but even versus Lunar Lake, like you said, was quite good as it is. To be able to drive the efficiency further was amazing.

Journalist 5: So I was really impressed with the integrated graphics FPS [...] so I guess my question is; Do you see [integrated graphics] being kind of the future of gaming laptops, not needing that GPU anymore because you have [...] more efficiency, less heat... How do you guys see the industry moving?

Client Product Team Lead: At the highest level, I mean, this is more like the [entry-level], right? Similar to [the RTX 4050]. Clearly, there is a lot higher [power] discrete graphics, mobile display graphics available.

So, we don't see that going away anytime soon. That said, this is the first step in terms of getting to that discrete graphics level. And over time, there is going to be also creation where a lot of AI systolics in that integrated graphics [segment] is going to help. So we see this as a new breed of things that we want to build and see where AI is going to enable [Intel] to take it to another level... But in the near future? It's not going to go away. You will still see discrete graphics.

Shalini Singh: No, just to add, basically, you know, we do expect this level of graphics performance to be able to replace entry-level discrete cards, right? In some cases, even something a little bit more than that, like a 4050, like 4060, whatever. But yes, like he said, the high-end, we don't expect that to go anywhere. So those will continue to be there, but [Intel will] absolutely challenge the entry [level] domain.

Mitch Lum: And we will have Core Ultra Series 3 laptops that have big discrete graphics in them [inaudible].

Damien Triolet: As Shani mentioned, at the entry level for discrete, it is a good point for integration, for the simple reason that memory capacity for entry-level discrete is relatively limited, and more modern workloads require way more memory. So integration is a thing about how to get more accessible at that.

Client Product Team Lead: Speaking of the previous question on memory shortages... [laughs]

Journalist 6: So is Arc built into the entire [Core Ultra] Series 3 components?

Mitch Lum: There are some SKUs that are Intel graphics, and then there are some SKUs that are Arc.

Journalist 6: What about Iris, is that Intel Graphics?

Mitch Lum: Iris is a brand that we EOL'd.

Journalist 6: Is that dead?

Mitch Lum: Yeah, we end-of-life'd that brand... probably three generations ago.

Journalist 6: I'm really up to speed on that, I was asking for a friend [laughs]

Client Product Team Lead: Just to the previous point, that 'X' is there for that reason. Anything with an 'X', like X9 [or] X7, is going to have Arc integrated graphics.

Mitch Lum: So if it doesn't, there's no Arc.

Client Product Team Lead: It'll be Intel Graphics, the 4 Xe version of the...[inaudible]

Mitch Lum: There's one nuance of that at the Core Ultra 5 level with the 10 Xe/ B370, the one place you might see that is in mobile workstations.

Damien Triolet: The smaller, non-Arc Intel graphics are using exactly the same technology, but [with a] smaller configuration, because of the smaller...

Journalist 6: The same as Arc, or Iris?

Damien Triolet: The same as Arc. Between Arc and Intel Graphics on Core Ultra Series 3 is the same technology. A smaller configuration on Intel graphics, and because of that, you're not going to get the type of experience that we want to be part of the Intel Arc [product line]. We want a certain level of performance to define.

Mitch Lum: So we bifurcated. There's a 4 Xe native configuration, and 12 Xe... Because not everybody needs that 'Arc' level of performance. If you think about your general purpose IT machine not doing a lot of 3D rendering.

So, customers have said, "Hey, can we get something that has more modest graphics", or systems that are putting in discrete graphics [...] So, on Intel Graphics, you're just running Office and that kind of stuff. I mean, you can actually run games at like, okay, [framerates], but it's just not really [something to] write home to your friends about.

Client Product Team Lead: So all the IPs in terms of the system. IPs, in terms of systolic AI [...] are still very powerful AI engines. So, it will be used for a lot of AI use-cases, and all the other basic graphics.

Tom's Hardware: I just want to clarify, it's Xe3 across the board, right?

Mitch Lum: Yes, absolutely.

Journalist 6: One of the numbers that jumped out was the battery savings [while] streaming, or energy efficiency. Can you shed a little more light on those processes? And, how much was there [inaudible].

Shalini Singh: Basically, the way we measured it... the process was simple. We just did a pure rundown, right? Just ran the workload until the time the laptop died.

Journalist 6: [What about] the engineering process behind getting that efficiency?

Damien Triolet: Part of the role [...] on the engineering side was to make the communication between the different engines, the media engine, the CPU, the display engine, more efficient. And, for example, when we do a lot of the video processing, we don't need to activate the GPU part at all. The media engine can communicate directly with the display engine. And these are the types of things that we use to reduce...

Mitch Lum: Architectural [changes] impacts graphics. Media and display would be a consolidated tile. Now, our graphics tile is for rendering and the [inaudible], whereas we have display and media on the actual compute tile, so you can power the IP that you don't even need at all, much more power efficiently.

Client Product Team Lead: and also, [regarding] the overall architecture, your P-Cores... Now we have [them] in the performance cluster, which has access to L3 cache, so when you need to balance power, you don't always default to the P-cores. You can run it on the E-cores. And we also bring the Lunar Lake efficiency cluster with it. So a lot of those are just building on all the technologies which we already have. And we are also doubling down on some platform technology. So Intel Intelligent Display [and] Intel Wi-Fi cut [a] significant amount of power when it runs. So all of those are going to be exclusive to Intel platforms, and they help displays a big part of [inaudible].

Shalini Singh: The low-power island basically adds to what he's saying. We started this journey back with Meteor Lake. [...] You know, the low-power island had 2 E-Cores, then with Lunar, we had doubled the quantity, and we had four [E-cores in Lunar Lake]. This time, we have eight [total E-cores, including LP-E]. So it's continuing to be like a significant portion, where workloads can now be sustained on that low-power island.

Journalist 7: Usually, we have a process watch, will we hear a little about EVO?

Client Product Team Lead: Yes. Short answer. In this generation, [...] our focus is going to be more quality than quantity. So there will be less EVO devices, but we will have stricter specs to showcase the best.

Journalist 7: You said, stricter, stricter, KPIs, like, what do they exactly look like?

Client Product Team Lead: Like in terms of resolution, in terms of battery life [...] instead of having two different cut-off points [For EVO SKUs], there will be [fewer] exceptions.

Journalist 8: When you guys were doing the comparisons between Lunar and Panther Lake, you said 77% faster gaming and 60% for other stuff, but we looked into the small print, it says that you're comparing the V SKUs to other H SKUs. One's more GPU efficient, one's more AI efficient, so [inaudible]

Damien Triolet: So if you go to the press deck, you're gonna see the comparison between all of them, right? In the keynote, we simplified. We take a few points. In general, you will see that both AI and gaming are faster [in Panther Lake SKUs] than the previous generation. The 70% plus [performance uplift in Panther Lake] for gaming, for example, applies to both Arrow Lake H and Lunar Lake, so you will see the same type of improvements. On AI in general, the improvement versus Alder Lake H is higher than versus Lunar Lake, especially on the NPU side. In Lunar, we're already at a relatively large NPU.

Mitch Lum: And maybe implicit to your question is the PL setting. So when we compare [the] Core Ultra Series Three versus the 288V. It's at 25 watts.

Shalini Singh: It's basically peak performance, right? So Lunar Lake's peak performance was about 25 to 30 watts.

Mitch Lum: So we compared it, apples to apples. So 25 watts versus Lunar Lake, 45 watts versus Arrow Lake.

Client Product Team Lead: And this is where the core name versus the [inaudible] company, all of Lunar Lake had the V-series. Arrow Lake had H. So when we compare talking Lunar Lake, it's always going to be V. right, and then that's why Damien said we have both numbers in the press deck, you'll see some Arrow Lake H.

Journalist 8: So all of these [CPUs] that are announced today, H?

Client Product Team Lead: We have H and non-H. So in Panther Lake, there are products that have an H, and there [are] products without an H suffix.

Journalist 8: Okay, what [are] the ones without the suffix?

Shalini Singh: You just have a number. You do not have [an] H. But think about it this way, the product transition for Panther Lake is basically [that] there were two products we announced last year, Lunar Lake and Arrow Lake, right? Arrow Lake was generally for higher performance, more cores, at a little bit of a higher PL1 (sustained power limit). Lunar Lake was more [about] power efficiency, battery life, good performance, and a lower power level. Both of those are basically transitioning to Panther Lake this year. So whether Yes, so whether you go with the suffix or...

Journalist 8: Why do you guys make this so confusing? So, now this series will have H, and...

Shalini Singh: Non-H. We're not doing any more V-series.

Journalist 8: [...] There's no more HX?

Client Product Team Lead: HX is in a higher category. [...] So the way to think about it is, U,V,Y,H, all of this creates an overlapping power level. With Panther Lake, we are going at a similar TDP. So all of Panther Lake starts at 15, and goes to 45 or 65 [Watts] depending on the SKUs. So what we did is we tried to simplify it, and in that range, you just have H or no [H suffix], that's it. So you don't have any more V's, no Y's, no U's. [...]

Mitch Lum: Just to tie up on that, so the parts with no suffix are basically eight-core based. You know how we show the die package combination? Those are eight-core based, the 16-core base all starts with H, so that's how you know you get more [inaudible].

Tom's Hardware: So, on gaming performance [of Panther Lake], you were comparing to Arrow Lake H, which I believe is not at ISO power...

Shalini Singh: It is.

Tom's Hardware: Is that ISO at 45?

Shalini Singh: Yes, 45. We're actually comparing to Arrow [Lake], we're comparing to Lunar [Lake], we're also comparing to Nvidia.

Tom's Hardware: Do you have numbers at peak power, for 115 [Watts] on Arrow Lake H, compared to 65 [Watts] on...

Shalini Singh: No, we're doing it at sustained PL1. So, 45 to 45, 45 to 25, and then for the competition, whatever their out-of-the-box sustained power is.

Tom's Hardware: Yeah, I was just curious for efficiency.

Client Product Team Lead: Arrow Lake H peak power, as in the high-end PL1 is 55...

Tom's Hardware: Max turbo?

Client Product Team Lead: 110 et cetera, will be the system power, if you're putting in discrete graphics. The CPU itself is 55 [this is incorrect, as many Panther Lake SKUs have PL1 of 65W].

[Session ends]

Stay tuned to Tom's Hardware Premium, where we'll be posting much more from CES 2026 over the coming days.

Audeze has updated one of the best gaming headsets we've tested — the Audeze Maxwell 2 sounds better for both gaming and music, has better noise-canceling tech in its mic, and is (hopefully) more comfortable and versatile, design-wise.

The big update in the Maxwell 2 is the inclusion of Audeze's SLAM technology, which, according to the company, "produces heightened spatial immersion and precision along with punchier bass response." This tech first debuted in the brand's flagship CRBN2 electrostatic headphones in 2024, and is now part of the Maxwell lineup. It seems to have the same excellent 90mm planar magnetic drivers as the original Maxwell — which offer an impressive frequency response range of 10 - 50,000 Hz (much wider than the typical gaming headset frequency response range of 20 - 20,000 Hz).

Tom's Hardware's Jeffrey Kampman had a chance to stop by Audeze's suite at CES and check out the headset, and he was especially impressed by the detail and positional audio — hopefully this will be even better for pinpointing directional footsteps or gunshots in games. As for music listening, he found the headset to be very transparent to the source material — he said the soundstage was perhaps a little stuffy (somewhat expected with the large, closed-back design) but that everything sounded great overall.

The headset also features improved AI-powered noise canceling technology for its detachable boom microphone — though, no word on whether this also applies to the built-in mics. Jeffrey said the sidetone sounded good, though noise-canceling was a little difficult to evaluate on the show floor.

Design-wise, the Maxwell 2 looks pretty similar to its predecessor — large, round plastic earcups with plush leatherette-covered memory foam earpads and a sturdy steel headband. It does have a wider, ventilated head strap, and Audeze says the earpads now have more space to accommodate a wider range of users. It's still a pretty hefty headset, however — it weighs 1.23 pounds (560g), which is almost 2.5 ounces (70g) heavier than the original Maxwell (1.08lbs / 490g). So we hope those nods to comfort end up paying off. The headset also has magnetic side plates that can be swapped out for ones with custom designs, though there are no alternate options available at launch. Like its predecessor, it features 2.4GHz wireless, Bluetooth, and wired connectivity, and gets up to 80 hours of battery life on either of its wireless connections.

The Maxwell 2 isn't the only planar magnetic gaming headset we've seen on the show floor: Asus also showed off its ROG Kithara, which is part of a collaboration with HiFiMan. The wired ROG Kithara features 100mm planar magnetic drivers and an open-back design, similar to the Audeze LCD-GX. It comes with a detachable full-band MEMS boom microphone and swappable 4.4mm, 6.3mm, and 3.5mm single-ended plugs, as well as a USB-C to dual 3.5mm adapter.

The Maxwell 2 is available to order now, and costs $329 for the PlayStation version and $349 for the Xbox version (both versions are also compatible with PC, Android, iOS, and Nintendo Switch).

MSI electrified CES 2026 with the debut of the long-awaited MEG X870E Unify-X Max motherboard—set to outshine the best motherboards on the market. Bursting with overclocking-focused features, the legendary Unify series finally storms onto AMD’s AM5 platform, after a four-year hiatus.

On the AM4 platform, MSI made waves by first releasing the budget-friendly MEG B550 Unify and B550 Unify-X, later stepping up with the premium MEG X570 Unify and X570S Unify-X Max. Now, with AM5, MSI is making an audacious leap straight to the X870E chipset—no whispers yet of more affordable models on the horizon. The hotly anticipated MEG X870E Unify-X Max is the fifth AMD chapter in the storied Unify saga.

True to the Unify legacy, the MEG X870E Unify-X Max makes an unmistakable statement with its commanding all-black aesthetic, punctuated by sleek, reflective highlights. As with all X-designated models, this powerhouse comes with just two DDR5 memory slots—purpose-built for memory overclocking. The minimalist two-DIMM layout improves signal integrity and slashes latency, ensuring rock-solid stability even as you push memory frequencies to the absolute limit. While the official specifications remain under wraps for now, expect the MEG X870E Unify-X Max to surpass DDR5-10000 effortlessly, provided your Ryzen CPU can keep pace.

Forget the old limitations of two-DIMM motherboards. With cutting-edge 64GB DDR5 modules now available, the MEG X870E Unify-X Max lets you pack in a colossal 128GB (2x64GB) of ultra-fast memory—more than enough muscle for power users, creators, and extreme enthusiasts alike.

Engineered for hardcore enthusiasts and relentless overclockers, MSI crafted the MEG X870E Unify-X Max with only the finest components. MSI has sculpted this motherboard from an eight-layer, server-grade PCB, anchored by a robust dual-rail power system. Its formidable 18+2+1-phase power delivery, paired with dual 8-pin EPS connectors, ensures your Ryzen CPU receives ample power even under the most extreme overclocking conditions.

The motherboard boasts dual PCIe 5.0 x16 expansion slots alongside two PCIe x4 expansion slots. The primary x16 expansion slot is equipped with MSI’s EZ PCIe Release, allowing you to lock or unlock your graphics card with a single button press. Storage is equally plentiful with five blazing-fast M.2 slots featuring Frozr heatsinks and tool-free EZ M.2 Clip II installation, plus two SATA III ports for legacy storage drives.

The MEG X870E Unify-X Max has two USB4 (40 Gbps) Type-C ports, one USB 3.2 Gen 2 (10 Gbps) Type-C port, and eight USB 3.2 Gen 2 (10 Gbps) Type-A ports. The motherboard also includes old-school PS/2 ports for overclockers as well as 5 Gigabit Ethernet and WiFi 7 wireless connectivity.

No Unify package would be complete without the MSI Tuning Controller. The compact remote puts real-time overclocking tweaks right at your fingertips, and also lets you monitor and diagnose your motherboard effortlessly using QCODEs.

MSI has yet to reveal official pricing or availability for the MEG X870E Unify-X Max. However, its recent announcement suggests eager enthusiasts won’t have to wait much longer for a full reveal of specifications and launch details. For context, MSI’s Intel-based MEG Z890 Unify-X currently retails at $699.99—so it’s safe to expect the X870E Unify-X Max to land in a similar premium price bracket.

At CES 2026 in Las Vegas, Keychron officially unveiled its Nape Pro trackball, designed in conjunction with Gizmodo Japan. The quirky-looking device is a trackball that nestles right up against any of the best gaming keyboards to give you instant access to a scrolling trackball and six programmable buttons without moving your hands.

Tom's Hardware stopped by Keychron in Las Vegas to get a closer look. The Nape Pro can sit either under your space bar or off to one side to perform the duties of a more conventional trackball/mouse. The Nape Pro is pretty diminutive, not taking up much more real estate than a regular mouse, and its rectangular shape actually makes it a bit more discreet for some setups.

As reported by Yanko Design, the Nape Pro's party piece is its OctaShift software, which can detect the orientation of the device (portrait or landscape) and adjust accordingly. YD reports the Nape Pro has a Realtek chip with a 1 kHz polling rate and a PixArt PAW3222 sensor to give the best wireless gaming mouse a run for its money. You can connect via Bluetooth, 2.4 GHz dongle, or USB-C. Firmware is ZMK, and Keychron also reportedly plans to release the 3D files for the keyboard. That means that users will be able to 3D print their own stands, caps, or other accessories.

The intense customization is a recipe that has served Keychron well in its other product offerings, so it's great to see that same philosophy on display here. The whole thing looks really unique and enticing, and I have to say that the idea of enhancing a desktop setup with a much more ergonomic trackball nestled alongside a keyboard is a great idea. Even as I type this, I'm flexing my thumbs in anticipation of how it might feel to reach for this trackball to do some scrolling.

Keychron hasn't shared any official information about pricing or release date for the Nape Pro. Elsewhere at CES, the company took the covers off its new Q UItra wireless keyboards with up to 660 hours of battery life and 8K polling. If that's not refined enough for you, the company also has a keyboard made of marble on display.

CES 2026 was dominated by AI and its accompanying hype, as evidenced by the keynotes from heavy industry hitters like Nvidia and AMD. But the seemingly insatiable AI-driven demands on the RAM and storage markets also meant that we saw no new GPUs and very few SSD or RAM launches at the show. And laptop and PC makers were very cagey about pricing – although we do know that new models will be more expensive. There was still plenty of innovative and eye-catching tech to be found in Las Vegas this year, even if we had to look in some unexpected places to find it.

Intel’s Panther Lake launch and AMD’s Gorgon Point did give laptop makers a reason to launch new models, and Dell took the opportunity to backtrack on last year’s rebranding, with impressive new XPS laptops. The lack of new GPUs meant that things were a bit boring in the realm of gaming PCs, but Lenovo unrolled some excitement with its expandable Legion Pro concept, and Asus decided RGB wasn’t enough, wrapping its latest prebuilt in hologram-generating fans. Not to be left out, ASRock jumped into the AIO cooling ring with a dual-pump concept with a hologram that hovers over your motherboard.

These are the best 18 products that the editors of Tom's Hardware found at CES 2026. — Matt Safford

Best CPU: Intel Core Ultra Series 3 "Panther Lake"

Intel Core Ultra Series 3 "Panther Lake"

CES 2026 was light on CPU announcements. We heard about Ryzen AI 400 ‘Gorgon Point’ chips, which are just a refresh of the existing Strix Point lineup, and the much-discussed Ryzen 7 9850X3D. Intel finally unveiled Panther Lake, which is a series it has been discussing for the past six months, and for years in the context of its 18A process. Panther Lake wins, but make no mistake; it doesn’t win by default. Even in a year light on CPU announcements, Panther Lake takes the crown not only because of its impressive performance targets, but also because of what it represents for Team Blue.

With the death of 20A on Intel’s road map, 18A has served as the inflection point, the point at which the ambitious manufacturing strategy Intel laid out close to five years ago actually reaches completion. Moreover, it represents the point in which Intel is bringing together the disparate approaches we’ve seen on mobile over the past several generations. It’s designed to be the AI PC enablement of Meteor Lake, the efficiency of Lunar Lake, and the performance of Arrow Lake-H, all wrapped up in a package that boasts a massive iGPU that Intel says can rival entry-level discrete graphics.

Across both desktop and mobile, Intel’s offerings have carried an asterisk for the past few generations. Lunar Lake is the most efficient x86 chip on the market, but it takes a clear backseat in performance. Arrow Lake is remarkable in its productivity performance, considering the efficiency and architectural approach, but it lags in gaming performance. Panther Lake, according to Intel’s claims, carries no asterisk. We’ll have to see if those claims hold up once the chips are available, but if there’s any announcement out of CES that holds some promise for the future, it’s Panther Lake. — Jake Roach

Read more:
🔴 Intel doubles down on gaming with Panther Lake, claims 76% faster gaming performance
🔴 Intel Panther Lake pre-release testing delivers over 80 fps in Cyberpunk, 100 fps in F1

Best GPU: MSI RTX 5090 Lightning Z

MSI RTX 5090 Lightning Z

In the midst of a disappointing CES for consumer GPUs, we appreciate that MSI went all-out with one of the most extreme custom RTX 5090s yet. The RTX 5090 Lightning Z removes most obstacles to overclocking Nvidia's ultimate gaming card to the limit, thanks to not one but two ATX12V-2x6 connectors and a crazy 40-phase VRM that's capable of delivering up to 1000W of power to the GB202 GPU with the stock VBIOS.

A full-coverage copper cold plate and a 360-mm radiator help remove all of the resulting heat, and a 3oz PCB puts plenty of copper on each trace for extra current capacity and heat dissipation. And if you're pondering a shot at the top of the HWBot leaderboards, MSI will grant you access to a warranty-voiding VBIOS that allows feeding up to 2500W of power into the GPU. It's the most ambitious RTX 5090 out there, and with 17 #1 spots claimed on overclocking leaderboards, it's easy to crown the Lightning Z the best graphics card (GPU?) of CES 2026. — Jeff Kampman

Read more: CES award outs MSI's monstrous 1600W RTX 5090 Lightning GPU

Best Laptop: Dell XPS 14 and XPS 16

Dell XPS 14 and XPS 16

It's rare that a company reverses course from a misstep. It's rarer that they apologize in the process. But for us, what's most important is that the new Dell XPS 14 and 16 look like they may be a return to form for the storied brand.

The new chassis designs are thinner and lighter than their predecessors, and they undo the previous version's biggest design flaws. The new XPS 14 and 16 have full-function rows, ditching the capacitive touchbar that has plagued previous XPS and Dell Premium notebooks. Dell also added etching in the glass palm rest in order to better outline where the touchpad ends. We also appreciate the innovations Dell is bringing to its XPS screens. There's a 1920 x 1200 non-touch display option with a 1 - 120 Hz variable refresh rate, which the company says is a first. That extremely low refresh rate should help save power when static images or text are on the screen. Alternatively, you can opt for a bright tandem OLED touchscreen ranging from 20 - 120 Hz.

Both systems will use Intel's Core Ultra Series 3 chips and are ditching discrete GPUs thanks to Panther Lake's powerful discrete graphics. We're excited to test these systems and see if all of the new design and engineering decisions pay off. — Andrew E. Freedman

Read more: Dell brings back XPS laptops

Best Gaming Laptop: Asus TUF Gaming A14

Asus TUF Gaming A14

With no new consumer GPUs at CES 2026, all eyes have been on integrated graphics. I expected to see someone be bold enough to put Intel's Core Ultra with 12 Xe Cores into a gaming laptop, but it was Asus using Strix Halo, where integrated graphics made it into a notebook.

Asus has done something similar before — it put a Strix Halo chip in the ROG Flow Z13 last year. But the TUF Gaming A14 is both a more mainstream system and offers a better form factor for gaming. This system goes up to 85W of total thermal capacity in turbo mode, and uses USB Type-C charging for 100W of power.

We've had versions of this laptop on our best list before, so we're excited to see what it can do with Ryzen AI Max+. But with no pricing yet, we're hoping Asus can keep it reasonable for gamers on the go. — Andrew E. Freedman

Read more: Asus puts Strix Halo in the TUF Gaming A14

Best Gaming PC: Asus ROG G1000

Asus ROG G1000

Maybe due in part to the rising cost of RAM, there weren't a ton of exciting prebuilt desktops at CES 2026. But you couldn't see the Asus ROG G1000 and forget it.

This gigantic 104-liter case is covered in LED-covered fans that create holograms, making this one of the most customizable — and unmissable — PCs around. You might think it's cool, you might think it's tacky, but you sure won't forget about it.

Asus packs this system with high-end parts, going up to an ROG Astral RTX 5090 and AMD Ryzen 9950X3D. The company puts the 420 mm radiator in a new chamber that sucks fresh air from outside the case (though we did see HP do that before).

There's no price or release date for this huge, hologram-covered tower, but it will surely be expensive. That's the price of turning heads. — Andrew E. Freedman

Read more: Asus ROG G1000 gaming PC is covered in holograms

Best Gaming Monitor: MSI MPG 341CQR QD-OLED X36

MSI MPG 341CQR QD-OLED X36

The MSI MPG 341CQR QD-OLED X36 is a 34-inch gaming monitor using fifth-generation QD-OLED panel technology. This new QD-OLED panel aims to right some of the wrongs of previous iterations, namely, blurry text and fringing. Thanks to a new RGB stripe subpixel layout, the MPG 341CQR QD-OLED X36 achieves its goal and then some.

The monitor boasts a typical brightness of 300 nits (1,300 nits peak) and uses a DarkArmor Film to boost black levels by an additional 40%. The 34-inch panel features a 3,440 x 1,440 resolution wrapped around an 1800R curve. And importantly for the enthusiast community, the refresh rate tops out at 360 Hz with a response time of 0.03 ms.

MSI has infused the MPG 341CQR QD-OLED X36 with two HDMI 2.1 ports, one DisplayPort 2.1a port, and a USB Type-C port (DisplayPort Alt Mode with 98W PD). When it comes to your peripherals, there are two downstream USB-A (5 Gbps) ports and a 3.5 mm headphone jack. The monitor also includes the MSI OLED Care 3.0 AI Care Sensor, which uses an NPU-based IC to detect whether a person is sitting in front of the monitor. It can then choose to implement power-saving routines or initiate automatic pixel-refreshing routines to prolong panel life. — Brandon Hill

Read more: MSI's fifth-generation QD-OLED gaming monitor embraces RGB stripe subpixels

Best Productivity Monitor: Dell UltraSharp 52 Thunderbolt Hub Monitor (U5226KW)

Dell UltraSharp 52 Thunderbolt Hub Monitor (U5226KW)

While we all love gaming monitors, sometimes we need something more geared toward productivity tasks. In that case, Dell is answering the call with the UltraSharp 52 Thunderbolt Hub Monitor. The first thing that catches your eye is the sheer size of the monitor, which in this case, stretches 52 inches across.

The next headlining spec is the resolution, which comes in at an incredible 6,144 x 2,560 and 129 PPI. In other words, the UltraSharp 52 Thunderbolt Hub Monitor makes the 45- to 48-inch Dual QHD (5,120 x 1,440) monitors seem rather quaint. The refresh rate is “only” 120 Hz, but we have to consider the number of pixels that we’re dealing with here. And we seriously doubt, even with today’s best graphics cards, that many people would attempt playing AAA games at maximum resolution.

The UltraSharp 52 Thunderbolt Hub Monitor uses IPS Black panel technology that helps it achieve a 2,000:1 contrast ratio, which is double that of typical IPS panels. And given its productivity roots, the monitor features two HDMI 2.1 ports, two DisplayPort 1.4 ports, four downstream USB-A (10 Gbps) ports, three downstream USB-C (10 Gbps) ports, one Thunderbolt 4 port, and a 2.5 GbE LAN port. There’s also a KVM switch, allowing the UltraSharp 52 Thunderbolt Hub Monitor to be used with up to four computers simultaneously.

Dell has managed to pack a lot into the UltraSharp 52 Thunderbolt Hub Monitor, and it is priced accordingly. It’s available now for $2,899.99. — Brandon Hill

Read more: Dell debuts monstrously large 52-inch 6K ultrawide monitor

Best Power Supply: MSI MPG AI1600TS PCIE5

MSI MPG AI1600TS PCIE5

MSI’s latest flagship power supply, the MPG AI1600TS, isn’t only 80-PLUS Titanium-rated, it boasts what the company calls “GPU Safeguard+,” which monitors every pin on each of the PSU’s two native 12V-2x6 power connectors. As soon as there’s an uneven load across each cable, a protection measure is tripped, alerting the user to the issue via software or a built-in alarm. Additionally, if you’re away from your desk when a fault is detected, your PC will power off after three minutes, preventing any further damage to your precious components. Given the number of melted power connector horror stories that we’ve published over the past year and change, this is a more than welcome new feature to protect your PC.

The AI1300TS PCIE5 comes packed with the same features, and sporting two 12V-2x6 connectors, it’s likely that if you plan to run or overclock a high-end graphics card, you’ll want the larger capacity variant. The PSU is PCIe 5.1 ready, in addition to sporting server-grade components like 105°C Japanese capacitors and SiC MOSFETs, which MSI claims generate less heat in operation. With extra additions like a zero-RPM fan mode, as well as stable power delivery, reducing fluctuation on the 12V rail to ±1%.

The PSU is set to ship alongside 12V-2x6 cables with a dual-color design, ensuring that you have a good visual guideline for when they are seated correctly, as well as pre-installed cable organizers to help you cable manage your next big build.

There’s no word on pricing or a release date for the MSI MPG AI1600TS PCIE5, but it’s sure to become a popular option amongst enthusiasts for its wide array of safety features. — Sayem Ahmed

Read more: MSI introduces GPU Safeguard tech on its latest PSUs to fend off RTX 5090 meltdowns

Best Motherboard: MSI MEG X870E Unify-X MAX

MSI MEG X870E Unify-X MAX

Motherboards were an interesting bunch for CES, with mostly refreshes on the menu. However, we saw a couple of options worth considering from Asus with the new Crosshair X870E Glacial and Dark Hero motherboards, but I have a passion for boards that lean towards overclocking (and chances are they do most everything else well, too). MSI’s new X870E Unify-X MAX blends overclocking, performance, and style, making it one of the best new refresh motherboards for AMD’s existing Zen 5 processors and upcoming models like the Ryzen 9 9950X3D2.

The Unify-X MAX is decidedly designed for overclocking with a 2-DIMM configuration and the included tuning controller. MSI mentioned memory support to over DDR5-10000, which is incredibly high for the platform (and likely accomplished via 8000-series APUs). That, combined with the robust 18-phase 110A SPS MOSFETs for Vcore, the limit is your cooling and patience for tweaking, not the motherboard. MSI makes pushing the limits easier with the included Tuning Controller for on-the-fly overclocking adjustments, saving time on reboots and reducing trips to the BIOS.

In addition to being dialed in for overclocking, it looks the part of a premium motherboard with oversized VRM heatsinks and large plate-style heatsinks covering much of the PCB. While it is designed for overclocking, you get some of the best features the platform offers, including 5 M.2 sockets (2x PCIe 5.0), Wi-Fi 7 and 5 GbE, USB4 (40 Gbps), and several EZ features that make installing and removing parts easier.

Pricing or availability were’t mentioned, but we expect the MEG X870E Unify-X MAX to fall in the premium mid-range space and land on store shelves soon. Keep an eye out for our review to see if it really measures up to the hype. — Joe Shields

Read more: MSI unlocks the full power of AMD CPUs with new MEG X870E Unify-X Max motherboard

Best Concept: Lenovo Legion Pro

Lenovo Legion Pro Rollable

Lenovo has been steadily pushing out foldable and rollable OLED laptops in recent years. But the Legion Pro is the company's first gaming-focused rollable and its first with a horizontally expanding screen. Starting out in a traditional, backpack-friendly 16-inch chassis, the motorized screen can expand to 21.5 and 24 inches.

As this rollable is still a concept, we didn't actually get to play any games on it. And the motorized mechanism still needs work to keep the expanding screen taught. But Lenovo has solved this problem with previous rollables, so I'm confident it can do the same with any future iterations of its gaming rollable.

As someone who never feels like there's enough screen in front of me, whether I'm gaming or working, the idea of being able to carry around a high-resolution 24-inch screen in a 16-inch sized laptop has me excited for the future. And with the rising price of RAM and storage and companies talking about re-releasing years-old hardware to meet consumer demand in a market where AI is consuming seemingly all of the cutting-edge silicon, I'll take whatever future-looking excitement I can find. — Matt Safford

Read more: Lenovo's CES concept devices include a rollable ultra-wide OLED Legion gaming laptop

Best Keyboard: Corsair Galleon 100SD

Corsair Galleon 100SD

The problem with keyboards — even full-size ones with extra macro keys — is that they just don't have enough keys for some people, which is why Stream Decks and macro pads are so popular (with a certain crowd, anyway). Well, Corsair acquired Elgato nearly eight years ago, so it's actually a little surprising we had to wait this long to see a Stream Deck integrated into a keyboard. But it's finally here in the Corsair Galleon 100 SD — a full-size gaming keyboard with a built-in Stream Deck. It's got 12 customizable LCD keys, two customizable dials, and a customizable full-color LCD screen (720 x 1280 pixel resolution) for stats or clocks or whatever it is people put on their Stream Deck screens. Like standalone Stream Decks, the LCD keys are also fully customizable — you can program to do just about anything, including multiple, timed actions (like a very powerful macro key), and you can also customize the display.

Of course, it's not just a Stream Deck, and Corsair is keeping keyboard enthusiasts in mind with the design. The Galleon 100 SD features an aluminum top plate, a gasket-mounted hot-swappable PCB, and six layers of sound-dampening foam in the case. It will ship with Corsair's pre-lubed MLX Pulse linear mechanical switches, and will feature a polling rate of up to 8,000 Hz and FlashTap SOCD, as well as a separate game mode and regular macro recording in Corsair's Web Hub (though it seems the Stream Deck will be managed through Elgato's Stream Deck app). Pre-orders are open now, and Corsair expects to start shipping around the end of January. The Galleon 100 SD will cost $349.99. — Sarah Jacobsson Purewal

Read more: Corsair's Galleon 100 SD is an LED-lit keyboard with a built-in Stream Deck

Best 3D Printer: Creality Sparkx i7

Creality Sparkx i7

Creality is a well-known 3D printer brand that regularly wins spots on our list of the Best 3D printers on the market, and the company’s latest Sparkx i7, which debuted at CES, is designed to bring the hobby to beginners and entry-level enthusiasts through a suite of easy-to-use features and software. We’ve had this 3D printer in our test labs and just released a full review, where it won our Editor’s Choice award, taking the top spot in our recommendations for the Best 3D printers for beginners from the Bambu Lab A1 mini.

Why? The Sparkx i7 couples leading-edge capabilities for the budget market with features designed to simplify 3D printing for beginners, which begins with the software. The CubeMe software is an AI-powered tool that automatically converts 2D photos into 3D-printable models, reducing the complexity of what can be, at times, overwhelming CAD programs used for model design.

The simplicity ethos also extends to the hardware, with a tool-free quick-swap hotend and filament cutter that are easily accessible, simplifying maintenance tasks. The printer hardware features a strong roster of capabilities, such as a 260x260x255mm build volume, print speeds of 500 mm/s with an acceleration of 10,000 mm/s, a 300 °C hotend, a built-in AI camera, and the ability to print with PLA, PETG, and TPU filaments. The printer also supports multi-color printing with a focus on reducing purging waste by 50%, addressing one of the traditional drawbacks of multi-color printing. — Paul Alcorn

Read more: Creality SPARKX i7 Review: Not just another color bedslinger

Best Mouse (Pointing Device?): Keychron Nape Pro

Keychron Nape Pro

The problem with trackball mice is that they're still mice — just mice that don't move. So they probably shouldn't be shaped like mice, right? Right (says Keychron).

The Nape Pro is a small wireless standalone trackball... device, designed to bring flexible trackball functionality to mechanical keyboard users. It features a large center trackball with a click wheel around it, surrounded by four programmable buttons and two side keys. Tom's Hardware's Matt Safford had a chance to check it out on the show floor, and he said the ball movement and wheel felt great — "some of the other buttons, not so much. But I think that's partially because those buttons are designed to be kind of blindly mashed from the side." (Fair enough.) The slim design means you can put it below or beside (or, I suppose, on top of) your keyboard, so you can move your cursor without moving your hands (too much). And, of course, the buttons are programmable.

The Nape Pro started out as a collaboration between Keychron and Gizmodo Japan, but it has yet to launch in the U.S. Pricing hasn't been confirmed, but a Keychron rep said it will probably cost "around $50." — Sarah Jacobsson Purewal

Read more: Keyboard giant Keychron unveils new Nape Pro trackball with programmable buttons

Best Headset: Audeze Maxwell 2

Audeze Maxwell 2

One of our favorite gaming headsets is getting a big audio upgrade that should make it better for gaming and music — even though it's already great for both. Audeze launched the Maxwell 2 at CES 2026 — this is the second generation of the premium audio brand's foray into gaming, and it comes with all sorts of improvements to, well, audio. The main improvement involves Audeze's "SLAM" technology (patent-pending), which allows for better spatial immersion and precision and was first introduced in the company's $5,995 flagship CRBN2. This should make the headset's 90mm planar magnetic drivers sound even better and make it easier for you to pinpoint the exact direction you're hearing those footsteps from. Tom's Hardware's Jeffrey Kampman had a chance to check out the headset in person and he confirmed that details and positional audio sounded spectacular and very transparent to the source material.

On the other side of audio, the headset also features an improved noise-canceling microphone system, which should be interesting to test out — we found the mic on the original Maxwell to be pretty good, and the AI-powered noise canceling to be decent, so (hopefully) it can only get better. The Maxwell 2 also has a new, wider suspension strap and a new earpad design with more space, for a more comfortable experience. It's still pretty hefty, though. Like the original Maxwell, it will have both 2.4GHz wireless and Bluetooth connectivity (as well as wired — Audeze is, after all, still an audio company), and it will get a solid 80 hours of playtime over either connection type.

The Maxwell 2 is available now, starting at $329 for the PlayStation version and $349 for the Xbox version (both versions support PC, as well as Android, iOS, and Nintendo Switch). — Sarah Jacobsson Purewal

Read more: Audeze launches Maxwell 2 headset, with better spatial immersion and bass

Best PC Case: CyberPowerPC MA-01

CyberPowerPC MA-01

Case aesthetics are always subjective, but if you like clean, swooping lines, curved glass, hidden cables, and interesting innovations, CyberPowerPC’s MA-01 is one of the most interesting PC enclosures in years. Arriving first as a prebuilt chassis, and later as a standalone case in three different colorways, the case features three knobs for dialing in exact shade, brightness, and effect for its RGB lighting, louvered vents to direct air intake directly toward key components, and a woven steel mesh top exhaust for cutting exhaust noise and resonance. It’s also generally the prettiest PC chassis I’ve seen since the launch of the wood-accented Fractal North in 2022.

That’s in part because the MA-01 is just oozing with examples of attention to detail, from the orange accents on the front-panel ports and the tactile RGB control knobs, magnetic brackets to hide bracket and screw mounts, and RGB that hides the actual lights, while bathing aspects of the case with warmer, indirect light. And the flowing lines surrounding the motherboard are actually angled vents designed to keep cool air moving over your warm-running components, rather than entering from one point and exhausting from another without actually doing much in the way of cooling. I’m not sure if the MA-01 will take the prettiest PC case spot on our best PC case list, but it’s the strongest contender in years, and I’m excited to see what CyberPowerPC does with any new cases in this lineup. Because you probably don’t name something MA-01 without there being a strong possibility for an MA-02 and MA-03. But let’s not get ahead of ourselves. The MA-01 looks to be one of the most interesting cases of 2026. — Matt Safford

Read more: CyberPowerPC's MA-01 comes with analog color control knobs, woven steel mesh, and slanted airflow

Best Cooler: Thermaltake Retro Ultra AIO

Thermaltake Retro Ultra AIO

In an age where AIOs and even heatsinks need to work hard to stand out, Thermaltake did so by looking to the past, drawing on classic tech aesthetics and modern design, with a detachable display that looks like one of my early PCs with an old-school CRT monitor on top. But that retro styling looks at home in the accompanying Retro 260 TG (MicroATX) and Retro 280 (Mid-tower) chassis, reminiscent of the 80s and 90s.

The throwback AIO comes in 240mm and 360mm sizes, both using single-frame fans (the two or three fans are one unit) with RGB lighting circling the outside. Details are scarce on the specifications and internals. The 3.6-inch TFT LCD is customizable, displaying system information, retro graphics, and your own custom theme via TT RGB Plus 3.0 software. Mounting and removal of the cold plate for maintenance is easy with Thermaltake's four-screw system.

I have to admit, the entire series tugs on the nostalgia strings and is one of the more unique designs we’ve seen at CES. Pricing and availability weren’t mentioned, but Thermaltake tends to offer solid value, which could turn the heads of some Gen X’ers interested in reliving their past with modern underpinnings. — Joe Shields

Read more: Thermaltake goes retro at CES 2026, shows off liquid cooler with CRT-themed display and 80s-style PC cases

Best Toy Tech: Lego Smart Play

Lego Smart Play

Lots of us love Lego – both as kids and adults – for its ability to spark imagination, develop problem-solving skills, and let us build worlds that exist both in our hands and in our minds. But experiments with Mindstroms and remote-controlled vehicle kits, Lego has largely kept the kind of tech hardware we love out of its mainstream products – until the company announced Smart Play here at CES 2026.

Set to debut in march in three Star Wars sets, Smart Play embeds a tiny ASIC into a 2x4 Lego brick, as well as a speaker, a color sensor, an accelerometer, LEDs, and a Bluetooth-based BrickNet to allow pieces and Minifigures to interact with each other, without the need for an app or a hub for setting things up or interacting with compatible tags, minifigures, and other Smart Play pieces. A wireless pad will power up the bricks when they need a recharge, although it’s not clear how often that might be.

It’s early stages for Lego’s new tech, to be sure. And the company seems to be taking a somewhat cautious approach to the tech by rolling it out first in just three sci-fi-themed Star Wars sets. But I’m interested to see what the company does with the platform in the future, and excited to get some hands-on time with Smart Play here in Las Vegas. — Matt Safford

Read more: Lego unveils new smart brick with embedded computer inside

Best Retro-inspired PC Case: Thermaltake Retro 260/360 TG

Thermaltake Retro 260/360 TG

Fresh off a trip back to the future, and building on what Silverstone recently started, Thermaltake entered the retro case market, showing off the new Retro 260 TG and Retro 360 TG cases that blend classic 80s and 90s PC look with modern features and specifications. The new chassis sports retro styling cues on the front that mimic a 3.5-inch floppy disk drive, and it even has two 5.25-inch bays, one of which is a mock optical drive. On top of that, the color is that familiar oxidized plastic off-white that really locks in the retro vibe.

The cases come in two sizes. The Retro 260 TG is the MicroATX option, supporting up to 240 mm AIOs in the top, while the 360 TG is the mid-tower/ATX case, supporting up to 360 mm AIOs in the top or the side. The Retro 260 TG holds up to nine 120mm fans, while the 360 TG handles up to twelve, so there’s plenty of airflow for modern components. The front I/O features modern connectivity, including a single USB 3.2 Gen 2 (10 Gbps) Type-C port, dual USB 3.0 (5 Gbps) ports, and front-panel audio—typical fare for the 21st century.

In line with the retro styling, Thermaltake also announced the Retro 240 and Retro 360 AIOs (which earned our best cooling award at CES) to perfectly match the classic appearance with a small CRT-styled LED on the pump. If you want more modern bling, both cases fit Thermaltake’s 6-inch LCD Screen Kit for the front, adding a touch of modern while still maintaining the retro aesthetic.

Unfortunately, additional details like pricing and the release date are few and far between, and we couldn’t pry the information out of them in the ballroom. Still, we expect to see these available to buy sooner rather than later. We love seeing these retro-inspired cases with modern features on the market, and Thermaltake has done a great job with these chassis (their whole retro ecosystem). At the same time, this makes those who cut their teeth on PCs in the 90s feel old! — Joe Shields

Read more: Thermaltake goes retro at CES 2026, shows off liquid cooler with CRT-themed display and 80s-style PC cases

Cooler Master just introduced a coolant distribution unit, or CDU, with a massive capacity of 900 to 2,500 watts. The Tom’s Hardware team saw the rather compact unit at CES 2026. The company says it is designed for CPUs, GPUs, and ASICs, balancing the ease-of-use of AIOs and the maximum cooling performance you get from open-loop systems. Each unit measures around 10 x 10 x 19.6 inches (255 x 255 x 497 mm) or around 32 liters and weighs between 12 and 18 kg (between 26 and 40 pounds), so you’re not going to fit this inside the case of your gaming PC. Instead, it’s built for specialized desktop systems that need robust cooling, like those that feature multiple GPUs for AI applications.

Another thing that makes the Aquagate MAX an excellent solution is its modularity and scalability. You can get this CDU with either a 240mm, 280mm, or 360mm radiator, allowing you to increase its cooling capacity without drastically taking up more space. And if that is not enough, Cooler Master says that you can connect multiple units in both serial and parallel connections, allowing you to expand your cooling capacity as you add more units to your system.

We don’t expect to see this cooling solution among the server racks you’d find in a data center (Cooler Master has the LTL In Rack CDU for that). Instead, this will probably be located at labs running high-end workstations or on AI and compute development systems where AI engineers are working on next-generation LLMs. Some overclockers might also find this a useful solution, although they might prefer liquid nitrogen cooling over this industrial-grade cooler. Aside from the Aquagate MAX and the LTL In Rack CDU, Cooler Master also showed off a Mini-Charge Tool designed to quickly refill cooling systems with coolant.

For gamers and consumers, Cooler Master released the GPU Shield Adapter Cable, which detects abnormal currents and adjusts power output in real time for improved stability, the MWE Gold V4 power supply with built-in GPU Shield capability, and a variety of fans, air coolers, and AIOs. There are also several PC cases of varying sizes, including the Q300L V3, Qube 540 SP, MasterFrame 500 Mesh V2, and the limited-edition Cosmos Alpha Gold, which only has 199 available units and features a 24K pure gold-plated CM logo.

The Consumer Electronics Show (CES) is well underway in Las Vegas, Nevada, and you can see our previous coverage of Day 0 and Day 1. We're now on to Day 2, with plenty of news still pouring out of the show.

Monitors are a big draw at CES

Whether you're looking for a monitor for gaming or productivity, n several interesting entries have been s announced this week at CES. On the gaming front, one of the most impressive monitors we've seen is the MSI MPG 341CQR QD-OLED X36. It's a 34-inch 1440p widescreen monitor with a 5th-generation OLED panel and an RGB stripe subpixel layout.

The switch from a triangular arrangement to a stripe allows for clearer text with less visible fringing around the edges. This text improvement doesn't come at the expense of light output, color accuracy, or black levels. The MPG 341CQR QD-OLED X36 also boasts a 360 Hz refresh rate and includes a plethora of video connectors, as well as a two-port USB-A hub.

On the productivity side, the UltraSharp 52 Thunderbolt Hub Monitor measures 52 inches across, with a 6K resolution (6,144 x 2,560) and a maximum 120 Hz refresh rate. The monitor can support up to four PCs simultaneously and includes a Thunderbolt 4 port, four USB-A ports, three USB-C ports, two HDMI 2.1 ports, and two DisplayPort 1.4 ports.

Dell has priced the UltraSharp 52 Thunderbolt Hub Monitor at $2,899. For comparison, Apple's 32-inch 6K (6,016 x 3,384) monitor costs $4,999 without a stand.

• Dell debuts monstrously large 52-inch 6K ultrawide monitor
• MSI's fifth-generation QD-OLED gaming monitor embraces RGB stripe subpixels

Intel's Arc B390 looks impressive in early benchmarks, but AMD isn't worried at all

Intel's Core Ultra Series 3 (Panther Lake) processors promise huge performance gains while keeping power efficiency in check. But Intel hasn't just focused on the CPU; its latest Xe3 graphics also deliver healthy performance gains over Xe2.

We had a chance to sample an early system with a Core Ultra X9 388H processor, which includes the 12-core Arc B390 IGP. We saw nearly 100 fps in F1 (High, XeSS Balanced), over 80 fps in Cyberpunk 2077 (High, XeSS Balanced), and roughly 57.4 fps in Spider-Man 2 (High, XeSS Quality).

That said, when we spoke with AMD representatives at CES about Xe3/Arc B390 performance, they seemed unbothered. Rahul Tikoo, AMD's SVP and GM of Client Product Group, stated, "Strix Halo, or Ryzen AI Max will kill it. [...] It's not even a fair fight at that point, because it's discrete-level graphics." He went on to throw in more jabs at Intel, adding, "There's a reason why they didn't compare it right there. [...] they compared their highest-end to our midpoint. And, oh, by the way, that 12 Xe [Panther Lake]... Wait until you see the price point on that. It's gonna be, you know. Enough said."

Game on!

• Intel Panther Lake pre-release testing delivers over 80 fps in Cyberpunk, 100 fps in F1 — Arc B390 offers playable 1080p frame rates with XeSS quirks in our hands-on tests
• AMD is unfazed by Panther Lake's big integrated GPU — 'It's not even a fair fight' to compare the Arc B390 to Strix Halo, AMD exec claims

Everything else from the show floor during Day 2

Here's the rest of the hot tech that we got a chance to take a look at on Day 2.

• Asus unveils cable-free AIO liquid coolers to streamline PC builds for a cleaner look
• Phanteks debuts striking PC case with multi-chamber design
• MSI introduces GPU Safeguard tech on its latest PSUs to fend off RTX 5090 meltdowns — warns users with pop-up and buzzer when abnormal current is detected on the 12V-2x6 connector
• Frore Systems' LiquidJet and AirJet Pak cooling systems in the flesh
• Be Quiet showcases its minty fresh Light Loop AIO cooler and Dark Perk mice at CES 2026
• Patriot shows off opulent Viper Xtreme 5 Aurum RAM kit with flashy finish as prices surge throughout industry

Patriot just dropped a plethora of new memory modules and SSDs, as well as several storage solutions for mobile devices and gaming handhelds. Despite the ongoing DRAM and storage pricing squeeze, the Tom’s Hardware team checked out Patriot's booth at CES 2026 in Las Vegas, Nevada, to check out 15 new DDR5 RAM kits, as well as a DDR5 SODIMM with a maximum capacity of 48GB (1x48GB).

Out of all the displayed modules, the company's Viper Xtreme 5 Aurum stood out with its delightfully opulent, flashy exterior. The company said that it designed the color specifically to blend well with both light and dark builds, ensuring that it wouldn’t stand out as either too bright or too dark. But, you might need to wear a pair of gloves to not litter it with pesky fingerprints. Patriot also showed off a prototype with the same eye-catching heat-spreader, running in a 2x24GB configuration, at speeds of up to 10,000 MT/s.

There are also four Viper-branded SSDs — the Viper VP4300 Lite PCIe 4.0 x4 SSD available in 500GB, 1TB, 2TB, and 4TB capacities, and the Viper PV563 PCIe 5.0 x4 SSD, which you can get in either 1TB, 2TB, or 4TB capacities. Additionally, the PV563 can be had with either a graphene strip or a full heatsink. Lastly, there’s the top-of-the-line Viper PV593, offering up 14,000 and 13,000 MB/s sequential read and write speeds, plus a random read speed of up to 2,000K IOPS, Patriot claims. We have no word on how much the memory or storage might cost you; however, as pricing for both DRAM and NAND storage remain extremely volatile. An IDC report suggests that PC prices may jump as much as 8% as a result.

Aside from these PC components, it also released a wide range of USB flash drives and enclosures for those that need portable memory on the go. The company also introduced a Storage Hub designed for mobile devices. This plugs in to your gear via a USB-C 3.2 Gen 1 connector, giving you one HDMI port capable of 4K@60Hz and a USB-C PD 3.0 port that accepts up to 100 watts of power. More importantly, it will give you extra storage options in 128GB, 256GB, 512GB, and 1TB capacities, allowing you to turn your phone or tablet into a proper productivity tool.

Nintendo Switch 2 owners are also getting another microSD Express option with the Patriot EP Express. This PCIe 3.0 x1 expansion card delivers sequential read and write speeds of up to 800 MB/s and 600 MB/s, allowing you to quickly load games directly from it. Aside from that, it also supports 4K and high-bitrate video recording, allowing you to use it with video recording and photography equipment as well, available in 256GB, 512GB, and 1TB capacities. The company also showed off a DDR5-10000 engineering sample with a 48GB (2x24GB) capacity. While this is still being tested, it gives us a glimpse of what Patriot has in store for us in the future. But you might want to start saving up now if you have your eye on any of the new products.

Memory prices are booming, pricing out many prospective PC builders in 2026. But, according to AMD's SVP & GM of the Client Business Unit, Rahul Tikoo, it's nothing to be worried about. During a roundtable interview with Tom’s Hardware at CES 2026 in Las Vegas, Nevada, we asked Tikoo how the company is responding to skyrocketing DRAM and storage prices, and how that might affect AMD going into 2026, as market intelligence firm IDC forecasts that PC shipments to fall by as much as 9% this year.

“We don’t see an issue there other than, you know, tightness leads to higher prices, eventually. So, from that perspective, I’m not seeing any impact to our business this year,” the executive told Tom’s Hardware. “I think that people who need the technology need the technology, and they’re going to buy the technology — though consumers might decide that they have a choice to make on how much memory, what CPU.” He also added, “Which is okay because we have a wide assortment of products available.”

AMD currently offers 215 SKUs, which are "AI PC" capable. There are also an equal or larger number of non-AI PC options available from the chipmaker. This means prospective PC buyers across a host of budgets and form factors may be able to find an AMD CPU to fit their build and budget. Old AM4 CPUs still in production have been topping Amazon’s sales charts, with another AMD executive even teasing the return of older Zen 3 processors.

Tikoo’s answer is logical — after all, AMD does not influence the memory chip supply chain, and they’re also seemingly not directly affected. Though the company sells Radeon GPUs, which are gaining ground on Nvidia’s RTX graphics cards, its market share is still small, sitting at just 7%. On the other hand, Nvidia dominates the GPU space, with a staggering 92% footprint. Therefore, it's likely that Nvidia is heavily affected, especially since Nvidia is reportedly no longer supplying VRAM to its GPU board partners. So, the only thing that it can really do is to offer a wider range of SKUs that utilize older DDR4 memory, which is currently slightly more affordable than DDR5.

Unfortunately, it offers little comfort to the average consumer — especially those who have already invested in a Zen 4 or Zen 5 chip, as the AM5 platform exclusively uses DDR5 memory. In line with what industry players are saying, you should only buy RAM kits or SSDs right now if you need them. If you don’t have an urgent requirement for more memory or storage, then we recommend you wait out the RAM-pocalypse before eyeing up a new upgrade, or find a solid bundle deal.

Although Tikoo doesn't see a major impact to AMD's business from DRAM shortages, the reality of higher pricing is hard to ignore. Estimates on when the DRAM shortage could ease range anywhere from six months to multiple years. As we begin 2026, we'll keep a close eye on memory pricing and where it's headed.

Intel has finally pulled back the curtain on Panther Lake, and in addition to a major process milestone with 18A, Intel is packing in the largest integrated GPU it has ever shipped in a handful of new X-series SKUs. These SKUs, including the Core Ultra X9 388H that Tom’s Hardware was able to test ahead of release at CES 2026, include 12 cores on Intel’s latest Xe3 graphics architecture and boast performance that can rival a discrete RTX 4050, at least according to Intel’s claims.

Testing the graphics performance in-person, Panther Lake is undoubtedly impressive, not only because of its raw capabilities, but also its clear efficiency. We aren’t able to publish power numbers today, but subjectively, Panther Lake delivered easily the coolest gaming experience I’ve seen out of a laptop, barely even getting warm during an hour and a half of heavy-duty benchmarks in a Lenovo reference design.

The silicon quality and process advancements are obvious, and Intel’s claims aren’t full of hot air. Panther Lake reaches that entry-level segment that Intel is targeting with its X-series parts, and does so with remarkable efficiency. The asterisk is software. Even in the benchmarking session, software issues crept up with a handful of popular titles Intel had pre-installed on the machines, which will only become more pronounced when extended to a larger sample of titles.

We’ll have full test results for Panther Lake once systems arrive on January 27 globally, including firm power and temperature numbers, along with productivity results. For now, we’re looking at some quick and dirty tests in games, along with the experience of using Panther Lake in the flesh.

Meet our Panther Lake test system

Intel set us up with a reference system made by Lenovo. You can see the system information above. This is an IdeaPad Pro 5 running the top-end Core Ultra X9 388H chip from Panther Lake, along with 32GB of memory. For testing, the system was perched on a stand for better airflow, and the games were running off of an external PCIe Gen 4 NVMe SSD connected via USB-C.

The native resolution for the display is 2880 x 1800, but I ran tests at either 1920 x 1080 or 1920 x 1200 depending on support in-game, along with XeSS upscaling (settings for each will be noted in the benchmarks below). This is in-line with the benchmark data Intel itself has shared on Panther Lake. Keep in mind, however, that upscaling works better with higher output resolutions. You’ll see a performance drop climbing up to native resolution, but with upscaling, the drop isn’t as severe as you’d expect.

Although we can’t publish power results, Intel confirmed that the Core Ultra X9 388H has a maximum turbo power of 65W (PL1) with a short-term limit of 85W (PL2). That’s power for the entire SoC, not just the integrated graphics. The performance results here for the SKUs sporting 12 Xe3 cores; the vast majority of the Panther Lake lineup only includes 4 Xe3 cores, with only the X-series SKUs getting the full 12.

As we dig into the benchmarks, keep in mind the limited time we had available for testing these systems. Most games only got a single benchmark pass as we tried to test as many titles as possible within an hour and a half. You should take the numbers as rough approximations of the performance you can expect ahead of launch, not the final word on the gaming prowess of Panther Lake.

Testing Panther Lake’s gaming performance

Now what you came for — the benchmarks. Broadly, you can expect frame rates nearing 60 fps in recent titles with Panther Lake at 1080p, with some assistance from XeSS upscaling. Multi-frame generation is available in titles that support XeSS 2 or newer through an override in Intel’s driver software. We didn’t look at frame generation when gathering performance results. However, in Spider-Man 2, we quickly looked at FSR frame generation and saw positive results. Given the limited availability of XeSS frame generation, FSR 3 (or newer) seems like the more practical route for frame generation on Panther Lake outside of tools like Lossless Scaling.

We tested two more titles, Clair Obscur: Expedition 33 and Battlefield 6, but we weren’t able to capture performance data; more on that in the next section.

Overall, the average frame rates are solid considering the settings. With upscaling assistance, 60 fps is achievable even in relatively demanding titles like Baldur’s Gate 3, while heavily titles like Marvel’s Spider-Man 2 and Doom: The Dark Ages fall slightly short of that mark while still providing a smooth experience for an efficient mobile platform. Actually playing a game like Doom: The Dark Ages brought out the experience Intel has been talking about. When you’re playing, the fluctuations below 60 fps aren’t apparent, and the complete lack of heat leads to a mobile gaming experience that’s refreshingly clean. Fan noise wasn't bad, either. In a room with about two dozen people with only a minor murmur, you couldn't hear the fans spinning.

Although the performance holds up, the software backing it has some issues. XeSS works and is available in a broad range of titles, but the quality of the implementation varies wildly. In F1 2025, it can go toe-to-toe with FSR 2 or 3 upscaling; in Cyberpunk 2077, the visual artifacts are immediately apparent. Intel still has a lot of work to do on XeSS, particularly when it comes to unifying the quality you can expect across titles.

We tested Spider-Man 2 with FSR frame generation, XeSS Quality, and 1080p resolution, which you can see the performance of in real-time in the video above. This provided a much better experience and clearly showed off the capabilities of Panther Lake without the drawbacks of Intel’s software ecosystem. Some of those drawbacks are too big to ignore. However, in many titles, moving over to FSR or Lossless Scaling can bring Panther Lake to life in a way that the limited support for XeSS 3 can’t.

The average frame rates are solid, but consistency is a bit different. There were titles where the consistency was great, such as Cyberpunk 2077, while others, like Baldur’s Gate 3, struggle. And there are still games that will prove too demanding for Panther Lake without heavy assistance from both upscaling and frame generation as evidenced by Monster Hunter Wilds.

Questions linger about Intel’s software stack

Intel has done a lot of work stabilising its Arc driver releases, and today, they’re much more robust than what we saw just a few years ago with the initial release of Alchemist GPUs. Still, I have some concerns about the software when it comes to using Panther Lake in a wider swath of games.

In addition to the titles above, we tried to test Clair Obscur: Expedition 33 and Battlefield 6, both of which we weren’t able to gather data for because Intel’s PresentMon — the monitoring tool available to us for testing — wasn’t able to hook into the game. Clair Obscur: Expedition 33 also reduced the resolution to just 900p each time we toggled to fullscreen mode; a well-known bug in this game that shows up across GPU vendors.

In areas Intel can more directly influence, there were still issues. Monster Hunter Wilds had an unplayable level of stuttering on the first benchmark pass, while the quality of XeSS in Clair Obscur devolved into a shimmering mess while falling short of a 60 fps average. These, thankfully, aren’t issues in the driver, but it takes the claims about Intel’s all-in-one graphics ecosystem down a peg when looking at the real-world experience.

Although we didn’t focus on frame generation, others in the room did; just behind our station, another journalist spent the better part of 45 minutes troubleshooting Intel’s multi-frame generation overrides in the driver without success. The performance of Panther Lake is solid, and I’m confident in it when combined with tools like FSR and Lossless Scaling. You can expect some bumps in the road if you stay fully within Intel’s ecosystem, however.

That shouldn’t overshadow the impressive performance of Panther Lake. The performance is stunning considering the hardware, and even moreso when factoring in power and temperatures. We’re looking forward to sharing those results with you when we can, but broadly, Intel’s claims about Panther Lake hold up. It’s a relentlessly powerful and efficient SoC that hits a sweet spot few other chips can touch. Hopefully our initial testing results hold up once we can dedicated dozens of hours to putting Panther Lake through the full gauntlet.

Frore Systems' seemingly magic apparatuses for both air and liquid cooling have long been in the news, claiming to offer performance up to several times that of existing designs. Some companies are already using Frore's designs, like Qualcomm with its Snapdragon X2 Elite reference platform, and many others are integrating the products as we speak. At CES, Tom's Hardware photographed some rather impressive live demos.

The company's wares are the AirJet and LiquidJet lines, each for a different cooling medium as their names imply. The AirJet Mini G2 is the building block, so to speak, of a cooling package that Frore calls an AirJet Pak and includes several Minis with vibrating membranes for near-silent or even fully passive operation. The company is aiming these at ultraportables, laptops, and mini-PCs — roughly any application requiring up to around 45 W. The units are dust-proof, too, and water-resistant.

One of the most impressive demos is this one, an AirJet Pak 5C (five AirJet Mini G2 units) cooling a Nvidia Jetson Orin NX Super machine (40 W steady TDP), with a 300g apparatus containing a cooling package not much larger than a 2.5" hard drive performing the same duty as a massive heatsink weighing in at a whole two kilos.

The same theme carries over to this Galaxy Book 5 Pro 14" that replaces its two-fan, 32-37 dbA bog-standard laptop cooling setup with a design packing four AirJet Mini G2s, offering an even higher TDP (24 W over 20 W), while simultaneously operating silently at the base noise measurement level of 27 dbA. As additional bonuses, theoretically, the AirJet-clad unit won't ever need cleaning, either. Not only is it much quieter, but the AirJet variant also goes significantly faster thanks to the higher steady-state TDP, as shown by the running Cinebench tests.

For servers and other high-performance applications, Frore makes the LiquidJet, essentially a really high-tech cold plate with 3D microscopic water channels that move water through a much shorter path than conventional designs, with far better cooling efficiency and lower pressure. Unsurprisingly, Frore is targeting mainly, but not only, at AI servers, where more efficient cooling directly results in lower power bills and better ROI.

Over at the Big Iron desk, Frore has a big honkin' Nvidia Rubin 2 SoC containing eight HBM modules plus I/O chips (in the center of the picture), for a grand total of 1950 W of heat that must be dissipated. The LiquidJet setup keeps this chip at a balmy 65-70°C, and the radiator-and-fan kit appears to be roughly bigger than an SFF PC, quite an achievement compared to what would otherwise be necessary. We'll bet that the fans weren't nearly as loud as those in a datacenter, too.

Here we see a large ASIC with a 1200 W TDP being cooled by one of its LiquidJet cold plates, remaining pretty chilly at 70-75°C. The inlet water temperature is just 34.5°C, warm enough to take a shower but not even qualifying as "hot".

Finally, a 600 W GPU (105° C), cooled by a LiquidJet with top-mounted inlet and outlets, holding itself steady at a nice 95°C. This displays how Frore precisely designs each block to better suit the hotspots of the chip(s) resting underneath, one of the key features for LiquidJet blocks. Be sure to click the pictures for high-res versions, and check out the gallery below for the whole set.

AMD's CES 2026 keynote was conspicuously missing the highly-anticipated Ryzen 9 9950X3D2. The CPU originally leaked alongside the Ryzen 7 9850X3D — which did get announced this week — while the true flagship was a no-show. At a company QnA in Vegas on Wednesday for CES, AMD hinted to media, including Tom's Hardware, that the 9950X3D2 is indeed on the cards, but that it doesn't have anything to announce right now.

At the media event, we asked AMD if the 9950X3D was "potentially one of the aces up [its] sleeve," to which Rahul Tikoo replied: " 9950X3D2? Yes. I wish I could announce future products today, but I can't, so no comments on that yet."

The 9950X3D2 is rumored to be the company's first processor to stack 3D V-Cache across both CCDs. This would allow for a whopping 192 MB of total L3 cache on the 9950X3D2, way more than even the 128 MB on its namesake brethren, the 9950X3D, which only has the extra L3 cache on one of its two chiplets.

In real-world usage, an increased cache pool could lead to marginal gains in gaming performance. The rest of the chip remains identical to the standard 9950X3D, with current rumors pointing toward a 200W TDP, a slight bump from 170W. It will also carry a max boost clock of up to 5.6 GHz with the same core layout.

AMD's non-commital answer notwithstanding, Alienware China shared a video on its BiliBili account, mentioning the 9950X3D2 coming to its Area 51 desktop. Following the publication of this article, Alienware clarified to Tom's Hardware that the Area 51 will ship with the announced 9850X3D, rather than the 9950X3D2.

Alienware didn't share any more details beyond that, and that's where a British system integrator by the name of "Sytronix" comes in.

The Ryzen 9 9950X3D2 shows up on the company's website, under the workstation page, as part of the "NexStation" system. There, the 16-core/32-thread config is listed along with an "X3D2 architecture" for the chip, claiming it offers improved cache management while running cooler than conventional CPUs. The 9950X3D2 is being paired with AMD's Radeon AI PRO 9700 GPU in the NexStation.

Those are two separate reports that tell us this elusive CPU not only exists, but that it's out there in the hands of some SIs. Why AMD chose to keep it close to its heart instead of revealing it, well, we don't know.

Who says you can’t have your cake and eat it too? That appears to be MSI’s mantra with the new MPG 341CQR QD-OLED X36 gaming monitor, which was on full display this week at CES. As its name implies, the MPG 341CQR QD-OLED X36 (say that three times fast) uses a 5th-generation QD-OLED panel with an RGB stripe sub-pixel arrangement.

The move to a vertical RGB stripe instead of the slightly odd triangular sub-pixel layout on previous Samsung QD-OLED panels allows for sharper text on screen. The increase in text clarity doesn’t come at the expense of the other advantages inherent to OLED panels (namely, rich color, ultra-fast response times, and deeper blacks).

In the case of the MPG 341CQR QD-OLED X36, we’re looking at full-screen brightness of 300 nits (1,300 nits peak), which is an increase from the typical 250 nits of preceding QD-OLED panels. MSI says that it incorporates an anti-reflective DarkArmor Film to boost the black levels by another 40 percent (VESA DisplayHDR True Black 500 certified), while delivering a 2.5x improvement in scratch resistance (a curious bullet point for a desktop monitor, but it’s appreciated).

MSI has also swapped out the rather large Y-shaped stand for a more compact flat base. This should help gamers reclaim previous desk space and make for a slimmer profile (the footprint has been reduced by 62 percent).

The MPG 341CQR QD-OLED X36 measures 34 inches across, features a 1800R curve, and has a resolution of 3440 x 1440. You’ll also find that the refresh rate is a speedy 360 Hz, and the response time is an OLED-typical 0.03 ms.

When it comes to connectivity, the MPG 341CQR QD-OLED X36 has two HDMI 2.1 ports, one DisplayPort 2.1a port, and a USB Type-C port (DisplayPort Alt Mode) that also supports 98-watt Power Delivery. You’ll also find a two-port USB-A (5 Gbps) hub and a 3.5 mm headphone jack.

The monitor also incorporates the MSI OLED Care 3.0 AI Care Sensor, which provides comprehensive protection for the display panel. This NPU-based IC can detect whether a person is sitting in front of the monitor; if no one is present, the monitor will initiate automatic power-optimization and/or automatic pixel-refresh routines to extend the panel’s usable service life.

At this time, MSI has not announced pricing or an on-sale date for the MPG 341CQR QD-OLED X36.

Asus’ latest ROG Strix LC IV series of AIO liquid coolers not only rival the best CPU coolers on the market, but also introduce the world’s first cable-less design. This innovation is made possible by a proprietary AIO Q-Connector that relies on pogo pins for seamless connectivity, and we got a good look on the ground at CES 2026 in Las Vegas.

The novel AIO Q-Connector is at the heart of Asus' breakthrough, featuring 11 spring-loaded pogo pins organized in two groups—nine for primary connections and two for additional controls. These precision connectors allow the new ROG Strix LC IV coolers to seamlessly receive power and control signals for the pump, ARGB lighting, and, in select models, an integrated 5.08-inch LDC display. Impressively, there’s no need for users to connect any cables to the radiator fans manually; the AIO Q-Connector discreetly transmits PWM and lighting signals through a cable cleverly concealed inside the cooler’s tube sleeving.

To fully benefit from the innovative cable-free design of the ROG Strix LC IV series, you must pair these coolers with a compatible motherboard equipped with the AIO Q-Connector. Currently, Asus has introduced four new AM5 motherboards with the new connector: the ROG Crosshair X870E Glacial, ROG Crosshair X870E Dark Hero, ROG Strix X870E-E Gaming WiFi7 Neo, and ROG Strix X870E-A Gaming WiFi7 Neo. Asus may expand the AIO Q-Connector to Intel motherboards in the future, but for now, AM5 users have a unique advantage in creating a next-generation, wireless cooling setup.

While the ROG Strix LC IV series is designed for a truly cable-free experience, Asus has ensured flexibility by including a modular design that allows users to connect standard cables to any motherboard, even if it lacks the proprietary AIO Q-Connector. However, given that the key selling point of the ROG Strix LC IV AIO liquid coolers is its advanced wireless functionality, it would be counterintuitive to invest in this premium product without leveraging its signature feature.

Asus has introduced three distinct base models for the ROG Strix LC IV series, each available in two color variants—providing a total of six configurations to suit various build aesthetics. The ROG Strix SLC IV 360 ARGB LCD and ROG Strix LC IV 360 ARGB LCD stand out with their vibrant 5.08-inch LCD displays, boasting a crisp 720 x 720 resolution. These displays enable users to monitor critical hardware stats, showcase personalized content, or display exclusive ROG animations, enhancing both functionality and visual appeal. Notably, the LCD panel connects to the pump via a dedicated five-pin pogo connector, ensuring reliable data transmission and a streamlined installation process.

The key distinction between the ROG Strix SLC IV 360 ARGB LCD and the ROG Strix LC IV 360 ARGB LCD lies in the tubing length. The SLC model features shorter tubing—an intentional design choice aimed at PC builders seeking a tidier and more organized interior. The small adjustment helps prevent tubing from hanging awkwardly across other components.

Completing the lineup is the ROG Strix LC IV 360 ARGB, which forgoes the LCD display and instead features an ARGB-illuminated logo on the pump housing. This model offers a more budget-friendly option for builders who want to experience the innovative cable-free design of the ROG Strix LC IV series without the additional cost of a display.

Another noteworthy feature of the ROG Strix LC IV lineup is the all-in-one fan assembly. The three cooling fans come fully pre-installed on the 360mm radiator out of the box. The approach, as silly as it may sound, eliminates the need for manual fan mounting, saving PC builders time. The upside of the design is that it allowed Asus to implement its Aura Sync lighting across the triple-fan setup, making it look like an animated billboard.

Asus has yet to announce official pricing or a release date for the ROG Strix LC IV series, keeping eager PC builders in suspense for now. However, the company has confirmed that all models in the lineup will come with a generous six-year limited warranty.

The Consumer Electronics Show (CES) has officially kicked off in Las Vegas, and we've already brought you extensive Day 0 coverage. We're back with our Day 1 coverage, complete with all the latest news on Lenovo's concept laptops with rollable OLED displays, a plethora of Asus Zenbook and ROG Zephyrus laptops, and the latest advancements in PCIe 5.0 SSDs aimed at more value-conscious offerings.

Lenovo continues its tradition of unwrapping wild concept PCs at CES

Lenovo is not afraid to show off concept designs for PCs that may or may not eventually reach the public market. CES 2026 is no exception, as we see with the Legion Pro Rollable. It features a seemingly mundane 16-inch OLED display.

However, when you open the monitor, both the right and left ends of the display unfurl to give you a larger 21.5-inch display in "Tactical Mode" or an even larger 24-inch display in "Arena Mode." The system is powered by Intel Core Ultra processors and RTX 50-Series graphics.

The second concept is the ThinkPad Rollable XD, which has a display that rolls out vertically, expanding from 13.3 inches to 16 inches.

• Lenovo's CES concept devices include a rollable ultra-wide OLED Legion gaming laptop
• Lenovo brings handheld-favorite SteamOS to the flagship Legion Go 2
• Lenovo's top ThinkPads and Yogas go all-in on Panther Lake

Asus blankets CES with new ROG and ZenBook laptops, keyboards, and gaming headsets

Asus debuted two new ROG Zephyrus laptops, which are available in 14- and 16-inch display sizes. The Zephyrus G16 is available exclusively with Core Ultra Series 3 chips, while the smaller Zephyrus G14 allows you to choose between Core Ultra or Ryzen AI processors.

Another item of note is the option of a 120 Hz 3K display on the G14 or a 240 Hz 2.5K display on the G16.

The TUF Gaming A14 managed to cram AMD's Strix Halo (Ryzen AI Max+ 392) into a relatively thin chassis. The display measures 14 inches across and has a 1600p resolution and a 165 Hz refresh rate.

If you're more interested in a thin-and-light productivity laptop, the Zenbook series has been refreshed. The S-Series comes with Panther Lake or AMD Ryzen AI 400 processors, up to 32GB of LPDDR5x memory, and up to a 3K OLED display with a 120 Hz refresh rate.

The A-Series is available in 14- or 16-inch display sizes, topping out at 1200p and 1800p, respectively. The smaller laptop is available with a Qualcomm Snapdragon X2 Elite, while the larger one can be equipped with the X2 Elite Extreme.

The company also showed off the ROG Cronox, a panoramic PC case that features a 9.2-inch display and fans that can be rotated to optimize airflow. Another interesting addition to the show was the ROG Kithara, an open-back planar magnetic gaming headset developed in conjunction with HiFiMan.

• Asus launches two new ROG Zephyrus laptops at CES
• Asus puts Strix Halo in the TUF Gaming A14 — strong integrated graphics in a very thin chassis
• Asus' Zenbook lineup gets 2026 refresh
• Asus unveils the ROG Cronox, its first panoramic PC case with a 9.2" screen and rotating fan bracket
• Asus teams up with HiFiMan for planar magnetic ROG Kithara gaming headset

PCIe 5.0 SSDs with QLC NAND shipping to OEMs, Phison debuts ultra-efficient PCIe 5.0 SSD controller

PCIe 5.0 SSDs have largely been relegated to the high-performance/enthusiast markets, but that's starting to change for the better. Micron announced that its 3610 SSD is shipping to partners using QLC memory. While it doesn't offer the breakneck speeds of its TLC-based products, the 11,000 Mbps performance is nothing to scoff at and well above PCIe 4.0 SSDs. Micron also plans to provide SSDs in 2230 form factor with capacities up to 4TB.

Phison is a dominant player in the PCIe 5.0 SSD controller market, and it's showing its prowess in the segment with the PS5037-E37T. While its previous flagship products have prioritized outright performance over efficiency, the PS5037-E37T is aimed at more value-minded customers without making any significant performance sacrifices. It can achieve sequential reads and writes of 14.7 Gbps and 13 Gbps, respectively, with active power consumption of under 2.3 watts.

• Micron announces 3610 SSD, the industry-first PCIe 5.0 QLC SSD available to OEMs
• Phison's new SSD controller sips a mere 2.3W and runs at 14.7 GB/s, addressing PCIe 5.0's power concerns

Everything else from the show floor during Day 1

Here's the rest of the hot tech that was either announced or that we got hands-on time with on Day 1.

• Broadcom announces two dual-band Wi-Fi 8 chips
• Corsair's Galleon 100 SD is an LED-lit keyboard with a built-in Stream Deck
• AMD Ryzen chief teases return of older Zen 3 chips to fight soaring RAM prices
• AMD hints at officially open-sourcing FSR 4
• MSI introduces GPU Safeguard tech on its latest PSUs to fend off RTX 5090 meltdowns — warns users with pop-up and buzzer when abnormal current is detected on the 12V-2x6 connector
• Phanteks debuts striking PC case with multi-chamber design
• MSI's new MPG CoreFrozr CPU air coolers feature a 6-inch LCD display

MSI has shown off some rather intriguing hardware so far at CES 2026, but something that really caught our eye while touring the show floor is the company's latest dual-tower CPU air coolers: the MPG CoreFrozr AP15 and AP17. Your eyes can't help but be drawn to the displays affixed to each unit.

The MPG CoreFrozr AP15 features a Digi-Display that provides real-time readouts of your system vitals. The LCD is split into four segments, showing CPU temperature, CPU load, and the fan's current RPM. There's even a weekly calendar function in the lower left-hand segment.

The MPG CoreFrozr AP17 takes things a step further with an absolutely massive 6-inch LCD. Not only can it display the same system information as its lesser sibling, but it also offers fully customizable output. You can even add custom visuals to give your system some added flair that matches the aesthetic of your gaming rig. The displays are also removable, though I don't really see the need to remove them if you're paying a premium for that functionality in the first place, but I digress.

When it comes to actual cooling performance, both feature dual cooling fans. However, their approaches differ slightly from there. The MPG CoreFrozr AP15 was specifically optimized to work with the AMD Ryzen X3D family of processors and features a copper base along with six heat pipes. The MPG CoreFrozr AP17 ups the stakes with a nickel-plated copper base and eight heat pipes.

MSI also designed both coolers to be user-friendly, especially in tight confines. As a result, there is sufficient clearance under the heat pipes to avoid interfering with your motherboard's DIMM slots.

At this time, MSI hasn't announced pricing for the MPG CoreFrozr AP15 or AP17. However, for reference, the Thermalright Peerless Assassin 140 is available for around $53.

PC components manufacturer Phanteks showcased one of the most striking PC cases at CES this year. At first glance, the new Exo 626s looks like an open-frame case; however, it comes with dedicated chambers for the PSU, motherboard, graphics card, and a 360mm AIO liquid cooler for the CPU.

The idea behind this case is that with each component having a separate chamber, airflow is easier to manage, and heat from one part doesn’t affect the others. From what we gathered on our visit in Vegas, the GPU chamber can fit large-sized graphics cards going as big as 4-slots. It comes with an RGB LED strip and a mesh finish at the top and the bottom, allowing the card to breathe easily. The PSU chamber at the bottom features a similar RGB strip along with some I/O ports at the front, including two USB Type-A ports, a USB Type-C, and a 3.5mm audio jack.

The AIO chamber also features a mesh finish at the front and is positioned next to the motherboard with a provision to route the water cooling tubes directly into the motherboard chamber. Lastly, the motherboard chamber comes with a glass panel at the front, while the back panel gives access to the AIO radiator and the rear side of the motherboard with a provision to mount a 120mm fan for additional cooling.

Unfortunately, the case is available as part of a pre-built only, as it is made in partnership with CyberPowerPC, with no confirmation on a standalone retail version at the moment. The pre-built will be offered with an Intel Core Ultra 285K CPU, an RTX 5080 GPU, an Asus TUF Z890 motherboard, 32GB of DDR5 dual-channel memory clocked at 6400 MHz, a 2TB WD Black SN8100 PCIe Gen 5 SSD, a Bitspower 360mm AIO liquid cooler with ARGB lighting, and a Corsair RM1000X 1000W 80+ Gold power supply unit. As for the pricing, it will be retailing at $3,500.

While the Exo 626s looks impressive, locking such an innovative case behind an expensive pre-built feels like a missed opportunity. This is exactly the kind of uncanny airflow-focused design PC enthusiasts would love to experiment with in their own custom builds. Hopefully, Phanteks takes note of the interest and considers releasing the Exo 626s as a standalone case for consumers and not just system integrators.

MSI just announced its new GPU Safeguard and GPU Safeguard+ tech available on its latest power supply units (PSUs). According to the company press release, this feature monitors the 12V-2x6 connector for abnormal current fluctuations, and it will then warn the user with a pop-up through the MSI Center (only for GPU Safeguard+) while the PSU itself will emit an audible warning buzzer. “This proactive alert system enables potential power issues to be addressed before they result in permanent hardware damage,” MSI said in its PR.

The Tom’s Hardware team visited the MSI booth at CES 2026, where it was explained that this protection is achieved through per-in-voltage monitoring. And if it indicates an anomaly, the only way you can stop the buzzing is by shutting your computer down, thereby cutting power to the GPU and helping protect it even before its power connector starts melting.

If it does not detect any response after three minutes of beeping, the PSU will force a black screen on your system to reduce the load on the graphics card while still continually beeping. MSI says that “In either case, the only way to get your PC up and running safely again is by shutting it down, unplugging the 12V-2x6 connector from your graphics card, inspecting for damage, and firmly re-plugging it if all’s well.”

This tech comes in four of MSI’s latest power supplies — the top-of-the-line MPG Ai1600TS and MPG Ai1300TS PCIE5 get GPU Safeguard+, while its mainstream MAG A1200PLS and MAG A1000PLS PCIE5 PSUs are protected with GPU Safeguard. The only difference between the two features is that the former displays a pop-up window alongside the warning sound, while the latter only gives an aural warning. Nevertheless, this should help protect your system from unexpected meltdowns, especially for high-powered GPUs like the RTX 5090 and even some Sapphire Nitro+ 9070 XTs.

Aside from these new PSUs, we also saw the MEG Maestro 900R panoramic PC case, with its triple-sided curved glass letting you see your build from every angle. There’s also the Coreliquid E16 360 AIO with its massive 6.67-inch 2K curved OLED display right on the CPU block. There’s also a plethora of other new components, like air coolers, motherboards, PC cases, displays, and GPUs, including the legendary MSI RTX 5090 Lightning.

Lenovo is updating its expansive productivity laptop lineup at CES 2026, including both its famed ThinkPad business notebooks and its consumer and creative-oriented Yogas. While Lenovo is aligning itself with all three big processor companies, Intel, AMD, and Qualcomm, its top-end systems all have one thing in common: Intel's new Core Ultra Series 3 "Panther Lake" processors.

Those systems are what Lenovo calls "Aura Editions," which the company uses to showcase AI-focused tools, like "Smart Modes" that can detect what you're doing and change settings, or "Smart Share," which lets you share photos and videos with selected devices.

ThinkPads and Business Laptops

The flagship ThinkPads are the X1 series. This year, the X1 Carbon and X1 2-in-1 are debuting new "Space Frame" chassis, which rearrange the interiors of systems so that components can be placed on either side of the motherboard. The company says this space efficiency also allows for higher sustained performance due to better cooling, as well as easier repairs with replaceable parts including the battery, USB ports, keyboard, fans, and speakers.

Lenovo says that these systems will have up to 20% increased heat dissipation due to updated thermal architecture. Both systems also come with a new 10MP camera with a 110-degree field of view.

Meanwhile, the ThinkPad X9 15p Aura Edition — a take on the laptop we called the prettiest business PC — is also jumping up to Panther Lake, including Intel Core X9 processors. The system features what Lenovo says is the largest ThinkPad haptic touchpad (135 x 85 mm), and will also get a 10MP camera.

These Aura PCs aren't cheap. They will all go on sale in March, with the X1 Carbon starting at $1,999, the X1 2-in-1 beginning at $2,149, and the ThinkPad X9 15p launching at $1,999. Interestingly, all of these laptops will come with both Windows 11 and Linux as configuration options.

In addition, Lenovo is bringing its motorized hinge-toting "Auto Twist" concept, which debuted as a concept at IFA in 2024, to a real laptop in its small business line, the ThinkBook Plus Gen 7 Auto Twist, which will go on sale starting at $1,649.

Yogas and Consumer Laptops

On Lenovo's consumer side, there are two dramatically different categories of device. The Yoga Pro 7i and 9i Aura Edition are both focused on creatives, with up to Intel's Core Ultra 9 386H processor and Nvidia GeForce RTX 5070. The Yoga Pro 9i has an option for a Tandem OLED display, a six-speaker sound system, and uses a power adapter going up to 245W (the 7i provides up to 140W over a standard USB Type-C connection). These will launch in Q2 starting at $1,799.99 for the 7i and $1,899.99 for the 9i.

The other extreme is the Yoga Slim 7i Ultra Aura Edition, a 14-inch white system coming in at less than 2.2 pounds. You'll still get up to an Intel Core ultra X9 with up to 12 Xe graphics cores and 9600 MHz RAM, as well as a new matte texture that Lenovo says will reduce fingerprints. This system will start at $1,499.99 in Q2.

Lenovo also has Yogas with AMD and Qualcomm processors. The latest AMD Ryzen AI 400 chips will power the Yoga Slim 7a, while the Snapdragon X2 Elite and X2 Plus will come to the Yoga Slim 7x. These are both 14-inch devices with 2880 x 1800 touch screens, designed for mobility and power efficiency. The Slim 7x will start at $949.99 in Q2, while the 7a is only being sold in select markets.

Lenovo continues to experiment with rollable-screen OLED laptops here at CES 2026, and its latest Legion Pro concept laptop is sure to pique the interest of on-the-go gamers. Its 16-inch display fits nicely in the Legion Pro 7i chassis it's built around. But open the laptop up, and its screen expands sideways, in both directions, to become either a 21.5-inch "Tactical Mode" screen, or an even more immersive ultra-wide 24-inch "Arena Mode" display. The current firmware has you hit Fn and the arrow left/right keys to expand or retract the screen. Hitting the keys twice skips the middle mode and completely extends or retracts the screen.

The ultra-wide form factor has been popular among desktop gamers for years, but has been extremely rare in gaming laptops due to size constraints. But Lenovo is trying to make that an issue of the past with the Legion Pro Rollable — or at least it will if the company takes this concept further and creates a retail version.

Of course, you'll need some serious CPU and GPU power to push those extra pixels. But the Legion Pro 7i that this concept is built around tops out with a mobile RTX 5090 and Intel's Core Ultra CPUs. While I of course don't know how much a finished version of this rollable concept will cost, I am sure there are quite a few gamers who would love to be able to carry around a 16-inch laptop that can deliver a screen up to 24 inches diagonally.

Lenovo's press release positions the Legion Pro Rollable as a device for esports competitors who can "carry small but train big" as they travel to and from events. And sure, I can see that, but I'm pretty sure there's a larger market as well for more casual gamers and road warriors who would love a screen-morphing portable with a screen that can get very large when you need it, while rolling back to a 16-inch size so it can fit in your backpack.

In a hands-on with the concept device, the screen looked very much like a product in development. The bezels were chunky and some of the screen edges didn't line up right. It works, but I didn't get to try a game on it. The main thing Lenovo seems to need to do is to work on the tension mechanism to hold the screen taught. They have this in the ThinkPad rollable concept (see below) with two motors and steel cables, so they know how to do it. But in the Legion, the screen looks a little wavy.

The sample I tried also made a pretty loud creaking noise, though Lenovo representatives told me it was a quirk of the unit I tried. I saw others that didn't have the same issue.

The other thing that I couldn't really tell was how much weight this might add. Probably not a huge amount, but the base laptop they used here was already a powerful and heavy system.

ThinkPad Rollable XD Concept

The ThinkPad Rollable XD is a more traditional vertical rollable concept that expands from a 13.3-inch screen to 16 inches. But rather than house the unseen parts of the screen in the base, this design has it wrapping around the top of the lid, giving you a dynamic "world-facing" display when the laptop is closed, or you haven't expanded the primary screen.

Of course, the screen needs to be protected, and Lenovo worked with Corning on a dual-sided Gorilla Glass Victus 2 screen cover with a 180-degree bend that wraps around both sides. And while there are other ways to control the expanding screen mechanism (like using your voice), you can swipe on the top edge of the display to move it up or down. When it's expanded, you can see the motor mechanism on the back. I like that the company left some of the complex engineering on display, though I doubt this kind of choice would make it to a final product.

Lenovo also touts AI features under the hood of the XD concept, like "live translation, voice assistant, multi-modal, and lid-controlled interactions." But I think one of the most interesting things about the XD concept is that it's a ThinkPad, unlike previous rollable concepts – and the ThinkBook Plus Gen 6 that the company brought to market last year. Perhaps the switch to arguably the company's flagship brand is an indication that the company sees a serious, productivity-focused future for the expandable screen tech it's been dabbling with for a few years now.

At the very least, housing the whole screen and the sliding mechanism in the lid should make it easier to take screens like this and add them to more laptops. And as someone who has taken a portable monitor and a mechanical keyboard to CES in previous years, pairing an expandable screen with one of the best laptop keyboards available has me at least a little more interested than I was before.

Lenovo is giving AMD’s Ryzen Z2 Extreme official SteamOS support for the first time. The Legion Go 2 launched in October last year, bringing the Z2 Extreme to the mobile gaming market and fixing many of the issues with Lenovo’s first handheld outing. Now, you can get it with official SteamOS support, combining the best aspects of the Legion Go 2 and Lenovo’s Legion Go S.

SteamOS is available directly from Valve for a DIY install, and Linux distros like Bazzite do a good job emulating Valve’s handheld interface. However, official support brings some perks. As we saw with the Legion Go S, SteamOS provides significantly better performance across games than Windows 11, even when comparing the exact same hardware inside the same chassis.

AMD’s Z2 Extreme is still at the heart of the Legion Go 2 with SteamOS. It’s an eight-core chip that leverages three Zen 5 cores and five Zen 5c cores, the latter of which trade peak clock speed for a space-optimized design. The Zen 5 cores can boost up to 5GHz while the Zen 5c cores top out at 3.3GHz. It defaults to a 28W TDP, though it has an adjustable TDP range from 15W up to 35W. The gaming grunt comes from the integrated RDNA 3.5 GPU, which sports 16 CUs.

Alongside AMD’s chip, Lenovo offers up to 32GB of LPDDR5x memory and 2TB of storage, plus an additional 2TB available through the microSD card slot. The memory is clocked at 8000 MT/s for the 32GB model and 7500 MT/s for the 16GB model. Lenovo is using the same OLED screen featured on the Windows 11 version of the Legion Go 2, which is an 8.8-inch screen with a resolution of 1920 x 1200, a 144Hz refresh rate, 10-point touch support, variable refresh rate, and a certification for VESA’s DisplayHDR True Black 1000.

Everything about the SteamOS Legion Go 2 is identical to the original model, short of the operating system. You still get dual USB 4 ports for charging or accessories, detachable controllers, a touchpad, six programmable buttons, and Hall Effect joysticks.

Lenovo says the Legion Go 2 with SteamOS should be available in June, starting at $1,199. The Legion Go 2 with Windows 11 is listed at $1,349 for the 32GB model, so the 32GB model of the SteamOS version should land somewhere around that price. It’s worth highlighting that the SteamOS version of the Legion Go S has a slightly lower price than the Windows 11 version, so hopefully that trend continues with the Legion Go 2. But with the way RAM prices are right now, who can predict what pricing might be like in six months?

Refreshed Legion laptops with new AMD branding

Lenovo is refreshing its Legion 5 and 7 ranges, and for the first time ever, properly noting configurations that come with an AMD CPU. Lenovo usually notes configurations with an Intel processor with an “i,” such as the Lenovo Legion 5i, while the AMD configuration would normally be called the Lenovo Legion 5. Now, Lenovo is adding an “a,” turning the Lenovo Legion 5 into the Legion 5a.

Naming aside, Lenovo is updating the Legion 7a, Legion 5i, and Legion 5a. The higher-end Legion 7a includes one of AMD’s new Ryzen AI 400 ‘Gorgon Point’ CPUs. The laptop carries up to an AMD Ryzen AI 9 HX 470, which features 12 cores (4x Zen 5 / 8x Zen 5c) and an RDNA 3.5 iGPU with 16 CUs. You can also pack in up to an Nvidia RTX 5060 mobile GPU, 64GB of LPDDR5x-8533 (capped at 8000 MT/s), and 2TB of storage split across two M.2 2242 PCIe Gen 4 slots.

Although the design of the Legion 7a remains mostly unchanged, Lenovo claims the laptop is 10% lighter and up to 5% thinner than the previous generation.

The Legion 5 range tops out slightly lower. The Legion 5i carries one of Intel’s new Panther Lake CPUs, the 16-core Core Ultra 9 386H. It’s not technically the flagship SKU, which is a title claimed by the Core Ultra X9 388H with its beefy integrated GPU. The Core Ultra 9 386H is almost identical outside of the GPU, however, sporting 16 total cores split across four performance cores, eight efficiency cores, and four low-power efficiency cores, along with a 4.9 GHz maximum boost clock on the P-cores.

For the Legion 5a, Lenovo is releasing two versions – one tops out with a new Ryzen AI 9 465 with 10 cores and 12 RDNA 3.5 CUs, while the other uses the older Ryzen 7 250; a Zen 4-based chip with eight cores. All three configurations top out with an RTX 5060, up to 32GB of DDR5-5600 memory, and 2TB of PCIe Gen 4 storage.

You can configure all of the laptops with one of Lenovo’s TrueSight OLED displays. The Legion 5 models use a 15.3-inch panel with a 2560 x 1600 resolution and 165Hz refresh rate, while the Legion 7a comes with a 16-inch panel with the same resolution and a 240Hz refresh rate. The Legion 5s carry a VESA DisplayHDR True Black 600 certification, while the Legion 7a holds a coveted DisplayHDR True Black 1000 badge.

Lenovo says all four variants should be available by April. The Legion 7a starts at $1,999, and the Legion 5a starts at $1,499 with Gorgon Point and $1,299 with Hawk Point Refresh. The Legion 5i is slightly more expensive, starting at $1,549.

In addition to Legion refreshes, the budget-focused LOQ range is getting a refresh with the LOQ15AHP11, starting at $1,149 with an AMD Ryzen 7 250, as well as up to an RTX 5060 mobile GPU and 32GB of DDR5-5600 memory. It’s a name bump alone, as this model already carries these same specs. Lenovo revealed the LOQ15IPH11, as well, which features one of Intel’s new Core Ultra 7 356H CPUs. However, Lenovo isn’t selling this model in the U.S.

Hot on the heels of one of AMD's David McAfee suggesting that the company might dust off the blueprints for its Zen 3 CPUs for another production run, another exec let out a hint about AMD's future plans with FSR 4. In an interview at CES, AMD's president of GPU Technologies and Chief Software Officer Andrej Zdravković suggested that the FSR 4 Redstone technology might be open-sourced. The statements come in the wake of an accidental release of the FSR 4 code earlier this year.

The hint came during a Q&A that Tom's Hardware attended when Chips and Cheese's George Cozma asked Zdravković if an open-source release was in the cards. The software head responded that the accidental public release in August was unexpected, but that it intends to release the source for the FSR4 library while keeping the core technology closed, so as not to give Nvidia engineers an advantage.

He continued by stating that AMD intends to "work as openly as possible," and was then asked to be specific about FRS 4's open release. Zdravković then said "that's the long-term plan," seemingly corroborating an earlier remark that "open sourcing is in [AMD's heart and mind]".

While the Radeon chief's words aren't a direct statement, it's likely that FSR 4 will see an official open-source release sooner rather than later. The August leak was by way of a GitHub repository that was part of FidelityFX SDK, which itself bears a broad MIT license, except for a handful of specific files.

Many have taken this to mean that although AMD pulled the source code down, theoretically it can't put the genie back in the bottle, as all of the published data, including the core AI model and its weights (a critical piece to replicate the technology), was arguably MIT-licensed for a brief moment in time.

Keeping FSR 4 as an exclusive selling point for 9000-series Radeon cards might not really be a realistic option anymore, either. As soon as they got their hands on the source code, industrious modders quickly tweaked the algorithm to work on RDNA 3 and older cards, and even on 3000-series GeForces.

While this comes with a hefty performance penalty versus running it on contemporary cards, many found the quality-versus-performance tradeoff very much worth it, given the significantly better output of the FSR 4 model. Don't be surprised if that FSR 4 GitHub repository pops up again soon.

Nvidia chief executive Jensen Huang referred to AI-controlled robots as "AI immigrants" at a question and answer session with reporters at CES 2026 that Tom's Hardware attended. Huang's comment related to how robotics may do jobs that others do not wish to take on.

Huang started off by saying that "having robots will create jobs," before turning on to discuss a global labor shortage, which he said is tens of millions of people.

"We no longer, as a population, will be able to sustain the economies that we would like to have," Huang said. "And so we need to have more, you know, if you will, AI immigrants to help us on the manufacturing floors and do the type of work that maybe we decided not to do anymore."

He followed up by saying that the "robotics revolution" will drive the economy forward, and that will allow for more jobs to hire more people. "We just need the economy to do well," Huang said. "We need inflation to stay low so that, you know, more jobs will be created, living will be more affordable. All of that's going to come with AI."

The comment didn't cause any gasps or reactions from the crowd of press at the event. Many will think the suggestion that human immigrants do jobs that others don't want to do is already an oversimplification of a complex situation. But it becomes even more complicated when comparing AI robots to the human condition, especially when there is fear about whether humans may have jobs at all.

Huang isn't the first tech leader to say that AI will create jobs. Many have done so with the idea that with agentic copilots, office workers will become far more efficient. He also said that "there are a lot of jobs that won't be replaced by AI for a very long time." But Huang's suggestion that physical, blue-collar labor in manufacturing or other areas could go may scare those who thought their woodworking skills, HVAC certification, or pure grit and determination would keep them working, no matter where they come from.

The CEO also said that he expected to see robots with human-level skills "this year," with work on locomotion, articulation, and fine motor skills.

"We don't just use our eyes, we also use touch," he said. "And the robot only has eyes, so it needs to have touch and so those fine motor skills are hard, hard to develop, but we're developing technology in that area, and I know the rest of the industry is doing so as well."

Nvidia says demand in China for its H200 data center GPU is "very high," but the company is deliberately avoiding timelines or announcements as it works through the final stages of U.S. export licensing, according to comments from CEO Jensen Huang at CES 2026. The remarks come nearly a month after the Trump administration said it would approve licenses allowing Nvidia to resume limited shipments of advanced AI accelerators to Chinese customers, subject to a 25% tax and additional oversight.

Speaking during a CES 2026 Q&A session with journalists, including Tom's Hardware in Las Vegas, Huang confirmed that Nvidia has already restarted its supply chain for H200 and is preparing to ship once the remaining regulatory details are resolved. He declined to offer dates or volume targets, instead emphasizing that actual purchase orders will determine how much hardware ultimately moves into China.

"The customer demand is high. It’s quite high. It’s very high," Huang said. He added that Nvidia has "fired up" its supply chain and that H200 systems are already flowing through production. The gating factor, he said, is the completion of licensing with the U.S. government, after which Nvidia expects orders to speak for themselves.

The restrained tone is notable given the political sensitivity around advanced AI hardware exports. Under the revised policy announced in December, Nvidia can sell H200 GPUs into China, but only under a licensing regime that includes a 25% charge and government review of shipments. It’s understood that Nvidia aims to begin shipments before Lunar New Year in mid-February, potentially using existing inventory to fulfil early orders while longer-term production ramps up. Those shipments, however, remain contingent on formal approvals.

Huang also made clear that Nvidia does not plan to trumpet individual deals or frame the restart of China sales as a major corporate milestone. “We’re not expecting any press releases or any large declarations,” he said, adding that the company expects to learn about market uptake through normal commercial channels.

This level of caution is understandable given the broader uncertainty surrounding the Chinese market. While Chinese customers are eager to regain access to high-end Nvidia accelerators, Beijing is heavily focused on promoting domestic alternatives. It has been suggested that Chinese authorities may consider conditions that encourage the use of locally produced chips alongside imported hardware, adding another layer of complexity for foreign suppliers.

Technically, the H200 remains a highly attractive option for large-scale AI workloads. Based on Nvidia’s Hopper architecture, it pairs the H100 GPU with 141GB of HBM3e memory and significantly higher memory bandwidth, making it particularly well-suited for training and inference of large language models. That capability gap is one reason Chinese customers have continued to pursue Nvidia hardware despite export controls and higher costs.

For now, Nvidia’s approach to China is to wait and see — the supply chain is warming up, and customer interest is strong, but the outcome depends on whether regulators on both sides give the final stamp of approval.

At a Q&A with the company in Las Vegas for CES 2026, Tom's Hardware put the question to CEO Jensen Huang regarding ways Nvidia could ease pressure on the consumer gaming GPU market.

Nvidia took to the stage at CES in Las Vegas this week, but new consumer GPU hardware was conspicuously absent as CEO Jensen Huang instead touted the latest and greatest the company has to offer in the realm of heavyweight AI computing. With DDR5 prices skyrocketing, SSDs not far behind, and Nvidia's flagship RTX 5090 now fetching an eye-watering $4,000 at some retailers, bad news is everywhere for PC builders.

GPU pricing is facing a squeeze from both ends, with both increasing RAM costs and likely dwindling supply causing a price increase on Nvidia's top GPUs. There are, of course, ways to address this, one notable option being boosting production of older GPUs that rely on older process nodes, less DRAM, and older technologies.

Sketchy rumors have been floating around about the return of the Ampere-based RTX 3060 in 2026. Most notably, AMD has teased the return of some Zen 3 AM4 chips to ease the strain on PC gamers looking for upgrades, revealing that spinning up old tech isn't beyond the realms of possibility. So we straight up asked Nvidia.

"Hi Jensen, Paul Alcorn from Tom's Hardware. The prices of gaming GPUs, especially the latest and greatest, are really becoming high, which might be due to some restrictions on supply and production capacity, one would assume. Do you think that maybe spinning up production on some of the older generation GPUs, on older process nodes where there might be more available production capacity, would help that, or maybe also increasing the supply of GPUs with lower amounts of DRAM? Are there steps that could be taken, or any specific color you could give us on that?

Huang: "Yeah, possibly, and we could possibly, depending on which generation, we could also bring the latest generation AI technology to the previous generation GPUs, and that will require a fair amount of engineering, but it's also within the realm of possibility. I'll go back and take a look at this. It's a good idea."

Huang's non-committal answer doesn't shed much light on the company's plans, but bringing back older GPUs is clearly not off the table, even if the answer suggests it's not something the company has previously considered. This may signal that Nvidia doesn't think the issue is getting out of control just yet. Supply will inevitably improve as time passes and chipmakers dramatically increase production on the latest process nodes, but we are likely years away from GPU supply reaching a healthy enough balance to feed both the consumer and AI data center markets. In the meantime, an increase in the number of GPUs built on older architectures and process nodes might be our only hope for affordable gaming GPUs.

On the software side, the possibility of newer AI-driven features that will boost performance to those older GPUs would also be a boon not only for those with existing gear but also for those who would be forced to grudgingly buy discrete older-gen GPUs as the only affordable option. And while lower VRAM capacities absolutely have an impact on gaming performance, newer AI-driven features like DLSS do help offset that enough to make them at least serviceable.

There remain trade-offs; Nvidia's latest DLSS 4.5 model cuts performance on older GPUs significantly, so as Jensen says, Nvidia would have to do an awful lot of work behind the scenes to make that prospect a reality. Ultimately, that could be helpful, at least until some unknown point in the future where a GPU with plenty of VRAM doesn't cost as much as a used car.

Micron just announced the 3610 SSD, which offers PCIe 5.0 speeds with the higher density of Quad-Level Cell (QLC) chips. According to the company’s press release, this chip uses its G9 NAND that lets it achieve competitive PCIe 5.0 performance while offering more storage space in the same footprint. Because of this, Micron says that this is the first SSD in the world to offer 4TB in a compact single-side M.2 2230 form factor, allowing manufacturers to pack in so much more memory in thin-and-light laptops and handheld devices. Aside from the greater storage density, it’s also touted to have 43% better performance per watt for better power efficiency and battery life.

“The 3610 SSD combines cutting-edge PCIe Gen5 technology, Micron’s most-advanced G9 QLC NAND, and a sleek, single-sided design to deliver premium performance, capacity, and power efficiency,” Micron Mobile and Client Business Unit senior vice president Mark Montierth said. “The 3610 will enable ultra-thin devices that meet the growing demands of on-device AI, immersive streaming, and performance-intensive workloads.”

The memory and storage chip maker has previously used G9 TLC NAND on its high-end 4600 PCIe 5.0 NVMe client SSDs, which deliver a blistering 2.1M IOPS random read and write speeds, and up to 14,500 MB/s and 12,000 MB/s sequential read and write speeds in a 2280 form factor. While these are the best speeds that you can get from the company, it’s only available in 512GB, 1TB, 2TB, and 4TB capacities. While the brand-new 3610 SSD offers relatively lower speeds, Micron says that they’re still good enough for AI applications, capable of loading a 20-billion-parameter model in under three seconds.

More importantly, using QLC technology on the 3610 SSD means that Micron can cram more chips in a single wafer, cutting production costs and making PCIe 5.0 SSDs more accessible. And even though the denser design of this chip means that it likely runs hotter than previous generations, it comes with host-controlled thermal management, allowing it to run under for extended periods without compromising reliability.

Even though the company markets this for productivity devices and AI applications, handheld gaming enthusiasts will likely be excited for this chip to hit store shelves, too, as it effectively doubles the largest available M.2 2230 SSD on the market. This means that users can install several AAA titles on their devices without having to worry about running out of space. We don’t have pricing and availability information for these chips yet, especially as they’re still under OEM evaluation. Furthermore, we hope that this technology alleviates the ongoing storage crisis a bit, especially as it allows Micron to eke out larger storage capacity from the same wafer size.

Phison’s SSD controllers already power many of the world’s best SSDs, and the company is set to extend its industry dominance. The new PS5037-E37T PCIe 5.0 SSD controller delivers blazing speed and remarkable power efficiency—engineered specifically for compact devices like laptops and gaming handhelds.

While Phison’s class-leading PS5028-E28 controller set the benchmark for raw performance, the new E37T takes a different approach by prioritizing exceptional power efficiency and affordability—without sacrificing impressive speeds. Manufactured by TSMC using an advanced 6nm process node, the E37T features a cutting-edge DRAM-less architecture and supports four NAND channels with transfer rates up to 4,800 MT/s. This innovative design enables the E37T to deliver a remarkable 38% performance boost over previous-generation PCIe 5.0 DRAM-less controllers, such as the PS5031-E31T.

The E37T delivers impressive sequential read speeds of up to 14.7 GB/s and sequential write speeds of up to 13 GB/s with an active power below 2.3W. According to Phison, the drive’s random performance is rated at up to 2,000,000 IOPS for both reads and writes. Compared to the flagship E28 controller, the E37T holds its ground in sequential performance: read speeds are only marginally slower, and write speeds are just about 7% lower.

However, the most notable differences appear in random performance, where the E37T registers about 23% lower random read and 33% lower random write speeds. Despite these differences, the E37T’s balanced approach to speed, efficiency, and affordability is what makes it unique.

At CES 2026, Phison delivered a show-stopping demonstration of the E37T prototype in action—surpassing even the company’s own lofty performance claims. The drive achieved blistering sequential read speeds of 14,926.78 MB/s and write speeds of 13,180.21 MB/s, all while maintaining an impressively low power draw of just 2.13W. For added real-world validation, a live CrystalDiskMark benchmark revealed the E37T-powered drive peaking at 4.47W under heavy load.

SSDs powered by the E37T controller will hit the market soon, ushering in a new wave of high-performance, energy-efficient storage options for compact devices. Engineered for versatility, the E37T will be featured in various drive form factors—such as M.2 2280, M.2 2242, and the ultra-compact M.2 2230—making it an ideal solution for everything from slim laptops to portable gaming consoles.

Phison E28 Roars Ahead: Next-Level Performance Gets Epic Capacity Upgrade

While the E37T grabbed headlines at CES 2026, Phison made waves with a major update on its flagship E28 controller. The company announced that next-generation E28-powered 8TB SSDs are right around the corner. With only a handful of consumer PCIe 5.0 8TB SSDs currently available—such as the WD Black SN8100, or should we say the Optimus GX Pro 8100, and Samsung 9100 Pro—Phison is set to seize a wide-open market with its E28-based drives. This move not only reaffirms Phison’s leadership in storage innovation but also promises to redefine what’s possible for high-capacity, high-performance SSDs.

We witnessed the E28 8TB SSD prototype in action—and it lived up to every bit of the hype. The drive shattered expectations, delivering blistering sequential read speeds of 14,919.18 MB/s and write speeds of 14,088.25 MB/s, soaring far beyond Phison’s own official numbers for the controller.

The biggest hurdle facing E28 8TB drives is the global NAND shortage and the insatiable demand created by the AI boom—a factor that’s already driven SSD and memory prices sky-high. As a result, these cutting-edge PCIe 5.0 8TB drives will likely come with a hefty price tag. For now, the E28 8TB SSDs could remain out of reach for most consumers. However, there is hope on the horizon: as the market eventually cools and the AI-driven frenzy subsides, SSD prices will begin to stabilize. When that happens, high-capacity drives like the E28-based 8TB ones will become more accessible, opening up exciting new storage possibilities without the premium price tag.

The proverbial news cannon that is CES is firing new products on all barrels, but every now and then, there's a stray remark that makes our ears perk up. In a round-table interview Tom's Hardware attended at CES 2026 in Las Vegas, AMD's David McAfee was discussing the sorry state of the ongoing chip crisis, and he let out a hint that AMD could bring back older AM4 desktop chips, presumably 5000-series Ryzen processors and APUs based on the Zen 3 architecture.

As even non-techies know by now, buying a brand new computer is a prohibitively expensive proposition, thanks mainly to the absurd prices of DDR5 memory. In addition, folks moving from machines that are just four years old will find themselves in the unenviable position of having to buy overpriced memory and a new motherboard on top of that, as the move to DDR5 also implies a socket change for both Intel and AMD chips.

When questioned about the rock-and-hard-place situation these users are in, McAfee stated that AMD "[is] certainly looking at everything that [it] can do to bring more supply and kind of reintroduce products back into the [AM4] ecosystem to satisfy the demands of gamers that maybe want that significant upgrade in their AM4 platform without having to rebuild their entire system", further adding that he thinks this is "definitely something [AMD is] very actively working on."

It should be noted that a remark by one person does not make for a company-wide mission statement, but at least at face value, this move would make full sense for both AMD and customers. Furthermore, the context for the aforementioned statement was that AMD's telemetry obtained through the Adrenalin software corroborated that a significant portion of users are still running 2000- and 3000-series chips.

In addition to that, McAfee noted that many of its retail partners are seeing higher numbers of CPU-only purchases, indicating that shoppers are buying new-old chips to grant a tangible speed boost to their existing machines in these troubled times, where just buying 32 GB of DDR5 memory, a new CPU, and a motherboard will easily bite you for over a grand.

While the scenario described is of someone upgrading just their CPU, older machines will likely have 8 to 16 GB of memory, thus asking for another DIMM or two. DDR4 prices have also been steadily rising, though Samsung has reversed its decision to stop DDR4 production, while SK hynix has reportedly increased DDR4 production at its Wuxi facility. Looks like, for the time being, keeping existing machines going is the only option many enthusiasts will have.

Old-timers around here will remember the ill-fated Optimus keyboards with LED displays underneath the keys. They were quite ahead of their time, but alas, never widely produced. The advent of streaming and the growing popularity of shortcut pads revitalized the notion, though, and Corsair's Galleon 100 SD combines the best of both worlds.

Describing the Galleon in one sentence is simple: it's a near-full-sized keyboard, except the usual numeric pad is instead an Elgato Stream Deck. The trick here is that it's all one unit, thus obviating the need for separate cabling, configuration utilities, and desk space.

The Galleon's Stream Deck contains 12 keys, each with its own LED display, resting underneath a main 720x1280 screen along with two dials. Each key is fully customizable, and although the Galleon seems targeted at streamers first, the pad can be used with many apps and games, from Spotify to Final Fantasy.

In fact, not only has Corsair posted multiple guides detailing the extensive software controls, but the Galleon also supports plugins from the Elgato Marketplace, so once you have a Galleon, you can sail there right away for customization booty. Notable application-specific plugins include Twitch, YouTube, Discord, OBS Studio, and a multitude of games.

As for the keyboard itself, it's a pretty fancy unit all of its own. It has pre-lubed Corsair MLX pulse linear switches, resting on a gasket-mounted PCB and six layers of sound dampening. The switches are hot-swappable, and you can employ three- or five-pin switches as the board supports both.

All that should provide quite a comfortable and quiet gaming experience, and it's further improved by the 8 KHz polling rate and Flash Tap SOCD. That's Corsair's take on prioritizing the last pressed movement key so that, for example, it's easy to counter-strafe in Counterstrike 2 by pressing the opposite direction without releasing the first.

The Corsair Galleon 100 SD is pricey at $349.99, but it seems like decent value, as a Stream Deck is about $135 all on its own. Corsair says these will ship between January 30 and February 2, essentially two-upping every other potential competitor by actually existing in the market and arriving with an application ecosystem ready to roll.

There are other options, but none quite as solid. The Finalmouse Centerpiece Pro keyboard is similarly priced but only expected to materialize in September, while the Flux keyboard seemingly never went into full production. And although the aforementioned Optimus keyboards from Art Lebedev looked fantastic, very few made their way onto users' desks.

Today at CES, Broadcom is expanding the family with two new dual-band Wi-Fi 8 chips that combine 2.4 GHz and 5 GHz radios into a single chip. The BCM6714 is the lower-end offering, with a 3x4 arrangement. That means it supports three spatial streams on the 2.4 GHz band and four on the 5 GHz band. The BCM6719 ups the ante by delivering four spatial streams across both bands.

And with that news, I can already see the collective eyerolls of the enthusiast community. When Wi-Fi 7 launched, the first devices to hit the market were more expensive tri-band solutions that supported the 2.4 GHz, 5 GHz, and 6 GHz bands. Later, more affordable networking products arrived, supporting only the 2.4 GHz and 5 GHz bands. While this allowed manufacturers to hit a lower price point, it also created confusion in the marketplace – just because a router was labeled Wi-Fi 7 didn't mean it supported the full feature set. That confusion will live on for another generation.

These two new entry-level chips join the BCM6718, which Broadcom announced last year, one of its first Wi-Fi 8 (802.11bn) compliant chips aimed at the consumer/residential wireless market. It's a 4-stream tri-band Wi-Fi (2.4 GHz, 5 GHz, and 6 GHz).

But that's not all; Broadcom also announced the BCM4918 Wi-Fi 8 APU, which is compatible with the BCM6714, BCM6719, and BCM6718. This system-on-chip is designed for high-performance computing and AI acceleration. Not only does it include an onboard neural engine for on-device AI/ML inference and acceleration, but it also features dedicated network engines to handle wireless and wired traffic, bypassing the CPU. It supports multi-gigabit Ethernet for powering your high-flying wired network. The chip also enabled Edge-AI processing and real-time network optimization.

Broadcom says it is already sampling the BCM6714 and BCM6719 to its "early access customers," pointing to general availability in consumer-grade networking products by the end of 2026, a similar timeline to MediaTek's. This seems doable, as we've already seen prototype Wi-Fi 8 hardware this week at CES, and Asus even told us that its first-generation Wi-Fi 8 products will launch later this year, with second-generation hardware coming in 2027.

Wi-Fi 8 is designed to improve reliability and lower latency rather than deliver a massive uplift in theoretical speeds (as had been the case with previous Wi-Fi iterations). Instead of obscene theoretical speeds that consumers likely won't see in the real world, the IEEE says that Wi-Fi 8 will deliver up to a 25 percent improvement in real-world speeds while reducing latency. Asus specifically highlighted improvements to mid-range throughput and showcased a 10 percent uplift in throughput over its fastest Wi-Fi 7 router with early prototypes.

Asus is going all out at CES this year, bringing a range of embedded devices like its new Zephyrus laptops, but also exciting PC components — such as the new ROG Cronox case. It's an e-ATX behemoth capable of sporting up to fourteen 120mm fans, along with the biggest GPUs money can buy today. It also just happens to feature a large 9.2-inch LCD, too.

That secondary display is at the bottom, just underneath the motherboard and above the fans. It shows system stats like temperatures and fan speeds, but it can be customized to play whatever animations or videos you want. The screen module sits on a hinge, so it can actually be angled outward to face you regardless of where your PC is situated.

We checked out this display in-person and, as you can see, it's pretty imposing. The case looks fantastic in person, and the screen is an FHD-class display with a resolution of 1920x400 (height cut short) running at 60 Hz. The screen looks very sharp, but the ROG logo makes some of the metrics harder to read, which may concern some users.

The case itself is built out of aluminium and features a panoramic design with tempered glass curved very subtly at the crease. Like other chassis in this segment, this gives you an aquarium-like aesthetic where every part of your build is in clear sight. That's why even the fans on the side can be rotated slightly toward the inside, and to ensure optimal airflow.

Speaking of, there's another fan at the bottom-rear responsible for directly blowing fresh air toward the GPU, so the card's fans can suck up fresh air at all times. Moreover, the case supports up to two 360mm radiators and GPUs up to 400mm in length can fit without issue, along with CPU coolers up to 180mm in height.

The ROG Cronox is also BTF-ready, meaning it's fully compatible with back-connect motherboards. Asus itself has new AMD motherboards and "wireless" AIOs that work without any cables, using pogo pins to communicate. And with the BTF 3.0 standard on the way, this might be the perfect time to hop on the cable-less look trend.

There's no word on pricing or availability, but the case does come in both black and white. There'll be an "ARGB" variant, which just means it will come with four of Asus' new ROG Eurux GR120 fans that you've seen in the pictures so far. These are the company's flagship offering with all the bells and whistles, like daisy-chain support, launching in conjunction with the Cronox.

Those who tuned into Nvidia’s CES keynote on January 5 may have found themselves waiting for a familiar moment that never arrived. There was no GeForce reveal and no tease of the next RTX generation. For the first time in roughly five years, Nvidia stood on the CES stage without a new GPU announcement to anchor the show.

That absence was no accident. Rather than refresh its graphics lineup, Nvidia used CES 2026 to talk about the Vera Rubin platform and launch its flagship NVL72 AI supercomputer, both slated for production in the second half of 2026 — a reframing of what Nvidia now considers its core product. The company is no longer content to sell accelerators one card at a time; it is selling entire AI systems instead.

From GPUs to AI factories

Vera Rubin is not being positioned as a conventional GPU generation, even though it includes a new GPU architecture. Nvidia describes it as a rack-scale computing platform built from multiple classes of silicon that are designed, validated, and deployed together. At its center are Rubin GPUs and Vera CPUs, joined by NVLink 6 interconnects, BlueField 4 DPUs, and Spectrum 6 Ethernet switches.

Each rack integrates 72 Rubin GPUs and 36 Vera CPUs into a single logical system. Nvidia says each Rubin GPU can deliver up to 50 PFLOPS of NVFP4 compute for AI inference using low-precision formats, roughly five times the throughput of its Blackwell predecessor in similar inference workloads. Memory capacity and bandwidth scale accordingly, with HBM4 pushing hundreds of gigabytes per GPU and aggregate rack bandwidth measured in the hundreds of terabytes per second.

These monolithic Vera Rubin systems are designed to reduce the cost of inference by an order of magnitude compared with Blackwell-based deployments. That claim rests on several pillars: higher utilization through tighter coupling, reduced communication overhead via NVLink 6, and architectural changes that target the realities of large language models rather than traditional HPC workloads.

One of those changes is how Nvidia handles model context. BlueField 4 DPUs introduce a shared memory tier for long-context inference, storing key-value data outside the GPU frame buffer and making it accessible across the rack. As models push toward million-token context windows, memory access and synchronization increasingly dominate runtime. Nvidia is seems to be taking the view that treating context as a first-class system resource, rather than a per-GPU issue, will unlock more consistent scaling.

This emphasis on pre-integrated systems reflects how Nvidia’s largest customers now buy hardware. Hyperscalers and AI labs deploy accelerators in standardized blocks, often measured in racks or data halls rather than individual cards. By delivering those blocks as finished products, Nvidia shortens deployment timelines and reduces the tuning work customers must do themselves. CES became the venue to outline that vision, even if it meant leaving traditional GPU announcements off the agenda.

No New GeForce cards

Suddenly, the lack of a new GeForce announcement becomes a whole lot easier to explain. Nvidia’s current consumer line-up of 50-series GPUs is still pretty fresh, and it continues to command prices in excess of $3,500 per unit. Introducing an interim refresh would carry higher costs at a time when memory pricing is at all-time highs, and supply remains tight. The company has also leaned more heavily on software updates, particularly DLSS and other AI-assisted rendering techniques, to extend the useful life of existing GPUs.

From a purely commercial perspective, consumer GPUs now represent a smaller (and, unfortunately, shrinking) share of Nvidia’s revenue and focus than they did even two years ago, let alone five. Data center products tied to AI training and inference account for the majority of growth, and those customers need system-level gains, not incremental improvements in graphics performance.

Lisa Su, during her AMD keynote, said it best: “There's never been a technology like AI.” CES, once a showcase for new PC hardware, has become a stage for AI announcements. This does not mean Nvidia — or AMD for that matter — is abandoning gaming or professional graphics. Rather, it suggests a lengthening cadence between major GPU architectures. When the next GeForce generation arrives, it is likely to incorporate lessons from Rubin, particularly around memory hierarchy and interconnect efficiency, rather than simply increasing shader counts.

Faith in CUDA

Nvidia’s system-centric approach inevitably invites comparison with rivals pursuing similar strategies. AMD is pairing its Instinct accelerators with EPYC CPUs in tightly coupled server designs, while Intel is attempting to unify CPUs, GPUs, and accelerators under a common programming model. Apple has taken vertical integration even further in consumer devices, designing CPUs, GPUs, and neural engines as a single system on a chip.

What distinguishes Nvidia is the depth of its software stack. CUDA, TensorRT, and the company’s AI frameworks remain deeply entrenched in research and production environments. By extending that stack everywhere it can, Nvidia increases the switching cost for customers who might otherwise consider alternative silicon. There are risks to this approach, and large customers are increasingly exploring in-house accelerators to reduce dependence on a single vendor, and complex rack-scale systems raise the stakes for manufacturing or design issues. Because of this, Nvidia’s ability to deliver Rubin on schedule will matter just as much as the performance metrics presented at CES.

Still, the decision to use CES 2026 to spotlight Vera Rubin rather than a new GPU points to where Nvidia sees its future. Let’s face it: We, and Nvidia, all know that the next phase of computing will be defined less by individual chips and more by how effectively those chips are integrated into scalable systems. Nvidia is therefore aligning itself with where the demand and investment are, even if that means placing less emphasis on the hardware that defined the company for decades.

At CES 2026, Asus revealed its new lineup of notebooks, retooling its Zenbooks with new chips from AMD, Qualcomm, and Intel. Prices and availability, however, remain a mystery.

Kicking things off is the Asus Zenbook S series. Both models are equipped with an Aluminum chassis and Windows Hello FHD IR cameras. The 14-inch S14's new configurations are equipped with Intel's new Panther Lake-based Core Ultra lineup, sporting up to 32GB of LPDDR5X RAM, and a swanky 3K OLED display, with a 120Hz refresh rate and WQXGA+ (2880 x 1800) resolution. At its heaviest, these models come in at 2.7 pounds, and in two colorways, which Asus has dubbed Scandinavian White and Antrim Grey.

Meanwhile, the Zenbook S16 comes equipped with AMD's latest AI400 mobile processors. However, aside from this, you still get the same resolution display, albeit slightly larger, in addition to an SD card reader, as well as a slightly larger battery. Of course, this comes at the tradeoff of weight, with the Zenbook S16 coming in at almost three and a half pounds, at its heaviest. The S16 also sports identical colorways to its smaller sibling.

Zenbook A14 & A16

The Asus Zenbook A14 and A16 are similar products, with one big difference: They're based on Snapdragon X2 Elite processors. The A14 is limited to utilizing the 18-core Snapdragon X2 Elite, meanwhile the A16 boasts up to a Snapdragon X2 Elite Extreme processor.

The A14 can come configured with up to 32GB of LPDDR5X RAM, with the A16 can cramming in up to 48GB of LPDDR5X, which may be handy for running local AI applications. Otherwise, expect much of the same when compared to the S-series above. Note that the 14-inch model has a lower-resolution display and a slower refresh rate, too. Both models sport an FHD IR camera, built-in speakers, and up to 130W charging via USB-C. Unlike the S14 and S16, the A14 and A16 are only available in Zabriskie Beige.

Asus Zenbook 14: AMD & Intel

For a more budget-oriented consumer, the standard Asus Zenbook offers slightly more modest specs, in addition to fewer frills compared to the aforementioned A and S lineups. Four SKUs have been announced, with one announced with Intel's Core Ultra 9 285H, based on the older Arrow Lake architecture, and a 'Quiet Blue' colorway. The Intel SKU offers a 120Hz display, whereas refresh rates are not mentioned for the following three AMD SKUs; you also get an additional Thunderbolt 4 port, compared to the AMD models.

The three AMD SKUs are only differentiated by availability, RAM capacity, and OS choice. If you buy at Walmart, expect to get Windows 11 Home, but if you buy from the Asus Store, you'll get Windows 11 Pro.

Aside from that, expect a 2.82-pound, lightweight chassis and a 14-inch OLED display capable of up to 400 nits of brightness. No refresh rates have been publicized for the AMD models here, so it may only carry a 60Hz refresh rate.

Asus Zenbook Duo

For those all about productivity while on the move, the Asus Zenbook Duo offers a large, dual-screen display, for much more screen real estate than a standard notebook, similar to its predecessors.

The laptop chassis is built using Ceraluminum, and it also comes with a built-in kickstand, made from the same material. Since you have to drive two displays, you also get a 99 WHr battery to make up for the increased power usage. You'll also get top-shelf 1000-nit OLED panels here, as well as a smaller design, which Asus claims offers a 5% smaller footprint, in addition to support for Asus Pen 3.0.

We're rapidly approaching a point where integrated graphics could potentially dethrone entry-level discrete GPUs. They're good enough for the best handheld gaming PCs, and now Asus is trying it in a laptop.

At CES 2026, the company is putting AMD's Strix Halo into the Asus TUF Gaming A14, a thin gaming laptop. It has tried this once before, putting a version in the Flow Z13 tablet. But with its standard clamshell design, the A14 is a much more appealing machine to more people. We've previously had versions of this laptop earn spots on our best gaming laptops list.

Asus' spec sheet says that the company is using a version of Strix Halo with 12 cores and 40 compute units. That lines up with the Ryzen AI Max+ 392, which AMD announced last night.

As far as gaming PCs go, the A14 is thin at 0.67 inches at its sleekest point, though it maxes out at 0.78 inches thick. The system weighs 3.6 pounds.

The system supports up to 32GB of RAM and 1TB of storage, as well as a 2560 x 1600, 165Hz display. All of this suggests that this version of the A14 won't be priced at a budget level and may be sold as a premium option. Asus didn't announce a release date or price as of publishing.

The system goes up to 85W of total thermal capacity with the system in Turbo Mode, which you'll want to be plugged in for. The system supports 100W of charging over USB Type-C.

Asus is selling the A14 in a "jaeger gray" color. The laptop features a keyboard with 1.7 mm of travel, a 16:10 glass touchpad, and two M.2 2280 slots.

We saw Strix Halo chips perform well in the Framework Desktop, albeit with more RAM, so we're excited to get our hands on this more mainstream laptop and test it to see just how well it works in a slightly more mainstream approach.

I did not have a partnership with Asus’ gaming arm and planar magnetic pioneers HiFiMan on my CES 2026 bingo card, but the ROG Kithara open-back planar magnetic gaming headset has just been announced. And after spending a few days with the fancy gaming cans before heading to Las Vegas, I can confirm they sound amazing – especially when I broke out a few CDs rather than relying on streaming. And at least to my head, they are extremely comfortable, despite weighing in at 0.93 pounds. That said, they will undoubtedly remain a niche product, as they come with all the trappings of audiophile headphones, like easily lost removable cables and adapters, a large, travel-unfriendly design, and a $359 price that – while reasonable for what you get – is pricier than our current pick for best gaming headset for audiophiles, Audeze’s Maxwell.

Asus ROG Kithara Specifications

Asus clearly wants this headset to feel like an audiophile investment. Opening the box, I was greeted by a removable “sound signature certificate” and a note signed in foil ink by Kris Huang, the company’s GM of the gaming and workspace gear business. Under this foam-backed sheet, the headset and its four connectors (3.5 mm, 4.4 mm, and 6.3 mm) present themselves, along with a USB-C adapter that lets you connect the headphone and audio jacks and plug into a PC or other device. The various connectors screw onto the balanced audio cable, which plugs into the bottom of each ear cup. This is all pretty standard stuff for high-end wired headsets (or in-ear monitors), but will probably feel fiddly and foreign to many coming from the gaming headset realm.

But that’s just for using the Kithera as a pair of headphones. If you want to use the mic (you know, to use the headset as a headset), you’ll need to use the shorter braided cable with its permanently attached mic and in-line controller for muting and volume adjustment. The microphone plugs into the left earcup, with the other jack connecting on the right side. This cable is OK, but at only about seven feet, it could be restrictive if you plug the headset into the jacks at the rear of your PC (as you generally should for the best possible audio).

The in-line controller on this cable is the cheapest-feeling part of the Kithera kit, by far. The front houses a large plastic slider for muting the mic (with some red plastic visible so you can tell you’re muted if you look down at it). And there’s a volume wheel on the side, which I’d honestly suggest just not using. Like the in-line controller as a whole, it feels cheap, and it is so sensitive that you can go from zero to max volume with just slightly more than one roll of your thumb. When starting at full volume, I found it difficult to get to a pleasing level of around 75%. As soon as I started to move the wheel, the sound output was too quiet. I quickly resorted to just leaving the volume roller at max and adjusting sound output levels from my keyboard.

In all, there are a lot of connectors, adapters, and cables here. I personally feel like I’d lose something important before too long. But Asus does at least include a full-sized sheet of translucent paper documenting how to connect and use everything. I’d probably lose that, too.

The headset itself feels quite solid, and the only nod to its gaming cred is the ROG logo embossed on each large earcup. Otherwise, the design looks a lot like a stealthier version of HiFiMan’s Ananda line.

I’m not a huge fan of the bare steel headband, which feels a bit utilitarian for something this inexpensive, but the leatherette and mesh interior band, combined with cushy ear pads, make the Kithara the most comfortable headset I’ve worn in… possibly ever. Two sets of ear pads are included, with a second velour pair letting more sounds in and out than the pre-installed pads. To me, both were similarly comfortable, though I might prefer the velour pads in the warmer days of summer, as they feel a bit more breathable. Sound quality aside, there’s a lot to be said for paying a premium for comfort – especially in a device that you may be wearing for several hours a day if you use the Kithera both for work, gaming, and media playback. So I like that these cans are comfy, and that Asus included two sets of ear pads – which are also fairly easy to detach and swap.

These are open-back headphones, though. So don’t expect exterior sounds to be muted or silenced like with closed-back and / or noise-cancelling headphones. And people near you will definitely be able to hear what you are listening to.

Personally, I prefer open-back headphones for their wider sound stage and lower general ear fatigue. For the same reason, I prefer using my AudioEngine PC speakers when working and gaming. But I mostly work from home, with nothing to annoy or distract me other than my cat – open-backs certainly aren’t for everyone, especially if you don’t have other audio solutions for travel or in-office work.

But as these are large, wired headphones, it isn’t likely that many people will regularly be taking the Kithara on their commute or long treks of travel. I have a pair of Sennheiser Momentum 4s for on-the-go listening, and sometimes use them at home, too. But the ROG Kithara headset sounds better and is more comfortable for long-term listening.

I didn’t have enough time to truly test the headset’s removable mic in games. But in a quick check using Windows’ Sound Recorder, it seemed plenty sensitive, but not great at minimizing plosives, despite its foam covering.

Although its name is a mouthful, the Asus ROG Falchion Ace 75 HE, debuting at CES 2026, ticks a lot of the boxes for a modern gaming keyboard. You get magnetic switches (Asus’ own factory-lubed HFX V2, to be specific), 8K polling, Rapid Trigger, a web-based configuration utility (important because all gaming peripheral software is awful), and of course, RGB. The only thing I find lacking here is a battery for wireless operation. But considering the high polling rate, magnetic switches, and plethora of RGB, you’d probably want to keep a charging cable close by, anyway.

The keyboard’s stand-out features are a programmable touch bar on the back left edge (which we’ve seen in previous Falchions), and a small jog wheel near the upper-right corner. The latter isn’t for volume adjustments or scrolling documents, but for fine-tuning the actuation point and rapid trigger sensitivity on all keys or a specific key, without having to jump into software or even enter the web interface. The idea, it seems, is to get the actuation point close to where you want it via the browser interface, then fine-tune it from there. That said, it would be nice if there were a tiny two-digit readout near the wheel that would actually tell you what number you’re setting things to with the wheel.

This feels like an important feature on a modern magnetic gaming keyboard. Because no one wants to keep jumping in and out of a game until things feel right. And when you aren’t gaming, you might want an easy way to turn the sensitivity down so you can type without the frustrations of super-sensitive switches. This works well with the keyboard’s layout, as well. At 75%, it’s nearly as compact as 65% models, while still offering up arrow keys and one column of useful keys to the right of the backspace button for work. There’s also an easily flippable switch on the rear right corner that lets you turn Rapid Trigger on and off.

While I didn’t get enough time with the keyboard before heading to CES 2026 to do a full review, I did spend a couple of days with it, playing Borderlands 4 and editing / writing as 2025 wound down.

The lubed ROG HFX V2 switches feel smooth and pleasing, and I noticed effectively no key wobble. Their actuation range, between 0.1 and 3.5mm, is a bit shorter than most other HE keyboards / switches, which have a range of 0.1-4mm, but I certainly didn’t miss the extra distance. I slightly prefer those in the Keychron K8 HE for their slightly softer sound and more cushioned feeling when bottoming out, but I’d have no major complaints switching to the Falchion Ace 75 HE and its ROG HFX V2 switches – for gaming or for work. And if I were meticulously dialing in actuation points for my favorite games, I would prefer the Asus ROG keyboard because it makes it easier to turn some of those features off when I’m doing things other than gaming, thanks to the adjustment wheel and the Rapid Trigger switch.

That said, I don’t particularly find the touchbar on the rear of the Falchion Ace 75 HE particularly intuitive or ergonomic. Perhaps I need more time to get used to it, and I do like that you can change it via the Web interface to control volume, media playback, lighting intensity, and other features. But you have to reach back behind the top edge of the keyboard and tap or swipe. There’s also a button back there that lets you toggle between modes. Using this feature feels awkward, and I sometimes found myself accidentally pressing the first few function keys when trying to skip a song on a playlist. I may find it more intuitive with time, but it feels like having the same touch area on the left side of the keyboard, rather than the back, would be a better choice, since you wouldn’t have to reach up and over the Function keys to use it.

The switches are also swappable, although with all magnetic keyboards, compatible options are pretty limited. Aside from the HFX V2 that comes with this keyboard, Asus lists an HFX V2X switch, alongside a “3rd Party Golden Switch,” which appears to be the same HFX V2 switch, made by TTC, just without the Asus branding. Asus also says that the switches from Asus’ other recent HE keyboard (the Falcata and Azoth 96 HE) will also work. In case you have multiple high-end keyboards from Asus and want to swap your switches around.

The Falchion Ace 75 HE feels very solid and premium, thanks to a metal top and bottom plate and PBT doubleshot keycaps. Asus says it will be available in black or white, for $219 at Best Buy starting in January, and comes with a nice carrying case for those inclined to travel with their favorite keeb. I just can’t help feeling that those keyboard travelers might also appreciate an internal battery, so they aren’t always tethered to a cable.

The Consumer Electronics Show (CES) has officially kicked off this year in Las Vegas, and there was no shortage of announcements, which began pouring in furiously last night. There were no fewer than three major keynotes, with Nvidia, Intel, and AMD each taking the stage to share insights into what they're delivering for 2026 and beyond.

New Nvidia GPUs are a no-show, DLSS 4.5 is incoming, Rubin GPU set to accelerate the AI boom

Perhaps the biggest news for enthusiasts from Nvidia was that the company didn't announce any new GPUs. This time last year, Nvidia launched the RTX 50 Series of GPUs at CES, and there were initial rumors of a "Super" refresh. However, it wasn't meant to be: the company tweeted before its big CES keynote that "No new GPUs will be announced."

With the ongoing memory shortages and Nvidia's increasing focus on its highly profitable AI ambitions, it's not entirely shocking news. But it breaks Nvidia's five-year streak of announcing new GPUs at CES.

However, it wasn't a total loss for gamers: Nvidia also introduced DLSS 4.5 along with Multi Frame Generation 6X. According to Nvidia, DLSS 4.5 is taking additional steps to improve temporal stability and reduce visible artifacts that can mar on-screen images (shimmering/flickering and trailing "ghosts" behind objects are common complaints), as well as anti-aliasing. Although Nvidia is extending support for DLSS 4.5 to its legacy RTX 20- and 30-series cards, the lack of Tensor Core FP8 acceleration leaves some uncertainties regarding performance.

Given the world's insatiable appetite for everything related to AI, Nvidia introduced its new Vera Rubin NVL72 AI supercomputer. It consists of six distinct chips: the Vera CPU, the Rubin GPU, the NVLink 6 switch, the ConnectX-9 SuperNIC, the BlueField-4 data processing unit, and the Spectrum-6 Ethernet switch. Together, you're looking at a single Rubin GPU delivering up to 50 PFLOPS of inference performance (NVFP4) and 35 PFLOPS of NVFP4 training performance fed by 288GB of HBM4 memory.

• For the first time in 5 years, Nvidia will not announce any new GPUs at CES — company quashes RTX 50 Super rumors as AI expected to take center stage
• Nvidia introduces DLSS 4.5 and Multi Frame Generation 6X at CES 2026 — updated models can generate higher-quality upscaled frames and more of them, dynamically
• Nvidia launches Vera Rubin NVL72 AI supercomputer at CES — promises up to 5x greater inference performance and 10x lower cost per token than Blackwell, coming 2H 2026
• Nvidia CEO confirms Vera Rubin NVL72 is now in production

Intel Panther Lake is on the prowl

Intel announced 14 SKUs in the new Core Ultra Series 3 (Panther Lake) family. The new lineup also includes three SKUs with an "X" designation: the Core Ultra X7 358H, Core Ultra X7 368H, and Core Ultra X9 388H. These chips feature a beefier Arc B390 GPU with 12 Xe3 cores (up from 8 Xe2 cores on the most potent Core Ultra Series 2 chips), which should offer similar performance to a discrete Nvidia GeForce RTX 4050.

According to Intel, the flagship Core Ultra X9 388H offers up to 60 percent higher multi-threaded performance (Cinebench 2024) than the Core Ultra 9 288V. In addition, gaming performance is reportedly up to 76 percent faster than the Core Ultra 9 285H.

Intel is aiming to deliver the performance of Arrow Lake-H with the efficiency of Lunar Lake with its new Panther Lake family. It looks as though the company is well on its way to meeting that goal.

• Intel doubles down on gaming with Panther Lake, claims 76% faster gaming performance — new X-series chips can match discrete RTX 4050

AMD makes the best even better with the Ryzen 7 9850X3D, Zen 5 and Strix Halo get refreshed

There’s a new top dog in AMD’s desktop gaming processor family: the Ryzen 7 9850X3D. AMD claims this is the “new fastest gaming processor,” offering a 400 MHz clock-speed advantage over the Ryzen 7 9800X3D. This is good enough for an average 7 percent uplift in gaming performance for the Ryzen 7 9850X3D over the Ryzen 7 9800X3D.

The company also announced the Ryzen AI 400 “Gorgon Point” family, which uses a refreshed Zen 5 APU. There are seven SKUs in the family, topped by the Ryzen AI 9 HX 474, which features 12 cores, a 5.2 GHz max boost clock, 36GB of L2+L3 cache, and 16 RDNA 3.5 GPU cores.

• AMD’s Ryzen 7 9850X3D promises 7% uplift over Ryzen 7 9800X3D – AMD fights itself with ‘new fastest gaming processor’
• AMD's Ryzen AI 400 series includes the first Copilot+ desktop CPU — Team Red refreshes Zen 5 APUs and Strix Halo
• AMD unwraps Instinct MI500 boasting 1,000X more performance versus MI300X — setting the stage for the era of YottaFLOPS data centers

Asus, Alienware, Dell, MSI, and more announce new laptops for 2026

With AMD and Intel both announcing new mobile processors at CES, new laptops are, of course, incoming as well. Asus, for example, has announced that its new laptops use a mix of Ryzen AI 400 and Core Ultra Series 3 processors, while MSI’s latest laptops are strictly Intel-only. HP even showed Qualcomm some love with its new OmniBook Ultra 14 that supports the Snapdragon X2.

• Asus launches two new ROG Zephyrus laptops at CES — 14 and 16-inch models come with latest AMD and Intel CPUs, and up to RTX 5090 GPU
• Alienware brings OLED to its gaming laptops for the first time in years — anti-glare OLED display boasts 240Hz refresh rate and 0.2ms response time
• Acer updates Predator, Nitro gaming laptops at CES 2026 – Panther Lake finds its way across the lineup
• MSI’s newest 16-inch Raider and Stealth gaming laptops debut — Panther Lake options, OLED panels, and familiar RTX 50 GPUs
• Acer refreshes Swift laptops with Panther Lake — claims 'world's largest haptic touchpad' on Swift 16 AI
• HP puts HyperX name on Omen gaming laptops — new systems get Intel and AMD's latest processors
• MSI’s Prestige ultra-thin laptops and convertibles embrace Panther Lake – up to Core Ultra X9 with standard OLED panels
• Dell brings back XPS laptops — ditches the capacitive touch bar, adds 1Hz display option, and upgrades 14 and 16-inch models
• HP's new OmniBook Ultra 14 gets Panther Lake and Snapdragon X2 options inside — exclusive variant of Qualcomm chip has 85 TOPS

Wi-Fi 8 is just around the corner

Just when we were getting comfortable with Wi-Fi 7, Wi-Fi 8 is currently testing in the prototype stage. However, we’re not expecting any big performance boosts with Wi-Fi 8 (unlike what we saw with Wi-Fi 7). Instead, Wi-Fi 8 is focused on improving reliability, improving short- and mid-range performance, and leveraging technologies like AI to better connect devices (and keep them connected).

MediaTek has announced its new Filogic family of Wi-Fi 8 chips that we can expect to see in various consumer electronic devices and networking products, and Asus had its prototype Wi-Fi 8 routers on hand to show real-world performance gains over Wi-Fi 7.

• MediaTek unveils Filogic 8000 Wi-Fi 8 family at CES 2026 – new chipsets expected to arrive later this year

Everything else from the show floor

Here's all the rest of the hot tech that was either announced, or that we got to have some hands-on time with at the start of CES 2026.

• Gigabyte's Aorus RTX 5090 Infinity takes aim at Asus' ROG Matrix
• Thermaltake goes retro at CES 2026, shows off liquid cooler with CRT-themed display and 80s-style PC cases
• Anycubic Announces 'First Print Guarantee' for burgeoning 3D printers
• Lego unveils new smart brick with embedded computer inside
• Keychron launches wireless Q Ultra keyboard series with up to 660 hours of battery life with 8K polling
• Asus ROG G1000 gaming PC is covered in holograms
• Acer announces new Predator-series wireless gaming headset and mouse
• Acer brings trio of Predator and Nitro gaming monitors to CES
• HP’s HyperX Omen 34 gaming monitor delivers V-Stripe QD-OLED tech
• Samsung's new 'Odyssey 3D' 6K monitor takes center stage at CES 2026, features solid eye-tracking
• OWC reveals 192TB of blazing 6.6 GB/s storage inside desktop data powerhouse
• Asus ROG and Xreal partner to deliver gaming-optimized AR glasses
• CyberPowerPC's MA-01 comes with analog color control knobs, woven steel mesh, and slanted airflow

With Asus grabbing plenty of attention with its limited-edition ROG Matrix RTX 5090 last year, Gigabyte has now entered the spotlight with its own flagship response — the Aorus GeForce RTX 5090 Infinity. Featuring a rounded look that resembles the light cycles from Tron, it adopts a double flow-through design with two cooling fans on either end, and a smaller third fan hidden at the center behind a mesh.

According to the company, the GPU features their Windforce Hyperburst cooling system that includes a separated PCB design. This is said to enable ‘penetrating airflow’ on both sides of the card’s backplate wherein all the hot air is exhausted towards the backside of the card. The smaller third ‘overdrive’ fan helps by providing an additional airflow that only fires up during demanding workloads.

Gigabyte has also used a range of advanced thermal materials to improve heat dissipation. These include superconducting heat pipes, composite metal grease for the GPU die — a hybrid compound that blends traditional thermal paste with liquid metal properties — and server-grade thermal conductive gel for other internal components. Together, these changes are said to keep temperatures in check while maintaining stable performance under sustained loads.

Thanks to the cooling system, Gigabyte has managed to make the card smaller compared to most RTX 5090 models in the market measuring 330mm in length, 145mm in width, and 65mm in height. While that should make it easy to install in most PC towers including a selection of mATX cases, it is still larger than the RTX 5090 Founders Edition. Speaking of which, the Aorus GeForce RTX 5090 Infinity comes with a 16-pin 12V-2x6 power connector in a vertical orientation similar to the Founders Edition.

The primary cooling fans also come with Gigabyte’s RGB Halo customizable LED lighting, which has been a staple on Aorus GPUs. Other features include a die-casted metal backplate, dual-BIOS functionality, and a four year warranty. Pricing and availability has not been disclosed yet, but expect the flagship GPU to launch with a premium price tag possibly upwards of $3000.

Thermaltake is taking a trip back to the 80s and 90s with its latest portfolio of products announced at CES 2026. The company unveiled two new PC cases with an 80s-style aesthetic, as well as a new AIO liquid cooler series featuring an LCD display fitted into a CRT-shaped cover that sits above the pump housing.

The 80s-themed cases Thermaltake showed off come in two variants: the Retro 260 TG and the Retro 360 TG. Both have modernly-designed layouts on the inside, but feature the classic off-white color scheme computers in the 80s were known for. The cases also have front panels that mimic systems of the 80s — stamped with a (fake) floppy disk reader and two 5.25" bays, one of which has a (fake) CD_ROM reader. (To clarify: neither case has any real 5.25" bays.)

As the names suggest, the 260 variant is smaller and supports micro-ATX motherboards and up to 240mm AIO liquid coolers on the top and front of the case. The 360 model has a larger interior that can accommodate standard ATX motherboards and up to 360mm AIO liquid coolers on the top and right side of the case. Thermaltake is also offering an optional 6-inch LCD screen for these cases, which sits where the fake 5.25" bays are on the front panel.

Thermaltake also showed off a pair of new AIO liquid coolers with the same retro theme as its new cases. The Retro 240 and 360 AIOs are new 240mm and 360mm liquid coolers, which each sport a 4-inch 720p LCD display on the front that's shaped like a CRT display. Both models come with a single fan frame, which is illuminated with aRGB lighting.

Thermaltake also showed off a keyboard and a mouse with the same theme. The keyboard has a compact 75-percent layout, Gateron G-Pro 3.0 mechanical switches, double-shot PBT keycaps, and 2.4GHz wireless, Bluetooth, and wired connectivity. The mouse is geared toward gamers, and has a PixArt PAW3395 26,000 DPI sensor, Huano transparent blue shell switches, and a 400mAh battery, as well as 2.4GHz wireless, Bluetooth, and wired connectivity. It offers up to a 1,000 Hz polling rate and weighs just 1.83 ounces (52g).

Nostalgia is always in — component-maker SilverStone has found success with its retro-style PC cases (it announced several throughout 2025), and even OEMs such as Maingear have retro-styled product lines of pre-built machines. That said, ThermalTake's products don't have all the fully-functional retro features, while SilverStone's products are more functional — for example, the FLP03 has three functional external 5.25" drive bays.

At CES, Asus has announced the ROG NeoCore, its first Wi-Fi 8 router concept along with results from what it claims is the world’s first real-world Wi-Fi 8 throughput test. The concept device looks like a D20 di from Dungeons and Dragons, with an industrial design along with some ROG branding.

Aimed at enthusiasts and gamers seeking high-end networking performance, the final specifications are still under wraps. The company, however, has confirmed that it will be built on its AiMesh platform allowing multiple Asus routers to connect and work together as a mesh network. It will also come with the Asus Ai Network Engine, which aims to optimize throughput and latency by monitoring traffic and allocating bandwidth dynamically.

“WiFi 8 is not just about chasing peak speed—it’s about making every connection smarter and more reliable. It enables seamless collaboration between smart homes, AI assistants, and cloud services, delivering stable performance anytime, anywhere,” said Tenlong Deng, Corporate VP and General Manager of Wireless and Networking at Asus.

As for the results, there isn’t any change in the overall data speeds, however, Wi-Fi 8 will primarily boost throughput with up to 2x higher mid-range throughput, 2x wider IoT coverage, and up to 6x lower P99 latency via smarter multi-AP/multi-client operation. The company has not shared details about the testing environment or system configuration behind these results, though it is claimed that the performance comparisons were conducted against comparable Wi-Fi 7 setups.

Asus is positioning Wi-Fi 8 to solve four core connectivity challenges - signal degradation over distance, weak two-way communication for low-power devices, densely populated network scenarios, and bandwidth inefficiency. To address signal degradation, Wi-Fi 8 should offer more consistent speeds as you move further from the router, helping maintain stable connections whether the router is indoors or outdoors. It also improves two-way communication for low-power devices like smart lights and controllers, ensuring they are always connected.

In crowded environments such as apartment buildings, nearby routers can interfere with each other. Wi-Fi 8 is expected to use smarter spectrum coordination to reduce interference and maintain stable performance. Finally, Wi-Fi 8 improves how bandwidth is managed, using more efficient spectrum use and smarter scheduling to reduce congestion, lower latency, and improve overall speeds.

Even though the Wi-Fi 8 standard is not expected to be finalized until late in 2028, Asus plans to release its first line-up of Wi-Fi 8 home routers and mesh systems later this year.

Nvidia CEO Jensen Huang announced during his keynote speech at CES 2026 that Vera Rubin is officially in full production. Vera Rubin NVL72 is Nvidia's next-generation AI-focused datacenter platform, and will feature up to 5x improved performance over Blackwell-equivalent products in some workloads.

Nvidia already showed off Vera Rubin at GTC 2025, but what's new is the fact that the Vera Rubin platform is now in production. Vera Rubin NVL72 takes advantage of an 88 Olympus core ARM-based CPU with "Spatial Multi-Threading," 128GB of GDDR7 memory, and a Rubin-based GPU with 288GB of HBM4 memory.

According to Huang, Vera's new multi-threading functionality enables each thread to have the full throughput of a single core, giving the chip the same processing capacity as 176 cores.

Accompanying these new chips are Nvidia's BlueField-4 DPU for offloading storage and security away from the CPU and GPU, Spectrum-6 Photonics Ethernet, and Quantum-CX9 InfiniBand NICs capable of up to 1.6 Tb/s. All of these chips are interlinked through Nvidia's high-speed 18TB/s NVLink interconnect.

As for the GPU — Nvidia claims its Rubin GPU is capable of 50 PFLOPs of NVFP4 inference, and 35 PFLOPs of NVFP4 training performance. These figures represent 5x and 3.5x the performance of Blackwell — all while increasing the GPU's transistor count by just 1.6x.

Nvidia has also made huge changes to Vera Rubin NVL72's chassis compared to previous generations — it's made the entire system fanless, tubeless, and cableless; the entire system is now 100% cooled with liquid cooling. Jensen claimed that installation times have gone from two hours on Blackwell-based counterparts down to just five minutes for Vera Rubin.

For more details, check out our main coverage of the Vera Rubin NVL72 platform. The new system will come out in the second half of 2026.

At CES 2026, Nvidia's gaming updates lean heavily on AI. The company is making its DLSS suite of tech even better with the new version 4.5 of its upscaling, or "Super Resolution" model, as well as an enhanced version of its Multi Frame Generation model that can support even more aggressive multipliers than the current 4x version.

The DLSS upscaler kicked off a revolution in gaming performance when its second version arrived in 2020, and the company has held a lead in upscaling quality ever since. In addition, Nvidia's introduction of Tensor Cores on RTX 20-series GPUs dating all the way back to 2018 means that the most recent DLSS 4 Super Resolution model, powered by a transformer architecture, still works with those products. At CES 2026, DLSS 4.5 marks the arrival of the second generation of the transformer architecture model.

Nvidia says it's constantly training DLSS to address edge cases where it might produce undesirable artifacts. The first case the company highlighted is highly distracting "shimmering" or flicker on static surfaces. DLSS 4.5 claims to improve temporal stability in these situations, resulting in a more solid-looking image.

Second, ghostly trails or after-images can form behind objects close to the player, like guns or swords. This is another case that Nvidia has spent training time refining in DLSS 4.5, and we should see reduced instances of this ghosting with the new model.

Finally, the company says that DLSS 4.5 should deliver better anti-aliasing performance in some titles, such as Indiana Jones and the Great Circle.

We had a chance to go hands-on with DLSS 4.5 across several games before the company's announcement this evening, and our experience suggests that DLSS 4.5 will indeed be an impressive improvement over the existing transformer model. What we didn't expect is that DLSS 4.5 actually makes certain lighting and particle effects look richer and more natural, too. We'll go into these points in more depth with DLSS 4.5 once we're home from CES and back at the test bench.

Nvidia says that DLSS 4.5 is more computationally intense than past models, but the increased resource demand will be offset to some degree by support for accelerated FP8 processing in the Tensor Cores of RTX 40-series and RTX 50-series graphics cards.

The company says that RTX 20- and 30-series cards will still be able to run the new model despite their lack of Tensor Core FP8 acceleration, but it doesn't make any guarantees about performance on that older hardware.

Even so, if DLSS 4.5 incurs a large increase in a relatively small portion of overall frame time, it might still result in a relatively small hit to performance, as upscaling models have to be relatively lightweight by design in order to provide their performance-boosting magic.

DLSS 4 and its transformer architecture already incurred a small but measurable performance loss on Turing and Ampere, so it's likely we'll see a larger dip with the new model on those cards. We'll need to do some testing and see whether the image quality refinements this model offers are worth the performance tradeoff on older hardware.

As with past DLSS Super Resolution updates, gamers will be able to pick and choose among model versions using the Nvidia App on Windows to pick the one that offers the best balance of performance and image quality for a given application. DLSS 4.5 should be available right away as an Nvidia App override, so check for updates and see whether new models become available in those override options (you'll see choices for models L and M in the selection dropdown.)

DLSS Multi-Frame Generation is also getting a couple of major improvements this year, although they're coming later than the updated Super Resolution model.

First off, DLSS MFG will now have multipliers ranging up to a whopping 6x, versus the current model's 4x. In part, Nvidia says this is possible thanks to the improved image quality provided by the DLSS 4.5 super resolution model, as well as the smooth frame pacing measures it's built into its hardware and software stack. Giving the MFG model better input data should in theory result in better image quality across generated frames, although we'll have to see how well stretching a single native frame out into 4 or 5 generated frames works in practice.

It's also worth remembering, as we've long cautioned and proven through dedicated testing, that DLSS MFG isn't a way of making unplayable frame rates into playable ones. If you're trying to boost average frame rates of just 30 FPS or below with MFG, your experience will still feel as laggy and unresponsive as it's likely to at 30 FPS.

Instead, MFG 5x and 6x will likely work best as tools for making the most of the ultra-high-refresh-rate monitors coming out this year and into the future. Frame rate isn't a perfect proxy for input lag, but if you have a solid 90 FPS to work with, MFG 5x and 6x could make it practical to use 360Hz or 480Hz+ monitors in tandem with DLSS Performance or Ultra Performance without crushing graphical fidelity in exchange.

Nvidia is also introducing a new dynamic mode for MFG that will automatically adjust the frame gen multiplier to maintain a target frame rate that's set in the Nvidia App. As long as the mode switch doesn't cause stutter or other perceptible issues in exchange for maintaining that target frame rate, a dynamic MFG could be a handy feature, but we'll need to try it out to see how quickly and smoothly it can handle the transitions between multipliers.

DLSS MFG with dynamic mode and extended multipliers will arrive in spring 2026, and as with the original MFG, it'll be an RTX 50-series exclusive.

For all that, some elements of DLSS 4 remain in the oven even a year after this family of features was first announced. Nvidia says that its Reflex 2 lag reduction technology with Frame Warp reprojection, which could further improve perceived responsiveness, remains targeted for a future release.

Artificial intelligence is arguably the hottest technology around, so it's not surprising that AMD used part of its CES keynote to reveal new information about its upcoming Helios rack-scale solution for AI as well as its next-generation Instinct MI400-series GPUs for AI and HPC workloads. In addition, the company is rolling out platforms designed to wed next-generation AI and HPC accelerators with existing data centers.

Helios is AMD's first rack-scale system solution for high-performance computing deployments based on AMD's Zen 6 EPYC 'Venice' CPU. It packs 72 Instinct MI455X-series accelerators with 31 TB of HBM4 memory in total with aggregate memory bandwidth of 1.4 PB/s, and it's meant to deliver up to 2.9 FP4 exaFLOPS for AI inference and 1.4 FP8 exaFLOPS for AI training. Helios has formidable power consumption and cooling requirements, so it is meant to be installed into modern AI data centers with sufficient supporting infrastructure.

Beyond the MI455X, AMD's broader Instinct MI400X family of accelerators will feature compute chiplets produced on TSMC's N2 (2nm-class) fabrication process, making them the first GPUs to use this manufacturing technology. Also, for the first time the Instinct MI400X family will be split across different subsets of the CDNA 5 architecture.

The newly disclosed MI440X and MI455X are set to be optimized for low-precision workloads, such as FP4, FP8, and BF16. The previously disclosed MI430X targets both sovereign AI and HPC, thus it fully supports FP32 and FP64 technical computing and traditional supercomputing tasks. By tailoring each processor to a specific precision envelope, AMD can eliminate redundant execution logic and therefore improve silicon efficiency in terms of power and costs.

The MI440X powers AMD's new Enterprise AI platform, which is not a rack-scale solution but a standard rack-mounted server with one EPYC 'Venice' CPU and eight MI440X GPUs.

The company positions this system as an on-premises platform aimed at enterprise AI deployments that is designed to handle training, fine-tuning, and inference workloads while maintaining drop-in compatibility with existing data-center infrastructure in terms of power and cooling and without any architectural changes.

Furthermore, the company will offer a sovereign AI and HPC platform based on Epyc 'Venice-X' processors with additional cache and extra single-thread performance as well as Instinct MI430X accelerators that can process both low-precision AI data as well as high-precision HPC workloads.

The Instinct MI430X, MI440X, and MI455X accelerators are expected to feature Infinity Fabric alongside UALink for scale-up connectivity, thus making them the first accelerators to support the new interconnect. However, practical UALink adoption will depend on ecosystem partners such as Astera Labs, Auradine, Enfabrica, and Xconn.

If these companies deliver UALink switching silicon in the second half of 2026, then we are going to see Helios machines interconnected using UALink. In the absence of such switches, UALink-based systems will use UALink-over-Ethernet (which is not exactly a way UALink was meant to be used) or stick to traditional mesh or torus configurations rather than large-scale fabrics.

As for scale-out connectivity, AMD plans to offer its Helios platform with Ultra Ethernet. Unlike UALink, Ultra Ethernet can rely on existing network adapters, such as AMD’s Pensando Pollara 400G and the forthcoming Pensando Vulcano 800G cards that can enable advanced connectivity in data centers that can already use the latest technology.