Ahead of Computex next week, Nvidia's social media accounts have begun promising "a new era of PC," along with the latitude and longitude of the Taipei Music Center, where CEO Jensen Huang will present his keynote for the event as part of GTC Taipei 2026.

While we don't have any idea exactly what's coming, it's intriguing to see who else is joining in on the game. The Windows X/Twitter account has shared the exact same message as Nvidia's, suggesting that we could see the long-rumored N1X laptop platform make its debut at Computex - and that it could be running Windows on Arm.

For background, N1X has long been rumored to be the mobile variant of the GB10 Superchip at the heart of the DGX Spark mini-PC, which boasts an RTX 5070-class GPU paired with 128GB of LPDDR5X memory and a powerful Mediatek-designed 20-core Arm CPU complex.

But the DGX Spark is an Ubuntu Linux-powered AI developer sandbox, not a jack-of-all-trades PC that can seamlessly run Windows apps, as the current crop of Windows on Arm platforms can. If Microsoft is putting its weight behind N1X, that could broaden the appeal of the platform for a more general computing audience by bringing the entire Windows app ecosystem to the platform.

Supporting N1X would also bring a powerful, advanced unified-memory-architecture AI computing platform into the Windows camp. None of Microsoft's other Windows on Arm partners have produced anything nearly as ambitious or powerful an AI foundation as the GB10 Superchip, so N1X laptops could be a major boost for the company's AI ambitions on Windows. Having that class of raw compute at its disposal could certainly inspire Microsoft to create new types of first-party local AI experiences that simply haven't been possible from the current crop of Copilot+ PCs and their relatively limited AI grunt.

But given what we know about GB10 already, the appeal of this type of system could be narrow at first. Because they share the same pool of LPDDR5X memory, the GB10 GPU enjoys just 273 GB/s of raw bandwidth, far less than that offered by more traditional laptops with dedicated GPUs that have their own pools of GDDR memory.

In our own experience, we've found that you can certainly game on GB10, but it's not the platform's strongest suit. So unless there's a major change in the platform's architecture or resources waiting in the wings, N1X PCs will likely need to deliver a new type of experience with their AI potential that's missing from current systems and platform architectures.

And N1X PCs will almost certainly be expensive amid the current silicon crunch. GB10 boxes are all selling for around $5000 by our reckoning, and that's partially because they include an exotic NIC that almost certainly won't make its way into any potential laptops powered by this platform. But massive pools of RAM and large SSDs don't come cheap right now, either, so we're still likely to be looking at pricey partner systems.

A broader product stack than the 128GB GB10 with smaller memory options and lower CPU and GPU resource counts could help make these systems relatively more affordable while still keeping them plenty powerful for local AI.

In short, there's still plenty we don't know about how an N1X-powered AI PC will look, but the fact that Nvidia and Microsoft could be teaming up to make it a Windows on Arm platform is a big deal in itself. We'll be on the ground at Computex 2026 very soon, and we'll report back with details on this potential development as we learn more.

Market research and tech analyst firm IDC reckons that global PC sales will reach 274 million units in 2025. Thus, we should be looking at a 4.1% uplift in PC sales compared to 2024, indicates IDC. This appears to be good news for the PC industry, and factors including a race to beat U.S. tariffs and Windows 11 are behind the upward revision, say the firm’s analysts.

We last wrote about IDC global PC shipment forecasts back in February, when the impact of U.S. tariffs and the cool reception to AI PCs were expected to be headwinds affecting the industry. At that time, IDC expected global PC sales to be up 3.7% (at 273 million units) compared to 2024 (262.7 million units were shipped last year).

IDC has adjusted its previous estimates upward after it observed that PC makers shipped larger-than-usual volumes to the U.S. ahead of tariff deadlines. “The 90 day pause and tariffs exemption applied to personal computers, combined with a definite level of uncertainty on what will happen after the 90 day pause, is motivating PC manufacturers to seize the moment and ship larger than anticipated volumes in the US,” explained Jean Philippe Bouchard, research VP with IDC's Worldwide PC Trackers.

That artificial boost to U.S. shipments might easily be cancelled out by an oversupply of PCs in subsequent months, with the PC maker's actions accentuating the ebb and flow of supply and demand.

Other negative sentiment regarding expected PC shipment volumes stems from things like worsening macroeconomic conditions, upward pressure on pricing, and degraded consumer sentiment, noted Bouchard.

Clock ticking on Windows 10

DeepComputing has developed a reputation as a pioneer of small form factor RISC-V PCs, being the first company to bring a RISC-V laptop to the market in 2022. The company is now accepting pre-orders for its newest product, a mini AI PC that can replace the mainboard of the Framework Laptop 13.

The new device comes with two names: either the "DC-ROMA RISC-V AI PC" or the "DC-ROMA RISC-V Mainboard II for Framework Laptop 13" according to its listing page. The mainboard is one of many members of DeepComputing's DC-ROMA line, a family of SFF/laptop PCs all built on the RISC-V architecture set. It is also DeepComputing's second replacement mainboard for the Framework Laptop 13, with both the RISC-V AI PC and its predecessor functioning either inside the Framework or on its own as a stand-alone unit.

The specs of the DC-ROMA RISC-V AI PC are slightly mysterious. The device will be based on the "world's first RISC-V chiplet dual-die connected AI SoC." The mystery SoC will hold a 64-bit 8-core RISC-V CPU running at up to 2 GHz, along with an NPU, GPU, and VPU to cover all of the AI bases. VPUs are an emerging type of microprocessor devoted to vision-based machine learning tasks, distinct from GPUs as they have no rasterization or video encoding abilities.

The device will reportedly provide up to 50 TOPS of AI performance, with 40 TOPS coming just from the onboard NPU. The 3D-capable GPU onboard supports up to 8K@50Hz encoding, with the VPU tagging along for those vision-dependent AI training tasks. The mini PC supports up to 64GB of LPDDR5 RAM and NVMe SSDs, and we'd expect to see a microSD slot for the OS, as is common on DC-ROMA devices.

The only SoC that advertizes the same specs (and makes the same bold "world's first" claims) is Eswin Computing's EIC7702X, which is two Eswin EIC7700Xs stuck together. The dual-CPUs, two 4-core SiFive Performance P550s, are rated for up to 1.8 GHz each, paired with GPUs based on Imagination IP. The EIC7700X, one-half of the EIC7702X we expect to find in the DC-ROMA PC, can be seen on SiFive's HiFive Premier P550 developer board, which we reviewed earlier this year and gave 3.5 out of 5 stars.

RISC-V is an open-source instruction set architecture, similar to x86 or Arm. RISC-V development has been very active in recent years, especially in China, where its open-source nature provides a path forward for Chinese engineers stymied by an ongoing trade war with the United States. The DC-ROMA RISC-V AI PC, while not likely to blow anyone away with its speeds, does provide an extremely novel promise. Small form factor AI performance on the RISC-V platform is a very narrow intersection of three already niche worlds, but it may be a desirable buy for those happy few.

The RISC-V Mainboard II for Framework Laptop 13 (aka the DC-ROMA RISC-V AI PC) is available now for pre-order from DeepComputing. The device will start at $300 on its release in Q3 2025, with a $9.90 deposit required today. If it follows the pattern of the first DC-ROMA Framework mainboard, it will likely be sold as a standalone board, inside a mini-PC enclosure, or pre-assembled inside a Framework Laptop 13. We're cautiously optimistic about the future of this release, and as always stand ready to applaud any and all headway made towards a brighter future for RISC-V computing.

Lenovo is set to unveil a series of new devices at the upcoming Mobile World Congress (MWC) 2025, with the ThinkBook Flip AI PC emerging as a standout. Renowned leaker Evan Blass has shared renders of this upcoming laptop, showcasing a unique rollable and foldable OLED display that reimagines traditional laptop design.

The ThinkBook Flip features a dual-hinge mechanism, allowing the screen to fold outward and effectively double its vertical size when fully rolled out. This design should provide users with an expanded workspace, ideal for multitasking or viewing extensive content without relying on external monitors. When the additional screen space isn't needed, the top half can be folded back, converting the device into a conventional laptop form factor. This flexibility also enables the upper portion to function as a secondary display, facilitating content sharing during presentations or collaborative sessions.

Unlike previous foldable devices such as the ThinkPad X1 Fold, which was essentially a foldable tablet with a Bluetooth keyboard, the ThinkBook Flip comes with a built-in keyboard. The device's minimalist design includes flat edges which should include essential ports like USB Type-C and USB Type-A, although it is not very clear from the leaked renders. Additionally, there is a webcam positioned within a notch at the top of the screen.

While specific hardware specifications remain undisclosed, it is expected that the ThinkBook Flip will be equipped with Intel's forthcoming Arrow Lake-H or Lunar Lake processor options. There is also speculation about potential AMD variants, possibly integrating Strix Point-based Ryzen AI 7 series chips. Essentially, it could end up being a high-performance device catering to both professional and creative users.

Lenovo has been experimenting with foldable and rollable display technologies for some time. At CES (Consumer Electronic Show) last month, the company unveiled the ThinkBook Plus Gen 6, featuring a rollable display that could expand vertically with just the wave of a hand.

Notably, the leaked ThinkBook Flip AI PC aligns with previous reports, which highlighted Lenovo's plans to introduce a range of new devices at MWC 2025. Among the anticipated releases are nine new laptops and a hybrid 2D/3D 34-inch curved monitor.

The monitor, bearing Lenovo's Legion gaming brand, is expected to target professionals and gamers alike, offering a larger alternative to the existing ThinkVision 27 3D. The leaked laptops include models such as the IdeaPad Slim 3x, featuring a 15-inch screen and likely powered by a Snapdragon X chip, and the ThinkPad T14s 2-in-1, marking the first 2-in-1 form factor in Lenovo's flagship ThinkPad T-series.

Other models, like the ThinkBook 14 2-in-1 Gen 5, ThinkBook 16p Gen 5, ThinkPad T14 Gen 6, ThinkPad T16 Gen 4, ThinkPad X13 Gen 6, and the Yoga Pro 9i Aura Edition, are expected to offer incremental upgrades with newer chips and components. Notably, the ThinkPad T14s Gen 6, a 14-inch laptop powered by a Snapdragon X chip.

According to a prominent analyst, Micron’s miss on Q3 earnings and guidance for the second quarter of next year might signal that the AI PC and smartphone supercycle isn’t happening. Most of the company's woes stem from a weaker market than expected for memory products for PCs and smartphones, and multiple reports from other market analysis firms have pointed out that the AI PC 'revolution' simply isn't happening. At least not yet.

In its latest financial report, the American memory manufacturer Micron reported Q3 revenue of $8.709 billion, somewhat below the $8.721 figure the market anticipated. Even worse, Micron's guidance for the second quarter of 2025 was $7.9 billion instead of the $8.98 billion that Wall Street expected. At the time of writing, its stock is down over 16%.

Calling the miss and forecast a “big whiff,” semiconductor analyst Daniel Newman said in an X post that it’s not the “beginning of the end for the AI trade” and companies like Nvidia, which has grown to be one of the world’s biggest companies since AI chips took off earlier this year.

Although high-bandwidth memory (HBM) is set to become a big market for Micron, expected to grow from $16 billion in total addressable market value this year to $100 billion by 2030, Micron’s primary source of revenue today is making memory chips for PCs and smartphones.

“However, the core business is contracting as PC and smartphone shipments lag AND Micron is dealing with customer inventory that is selling off slowly leading to even lower booking/sell-through in this and the next quarter, “ Newman said. “Bad news is the AI PC and AI smartphone ‘supercycle’ has more or less been a bust.”

In 2024 and 2023, hopes were high that the PC industry would be supercharged by demand for AI PCs thanks to their new, AI-powered features. However, it seems that things didn’t shake out that way. One report by IDC Research in September said AI isn’t driving demand for AI PCs. Instead, it's driven by the general desire to upgrade, as new chips with AI hardware also feature faster CPU and GPU cores.

Trendforce published a report last month that reached a similar conclusion: Customers were not interested in AI PCs. The market analysis firm says that users upgrading to Windows 11 PCs from older Windows 10 devices will drive more sales than AI in the new year.

Qualcomm’s apparent struggles with its new Snapdragon X chips for Copilot+ laptops indicate that demand for AI PCs is probably not great. In Q3 of this year, Qualcomm captured a mere 0.8% of the PC market, with only 720,000 units sold. Intel’s interim co-CEO Michell Johnston Holthaus claimed last week that return rates for Snapdragon X laptops were relatively high, a charge Qualcomm denies.

The Arm chip company seems to be looking forward to upcoming Snapdragon X PCs that will bring the cost-to-entry down to $700 without reducing performance on the NPU, a significant source of AI prowess on Qualcomm’s PC chip. The cheapest Snapdragon X laptops cost $1000, so a $300 reduction is substantial.

However, if demand for AI PCs were great, it would seem hasty for Qualcomm to move on to lower price brackets so soon, mainly if these upcoming PCs feature the same top-end NPU present in the Snapdragon X Elite and Plus. Focusing on more budget-friendly parts of the market makes sense if there’s no more appetite for AI PCs at the higher end, but it's also clear that Qualcomm is struggling to build scale.

The AI PC seems to have a major issue because there’s not much use right now for running AI locally on a PC. AI software that runs directly on an end user’s device is more in the enthusiasts' wheelhouse. Meanwhile, popular services like ChatGPT run in the cloud and don’t require a brand-new AI PC.

As the AI PC becomes the default PC, having AI hardware will eventually cease to be an essential selling point, much like how multicore CPUs, integrated graphics, and solid-state drives have become standard for PCs. Consumers probably won’t pay extra for an AI PC, but they probably won’t buy a non-AI PC either since that implies it’s using an old chip that predates the AI revolution. In other words, AI becomes table stakes. 

However, the lack of demand for AI PCs and smartphones isn’t necessarily a sign that the whole AI industry is about to implode. “The good news is HBM is selling well and growing fast meaning the AI chip demand is still in tact,” Newman argues. “NVIDIA, Broadcom, AMD, Marvell, etc. Panic not. HBM is growing fast.”

Of course, it’s not great news that consumers aren’t enthralled by running AI locally on PCs and smartphones. However, there’s still a big market for running AI in the cloud, especially since cellular devices make it easy to connect to the internet. Dropping demand for HBM would be a much more significant indicator of something going wrong since HBM is critical for AI chips made for the data center.

However, if AI turns out to be a bubble and pops, this might be the first warning sign: For now, there is no explosive demand for AI on running on end-user devices.

Despite Qualcomm's ambitions and high hopes to control over 50% of the Windows market in the next five years, recent statistics show that its first-gen X Elite offerings are failing to gain consumer traction. TechRadar reports that of all PCs sold in Q3 24, Qualcomm's chips powered 720,000 units - capturing just 0.8% of the entire market. In other words, one out of every 125 devices sold in the last quarter featured an X Elite processor, which pales in contrast to Intel, AMD, and even Apple on the Arm front.

Undoubtedly, AI PCs are on the rise, primarily because virtually every new processor comes equipped with a Copilot PC+ compliant NPU. Per the report, 13.3 million AI PCs were shipped in Q3 this year - a 20% increase compared to last quarter. For the uninitiated, this stat refers to PCs featuring an NPU-equipped processor, like Intel's Meteor Lake and Lunar Lake offerings, AMD's Phoenix, Hawk Point, and Strix Point series, and Qualcomm's X Elite SoCs.

Despite a 180% surge in X Elite sales from Q2 24 to Q3, Qualcomm's chips power less than 1.5% of Windows PCs. This lackluster reception is primarily due to unoptimized software or the lack thereof. For example, Microsoft's tardy release of an official Windows 11 ISO for Arm devices left early X Elite adopters in the cold for months. Additionally, Linux support isn't as robust as existing x86 alternatives; longer battery life isn't sufficient or incentivizing enough to drive adoption.

Qualcomm is preparing to launch budget X Elite laptops next year to somewhat alleviate this situation. On that note, the fruits of Qualcomm's acquisition of Nuvia are materializing—both in terms of high-stakes litigation and powerful mobile SoCs, namely, the Snapdragon 8 Elite designed using the Oryon 2 architecture. Rumors suggest that the X Elite's successor will jump directly to Oryon 3, aiming for a 2025 reveal/launch.

In any case, compatibility will always be an issue with these Arm SoCs (in Windows) - be it any chip maker. Apple's sheer market share and dominance compelled developers to write and optimize software for its silicon. Collaborating with Microsoft is vital for Qualcomm in this regard, however, it'd still have to compete with other x86 players in the Windows space; Intel and AMD.

If you own an "AI PC" from Intel or AMD, that means you have an onboard Neural Processing Unit (NPU) alongside your typical CPU cores and graphics. And if you weren't aware, it seems — at least for Intel NPUs — you'll have yet another driver to potentially update on a regular basis. Even if you don't bother with Windows' AI PC features, these driver updates include security fixes alongside fixing issues or adding features, so it's probably a good idea to be diligent with these updates.

If you're using an Intel NPU on a Windows 10 PC rather than a Windows 11 PC, Intel also recommends that you outright disable the NPU within the BIOS rather than keeping it or its drivers enabled. What happens if you don't do this isn't clear (probably nothing for most situations), but leaving a potentially unusable NPU enabled could serve as a vector for attack or contribute to system instability.

Anyway, the latest Intel NPU driver was released on October 11. Intel NPU driver version 32.0.100.3053 adds support for OpenVINO 2024.4 and Arrow Lake CPUs. These are minor but expected changes ahead of the full launch of Arrow Lake CPUs that's scheduled for October 24. We'll have a full review of some of the processors then, where we'll have independent testing of Intel's next-gen desktop CPUs, the new socket, and the new architecture.

Arrow Lake and its Intel Core Ultra design should completely do away with the hardware issues that infamously plagued 13th and 14th Gen Intel 'Raptor Lake' CPUs, so at least we have that to look forward to. Of course, being a brand new platform and architecture, there plenty of potential for other problems to emerge over time.

But getting back to the NPU drivers. Fixed Issues include Microsoft Copilot performance and accuracy improvements, along with aligning the "UD40 Driver version number." There are also fixes specifically for Lenovo View Video enhancer,  a krisp model loading issue, and enabling and optimizing face/eye tracking and 2D/3D conversion functionalities. Lenovo View Video Enhancer, for those unfamiliar, is a component of the Lenovo View device driver (for Lenovo devices but especially AI PC-enabled ThinkPads) that provides better video call features, including quality adjustments, background concealing, and even little health warnings.

Intel has already established a regular cadence for NPU driver updates, looking at the history of releases. There's basically a new driver revision every month, sometimes enabling new devices (like the Keem Bay video processing unit), more often just addressing a few bugs. But we'd expect NPU driver updates from Intel's competitors to be just as important in the coming years, particularly if the AI PC features and Copilot+ really take off.

According to reports from leakers on the Chinese social media site Weibo, Mediatek and Nvidia are collaborating on a 3nm AI CPU. IT Home shared a report from user "Mobile Chip Expert" today, claiming that the CPU is entering the tape-out phase of production this month, with mass production on track for late 2025. 

We've reported on rumors of potential MediaTek and Nvidia collaborations for a few months now, so as usual, take this new report with a serving of salt. Previous rumors of a PC chip collaboration were the companies working on an AI card processor. However, today's reports are connected to rumors of an Nvidia/MediaTek AI processor that started in May.

The MediaTek AI processor is expected to be paired with an Nvidia GPU. The post also names Lenovo, Dell, HP, and Asus as prospective customers looking to adopt the processor in OEM hardware. The chip has also been linked to a rumored $300 price tag. 

Today's report curiously only refers to the rumored chip as an "AI PC" CPU, surprising as both previous leaks about the collaboration and MediaTek's stable of products would point to a mobile chip or SoC being more likely. MediaTek specializes in mobile products and typically employs AMD as a hardware partner. AMD uses MediaTek’s Wi-Fi 6E solutions in its mobile platforms, and MediaTek and AMD have also been connected through rumors of a 5G notebook for some time. 

It would not be surprising if the MediaTek/Nvidia chip transitioned to only an AI PC solution and moved away from being a mobile phone offering. After the big waves and lessons learned from the Qualcomm Snapdragon X Elite launch this year, there is room for another Windows-on-Arm competitor from a primarily mobile company like MediaTek. One of the most prominent problems reviewers had with the performance of Snapdragon X-powered computers was their lack of graphical strength, which Nvidia can quickly solve. Nvidia's name recognition alone will do much to boost the MediaTek chip, whether it is Nvidia's return to mobile graphics or a different venture into laptop-form-factor AI PCs.

MediaTek and Nvidia's only officially announced collaboration is the Dimensity Auto Cockpit platform, an SoC for cars that license Nvidia's graphics IP and Nvidia Drive OS. The chip seeks to be a front-runner in automotive chips by offering full support for infotainment platforms, including AAA gaming with RTX graphics and safety features such as HDR multi-camera support. So far, no US-based cars seem to have adopted the platform.

Processors with AI capabilities have recently become popular, especially with the launch of the Qualcomm Snapdragon X, AMD Ryzen AI 300, and the Intel Core Ultra 200V processors, which feature NPUs that deliver 40 TOPS and up. However, IDC Research says that the adoption of these AI PCs, especially the Windows Copilot+ PCs, is largely driven by the need to purchase new computers or upgrade existing laptops—not because of their AI capabilities.

“While AI has been a buzzword of late, it has yet to be a purchase driver among PC buyers,” says Worldwide Mobile Device Trackers research manager Jitesh Ubrani. “Businesses certainly recognize the importance of AI though many struggle to see the immediate use case and instead are opting for AI PCs as a means to futureproofing.” He also said that most retail customers do not appreciate the advantages that the on-chip Neural Processing Unit (NPU) brings.

Laptop manufacturers still need to spell out the advantages of an NPU so that buyers can appreciate the capabilities of an AI PC for day-to-day use. Despite Microsoft’s aggressive marketing of its Copilot+ PCs, short-term adoption remains somewhat modest. That said, since all new AMD, Intel, and Qualcomm processors have built-in NPUs, we expect AI PC sales to grow exponentially in the coming years.

Tablet sales growth outpacing traditional computers

Although the sales of AI PC are expected to grow slowly but surely, computers in general are expected to remain flat in 2024, with growth slowing down to just 0.3%. This number is pulled down by the economic challenges experienced by China, as the global market, excluding the East Asian power, is expected to grow by 2.8%.

However, this slowdown in PC sales is offset by the growth of tablet sales. The latter is expected to hit a 7.2% increase this year, largely driven by the refresh cycle as companies release new models with better performance numbers and larger screens. What’s interesting is that this growth isn’t driven by the traditional tablet powerhouse, the iPad, but by the aggressive moves of Android tablet makers.

“The first half of 2024 showed how aggressive some of the Android vendors have been in competing in the tablet market, maximizing on the opportunities for sustainable demand for tablet devices in emerging regions such as Asia/Pacific (excluding Japan), the People’s Republic of China (PRC), and Central and Eastern Europe,” said IDC Mobility and Consumer Device Tracker senior research analyst Anuroopa Nataraj. “Adoption rates are expected to climb in the short term as the market continues to evolve with newer tablet models with upgraded chips, larger screens, and connectivity. And the device replacement cycle is only going to push the growth further.”

Nevertheless, the long-term growth of tablets is expected to slow to near 0% by 2028, especially as these devices still do not offer the whole performance that most users expect from laptops. Furthermore, the increasing efficiency leading to extended battery life and thinner form factors delivered by the latest generation of x86 and Arm processors, as well as the inclusion of touchscreen interfaces on many new PC models, might make purchasing a tablet a moot point for many buyers in the future.

Lenovo’s been at the forefront of convertible laptop designs since its first Yoga laptop shipped in 2012. But here ahead of IFA in Berlin, its latest Auto Twist AI PC concept attempts to take things into sci-fi territory, with a motorized “Dual Degree of Freedom” hinge that can track your movements to make sure you always have “optimum display visibility.”

Of course, this means your webcam will also keep you in frame for meetings, presentations, or video calls with friends and family. That could be convenient for those who don’t like to sit or stand in one place for calls, and it’s certainly striking to see a laptop screen follow you around the room. But it’s also hard not to be reminded of 2018’s Facebook Portal, which similarly had a screen and camera that followed users around the room.

There is also no denying that the screen wobbles noticeably as it moves, despite a Lenovo representative telling me the company had done a lot of work to reduce it. Still, it’s impressive that the company was able to fit motors into the base of the screen without giving it a huge bump. There is a slight bulge on the bottom, but it’s barely noticeable — especially when the screen is moving on its own.

Of course, the Auto Twist AI PC concept is a full-blown laptop, and its motorized hinge can do a few other interesting things. At Lenovo’s event, saw the laptop responding to voice commands, like “tablet mode,” “laptop mode,” and “close laptop.” At one point, the screen did get stuck trying to convert to tablet mode, with the lid seeming to get stuck up against the keys. But after a quick reorientation and re-launching of the software, I didn’t see any other issues in terms of movement. The lid is designed to automatically close itself when you step away from the laptop.

Hopefully that last bit will be adjustable in terms of time and sensitivity. No one wants their laptop to close when they turn away for a few seconds to talk to a colleague. And if you’re working from home, you probably don’t want to have to open your laptop back up every time you get up to get a drink or a snack.

If you’re the slouching type, the Twist AI PC is also ready to nag you about it. Lenovo says the laptop has “wellness enhancements” designed to promote ergonomic computing and reduce strain. Honestly, most of us could use some help there. But I’m not sure I want my laptop automatically reminding me how tired I am by adjusting the screen angle when I’m hunched over at the end of the day (fine, also sometimes at the beginning of the day), or sending me notifications to sit up straight because the camera can tell that I’m slouching.

This is, of course, a concept laptop. So you probably don’t need to worry about your next company-issued ThinkBook sending TPS reports to your corporate overlords about your bad posture or how many times you paced across the meeting room during your last presentation about quarterly financials. 

But some of these features may eventually filter down to buyable devices. I don’t hate the idea of a laptop that can open and close on its own – so long as the motors and mechanism are robust enough that I don’t have to worry about them breaking. The Auto Twist AI PC did look a bit fragile, but it’s still a concept. I’m sure if Lenovo ever releases a device like this with a motorized hinge, it will iron a few more kinks out before its launch date. 

The noise around AI in general and AI PCs in particular is quite significant these days, as everyone is trying to take advantage of the buzz word and present their products as AI-capable. Technically, 14% of PCs shipped globally in the second quarter contained a neural processing unit (NPU), formally making them AI PCs, reports Canalys.

The latest CPUs from AMD, Apple, Intel, and Qualcomm contain NPUs. As a result, 8.8 million AI-capable PCs were shipped in Q2 2024, comprising 14% of total PC shipments for the quarter, reports Canalys. As AMD, Intel, and Qualcomm ramp up production of their latest processors with NPUs, analysts believe shipments of AI-capable PCs are poised for rapid growth in the latter half of 2024 and into 2025.  

In the Windows segment, AI-capable PC shipments surged 127% sequentially in the second quarter. Lenovo, the world's largest PC maker, introduced its Snapdragon X Elite-based PCs, including the Yoga Slim 7x and ThinkPad T14s, boosting its AI-capable share to about 6% of total Windows PC shipments, reflecting a 228% growth. HP followed closely with an 8% share, launching the Omnibook X 14 and EliteBook Ultra G1 alongside its Core Ultra devices. With just under 7% share, Dell launched Copilot+ PCs across several lines, though availability was staggered.

"A key benefit from AI-capable PCs that has materialized for PC OEMs is the growth boost within their premium offerings," said Ishan Dutt, Principal Analyst at Canalys.

He continued, "Windows PC shipments in the above $800 range grew 9% sequentially in Q2 2024, with AI-capable PC shipments in those price bands up 126%. As the range of features from first- and third-party applications that leverage the NPU increase and the benefits to performance and efficiency become clearer, the value proposition for AI-capable PCs shall remain strong. This is especially important over the next 12 months as a significant portion of the installed base will be refreshed as part of the ongoing Windows upgrade cycle."

Meanwhile, only Qualcomm's Snapdragon X Elite processors comply with the requirements of Microsoft's Copilot+ PCs that need a 45 TOPS NPU. AMD and Intel must ship their Copilot+ platforms, the AMD Ryzen AI 300-series, and the Intel Core Ultra 2-series 'Lunar Lake.' Copilot+ requirements will likely be the standard for AI PCs, so contemporary AMD and Intel processors have an NPU; software developers will likely target Microsoft requirements, so it remains to be seen how many AI PCs shipped today will be AI-capable in a year or two.

Apple currently leads the AI-capable PC market, as all of its M-series chips have packed an NPU since 2020, and by now, there are tens of millions of Macs with a neural processing unit capable of launching AI workloads. However, while Apple's own programs use the company's NPUs, a limited number of third-party programs can take advantage of those units. Canalys believes that the introduction of Apple Intelligence, now in beta in the U.S., is set to enhance AI functionalities for Mac users, potentially scaling quickly once fully launched, given its compatibility with most existing Mac devices.

Analysts from Canalys believe the AI-capable PC market is on track to ship around 44 million units in 2024 and 103 million in 2025. The combination of new product releases, wider price range availability, and growing demand for AI-enhanced features is expected to drive this growth further, solidifying AI's role in future PCs.

With AI PC hype at a fever pitch, chip vendors Intel, AMD, and Qualcomm are engaged in a fierce dogfight. Meanwhile, relative newcomer Cixin Technology hopes to offer China a domestic alternative for AI performance with the Cixin P1, a brand-new SoC for AI PCs. 

The Cixin P1, announced yesterday in a press conference, is an Arm processor built on the 6nm process, with a 12-core CPU at 3.2 GHz, 10-core integrated graphics, and an NPU capable of 30 TOPS — which pushes the entire package to a total of 45 TOPS. The CPU is built on the Armv9.2-A architecture and runs a configuration of 8 performance cores and 4 efficiency cores. Little info about the chip has made it out of China, so we're relying heavily on coverage from the Electronic Engineering Times China. 

Cixin declined to share specifics on the CPU, GPU, or NPU used in the P1 SOC. Based on the specs it did share, and ARM's presence as a guest at the launch event, it is highly likely that Cixin is licensing Arm's CPU and GPU designs. The Cortex-A or Cortex-X series are RISC processors designed by Arm; the onboard CPU is likely a new variant of these. 

The GPU is an Arm Immortalis GPU. Speculation suggests that the specific GPU used is the Immortalis-G720 — Arm's 5th-gen mobile GPU, designed for mobile gaming and AI performance. The NPU likely comes from a partner other than Arm, since Arm's largest licensable NPU only scales up to 10 TOPS. 

The Cixin P1 has a modern slate of features, including support for LPDDR5-6400 memory and 16x PCIe 4.0 expansion for dedicated GPUs/AI accelerators, and it can output at up to 4K 120FPS. The P1 was built to be a chameleon of a processor, developed under a strategy of "one chip for many uses". This applies to its firmware, which is uniquely compatible with Windows, Android, Kirin, Tongxin, and a host of other operating systems, foreign and domestic. 

"One chip for many uses" was also the guiding light in determining Cixin P1's use cases. "Notebooks, mini PCs, all-in-one computers, desktops, home entertainment consoles, enterprise edge hosts, etc.," were cited as potential uses for the P1, and launch attendees saw the processor power notebooks and desktops. They also saw the many uses in person, as the company had benchmarks like 3DMark06 and the popular video game Genshin Impact running on the floor, along with other productivity and AI LLM demos. No numbers, such as benchmark results or Genshin's frames per second, were shown — though reporters qualitatively describe demos as running smoothly. 

The Cixin P1 is a very unique processor. While it does not fulfill Beijing's goals of being a fully home-grown processor because it uses Arm CPU and GPU designs, the chip is specifically designed for the needs of the Chinese market and stands ready to enjoy some level of success. It also sadly does not meet Microsoft's arbitrary "AI PC" requirements. The P1's NPU on its own is only 30 TOPS — shy of the 40 TOPS target — though it's likely that Cixin could not care less about Windows Copilot+ certification.

Cixin is a very young company, first established in 2021, that has been growing thanks to investments from 15-20 public and private investment partners listed on its website. This freshness to the Chinese tech scene may limit widespread adoption early on, so we'll have to wait and see if Cixin manages to succeed in the huge, hungry tech market seeking separation from U.S. interference.

Artificial Intelligence (AI) is the tech term of the moment, and everyone in the PC space wants a piece of it. Every major computer company, from the ones that make the best ultrabooks and laptops to those that put together the best gaming PCs and desktops, as well as most component manufacturers, want to say they're offering an AI PC.

But what is an AI PC, exactly? What does it do differently from the computer you already have? Do you need one at all?

Here's how to cut through the noise to learn what an AI PC actually is:

The many definitions of an AI PC

Since every PC company and component manufacturer wants to say it is offering an AI PC, there are a ton of different definitions out there. Much of the press and the industry has seemingly coalesced around Microsoft's definition, which Intel shared at an AI PC developer program showing off its Core Ultra "Meteor Lake" processors. That definition had three requirements, suggesting that an AI PC:

• Includes the latest NPU (neural processing unit), CPU, and GPU
• Comes with Microsoft Copilot
• Has the Copilot key on its keyboard

This definition did rule out some existing PCs that had AMD and Intel's NPU and Copilot, but hadn't included the Copilot key. Most major laptop releases since then have included that key. If you take the Copilot key away, previous Windows on Arm devices, like those running Qualcomm 8cx would also qualify, as those have NPUs and can run Copilot, too. The Copilot key is little more than a branding exercise as it simply launches Copliot by simulating Shift + WIndows + F23, and one could just hit Windows + C to get the same efffect.

Copilot+ PCs: Microsoft’s New Standard

Since then, Microsoft has introduced its Copilot+ PCs, with laptops from Asus, Dell, Acer, Samsung, HP, Lenovo, Samsung, and Microsoft's Surface brand. Those all use Qualcomm's Snapdragon X Elite and Plus processors at the moment, though Microsoft has said upcoming AMD Strix Point (aka Ryzen AI 300) and Intel Lunar Lake chips may also fit the bill. To be considered a Copilot+ PC, laptops need to have at least 16GB RAM, 256GB storage, and an on-board NPU that's capable of 40 TOPS (trillions of operations per second). The Qualcomm Snapdragon X Elite chips support 45 TOPS on the NPU.

Copilot+ PCs, the first of which will launch on June 18, 2024, will have a series of four unique Windows AI features that other PCs cannot access. These include Cocreator (image generation in Paint), Windows Studio Effects (webcam blurring and special effects), real-time translation and captions for audio and Recall. Recall, the controversial feature which keeps a record of almost everything you do on your PC so you can remember it, was just pulled from the Copilot+ launch date build of Windows.

So, anyone who doesn’t buy a Snapdragon X-powered laptop will have to wait on those special Windows features, likely for many months. AMD recently confirmed that its Ryzen AI 300 PCs won’t be getting the Copilot+ features when they launch later this year, but will eventually. Desktop users are left out of the cold until at least the launch of Intel Arrowlake in Q4. And anyone with a current-gen laptop or desktop is probably left out permanently.

Given the weak set of Copilot+ features, anyone who is actually paying attention probably isn’t that sad about having to miss them or wait for them. There are many other ways to get an offline (or online) AI image generator, to do real-time translation and to blur your webcam background. Recall is somewhat unique, but many people won’t want it, because of the privacy risks associated with taking constant screenshots of your work.

Intel and AMD define AI PC more Broadly

Intel, on its website, has taken a more general approach:  "An AI PC has a CPU, a GPU and an NPU, each with specific AI acceleration capabilities."

AMD, via a staff post on its forums, has a similar definition: "An AI PC is a PC designed to optimally execute local AI workloads across a range of hardware, including the CPU (central processing unit), GPU (graphics processing unit), and NPU (neural processing unit)."

Who is making AI PCs?

If you're following the definition that most of the industry is using: CPU, GPU, and NPU, then the answer is most laptop manufacturers are making them. They include Dell, HP, Lenovo, Asus, Acer, Samsung, Microsoft, among others.

These are laptops with Intel (Core Ultra "Meteor Lake"), AMD (Ryzen 7040 or 8040), or Qualcomm (Snapdragon X Elite or Plus) processors.

What is an NPU, exactly?

A neural processing unit, or NPU, are processors that specialize in parallel computing specifically for the purpose of AI workloads (GPUs often also use parallel processing in order to render advanced graphics). Intel, AMD, Qualcomm, and Apple have these attached onto the processor with the CPU and integrated GPU.

They're also highly efficient, allowing for longer battery life than running these processes on a CPU or a GPU (even if, in some cases, those might be the more performant options). NPUs run matrix math, allowing them to do things such as video decoding, upscaling and background removal at a fraction of the power. 

An NPU's performance is measured in tera operations per second, or TOPS. Intel’s Meteor Lake processors and their NPUs can only do about 10 TOPS while Qualcomm’s Snapdragon X processors, AMD’s Ryzen AI 300 processors and Intel’s Lunar Lake chips will all deliver 45 or more TOPS from their NPUs. 

What about desktop PCs? What about GPUs?

At the moment, every PC with an NPU is a laptop (or, in some cases, tablets or gaming handhelds). Intel has yet to put an NPU on a desktop chip, but AMD first did back in January with the Ryzen 7 8700G and Ryzen 5 8600G, each of which support 16 TOPS via Ryzen AI. (Those technically based on mobile designs, but with more power and in a desktop form factor). A big reason to include the NPU on mobile devices is that they are extremely power efficient, which helps with battery life. Desktop systems don't have batteries, so processor manufacturers can still pump their processors using more wattage.

Additionally, desktops are more likely to have discrete GPUs, which have also proven to be extremely adept at certain AI tasks (though this is mostly high-end parts, like the Nvidia GeForce RTX 4090, which has 24GB of RAM to work with).

In fact, in a recent blog post, Nvidia's vice president of consumer AI, Jason Paul, suggested that the company started the AI PC boom all the way back in 2018, when it launched its first GPUs with RTX tensor cores and DLSS with the RTX 20-series and Turing architecture. That's yet another different definition from the ones laptop and CPU companies are making.

While this doesn't fit the conventional, NPU-focused definition that many are working with, there are a number of companies putting the AI label on desktop PCs, too.

Newegg, for example, is selling its ABS desktops (which start with consumer-grade parts like the Nvidia GeForce RTX 4070 Super and an Intel Core i5-14400F for $1,800 and goes up from there), on an AI PC page alongside laptops using Intel NPUs, as well as desktop parts ("AI CPU," "API GPU"). MSI lists a number of its desktops using 14th Gen Intel processors as "AI Gaming Desktops" because of software that the company includes.

The desktops that make more sense to call AI PCs are workstations that have the power to train some models. Maingear, for instance, sells its Pro AI systems with Intel Xeon chips and Nvidia's RTX 5000 and RTX 6000 Ada GPUs. These range from $28,000 to $60,000 and are definitely not for people looking just to generate images or photos.

Are Macs AI PCs?

Apple is introducing its take on the AI to Mac laptops and desktops this fall. The company is launching a beta version of  Apple Intelligence with generative writing, images, custom emoji, and a more capable version of its Siri assistant in macOS Sequoia this fall (as well as iPadOS 18 and iOS 18).

Apple will support Intelligence on Macs and iPads using its M1, M2, M3, and M4 families of processors. On the iPhone, it will support the A17 Pro on the iPhone 15 Pro and Pro Max (and presumably whatever is in the next iPhone this fall).

Because Apple uses its silicon across all of its Macs, its desktops, like the iMac, Mac Studio, and Mac Pro also have NPU (or as Apple refers to them, Neural Engines). That means macOS will have AI-specific features on desktop running on NPUs before Windows.

Do I need an AI PC?

Right now? "Need" is a strong word.

AI features are still in their infancy. In many cases they're still in beta. Many popular chatbots, like OpenAI's ChatGPT and Google Gemini, are totally cloud-based, as is most of what Microsoft Copilot does.

That's not to say there's no features you'll miss. On Copilot+ PCs, image generation built into Windows 11 and Restyle photo editing are exclusive to the new Snapdragon systems. Some other software, like the Adobe Photoshop, Lightroom, and Express as well as Da Vinci Resolve will use NPUs for some AI features and the cloud for others. Those NPU features may be slower — or non-existent on older computers.

NPUs are also being used to power background blur and other camera effects on some PCs, though you don't inherently need an NPU for that kind of work (though it does free up the CPU and GPU).

But across the board, these features are still just rolling out, and it's unclear which will be the most useful to you. Local AI is more secure, as you don't have to send your information to another company's servers, but if you're using it in limited scenarios, the cloud functionality the most popular apps currently offer should more than suffice.

There's a lot of hype around AI. And while it has some legitimately cool uses, there are still plenty of places where it's unclear how much people will want to use it. If your current PC is still doing what you need it to do and is getting security updates, it may be worth waiting as more powerful tech comes out to support, presumably, more local AI tools, and see what you actually need. 

It's already clear that companies will lock certain features to newer PCs. While early adopters may jump at the chance to try them, there's also no harm in letting those people be the beta testers (and many of these AI features are being labeled as beta by the companies that make them) and grab something when you know what you want.

Are AI PCs more secure?

One of the biggest pushes for AI on a laptop, rather than in the cloud, is security. After all, running an AI workload on an NPU in your computer means you don't necessarily need to send your information to the cloud.

That being said, that also means AI features need to be built securely, too. After security researchers discovered how easy it was to steal data from, Recall, Microsoft’s new AI feature that takes screenshots of all your activity, for later reference was pulled from the initial Copilot+ set of features as Microsoft promised more security and testing with Windows Insider members.

If you are running a business and having an LLM manage top-secret corporate data, having it processed locally would be more meaningful. But most of the AI features currently being marketed are not mission-critical business tools. Perhaps if Microsoft 365 Copilot ran completely locally, that would be a plus for some companies.

No matter what type of PC you're using, you should still adhere to good security practices. Other humans are still outsmarting us there.

Bottom line

Right now, the term “AI PC” is still somewhat vague. CPU vendors and Microsoft are using the term to sell you new computers (currently, new laptops only) that have powerful NPUs built into their processors. Most of the generative AI features people actually use today (chatbots, image generators) are free to use in the cloud, making them less than “must-haves” in their local form.

But NPUs do promise to save battery life by performing some common tasks such as video playback at much lower power. Some web browsers, such as Edge, can use the GPU today to do AI upscaling of videos, but soon that will be offloaded to the NPU. For creative professionals who are doing audio, video or photo editing while unplugged, the NPU will tackle tasks such as background noise removal at much lower power than your CPU or graphics card; of course, the software has to be optimized to do so. And, in the near future, we’ll see more tasks transferred to NPUs, which in turn, will increase system efficiency even more. 

So, in the end, the killer feature of AI PCs (at least if they are laptops) could really be longer endurance. If your laptop goes from lasting 12 hours on a charge to lasting 20 hours on a charge because of NPU usage, that could be a huge deal. But the generative features are still in their infancy, so if you're not ready to upgrade, there's still time to wait and see what evolves.

Correction, June 17, 3:44 p.m. ET: This article has been corrected to point out that AMD has NPUs on the desktop with its Ryzen 8000G processors.

Framework made a snarky post on X last May 29, asking if it was the only laptop maker that doesn’t slap AI branding all over its landing page. This was probably made in regards to the upcoming Computex 2024, which started on June 3, and had AI everywhere. The company then made a follow-up post a couple of weeks later, saying that May was its highest revenue month since Framework started, meaning its non-use of AI branding isn’t having a negative effect on the company’s sales.

Despite that, Framework’s current laptop offerings already have AI capabilities. The Framework Laptop 16 can be had with a Ryzen 7 7840HS or a Ryzen 9 7940HS, with both chips capable of hitting 10 TOPS. On the other hand, you can pick between the Intel Core Ultra 5 125H, Ultra 7 155H, or the Ultra 7 165H process for the smaller Framework Laptop 13, with the best Intel chip capable of hitting up 34 TOPS.

Ever since AI entered mainstream consciousness, many companies have taken advantage of the mystique and mystery of this technology, with some adding AI branding to their products given the chance. For example, AMD’s mobile chips for the Ryzen 7000 and 8000 families had the 70XX and 80XX naming scheme. However, its latest Strix Point chips, which we expected to be named Ryzen 80XX, instead became Ryzen AI 300.

HP created a new AI Helix logo to differentiate its laptops that with 40+ TOPs per second and has Windows 11, and Microsoft even launched its Copilot+ PC branding last month. Even Cooler Master China tried to get in on the bandwagon by announcing an ‘AI Thermal Paste’, although the company later clarified the issue saying it was just a mistranslation.

AI technologies are useful in several applications, like medicine. For example, some scientists used an LLM to interpret thoughts, giving us the possibility to build an interface that would allow a quadriplegic to operate an machine that will help them regain their mobility and independence. However, AI (or at least its current iteration) isn’t at the level yet of what most people think of — a living, thinking machine that can reason just like a person. But marketing people are using our preconceived notion of what AI is to sell us machines.

Sure, the NPUs and other features on these AI devices (except for the thermal paste) are technically ‘AI’, but they’re not yet the AI we see in movies and TV shows like Star Trek, Star Wars, or (and we hope this will never come to pass) Terminator. So, while marketers and ad agencies are free to use the AI branding all they want, we, the consumer, should remember that AI is still in its infancy on the PC.

Hoping to simplify the ever-expanding world of the PC buyers' market, HP put out a press release disclosing its new Omni-, Elite-, and Pro- prefixed naming scheme across its entire line of laptop and desktop PCs. This new naming scheme is being introduced alongside HP's own AI PC logo dubbed an "AI Helix", and an even-odd numbering scheme with even numbers denoting PCs for commercial users and odd numbers denoting PCs for consumers.

Meanwhile, the "Ultra" naming is meant to denote "the most premium offerings or devices that redefine traditional form factors". Below, we've included some more official imagery and a table breaking down all the tiers established by HP's new PC branding scheme.

All New HP PC Models

Besides the new PC branding, HP is also pushing the AI PC branding with its new AI Helix logo, pictured above. According to HP, this means "they are built with unique HP AI experiences, such as HP AI Companion and Poly Camera Pro", have an NPU capable of 40+ TOPs per second, have protection against "AI-generated threats", and a "next-gen OS", which we're pretty sure just means Windows 11.

As nice as it is for HP to put in this effort to streamline its branding and clarify its offerings a little better, it's difficult to ignore some of the more flowery language presented in the original release. When explaining the shape of its AI PC logo, HP claims it is "reminiscent of the structure of DNA, symbolizing our commitment to enable the DNA of AI into our devices. Just as DNA is at the core of life's evolution, we believe AI is the driving force behind technological advancements already enhancing human potential."

While that's mostly marketing speak, it's important to remember that as it currently stands, the greatest-scale applications of AI are being seen in "generative AI" which has rightfully drawn incredible scorn from skilled artists and laborers whose livelihoods are openly being targeted. Not that AI can't be used ethically or isn't cool, of course.

Ever since Intel made an initial announcement about “AI PCs” for Windows in October 2023, everyone’s been guessing about what’s involved in putting local AI horsepower on a properly-endowed laptop or desktop PC. Back in December, Intel began shipping its first “AI PCs,” a series of laptops with its Core Ultra “Meteor Lake CPUs,” the first to come with built-in Neural Processing Units (NPUs). And many consumers bought one, with hopes that Microsoft would add new local AI features that take advantage of this added processing power.

As of today’s (May 20) Microsoft event, the waiting and wondering are mostly over as the software giant unveiled a new class of computer, the Copilot+ PC, that will get a series of exciting and exclusive Windows AI features. Unfortunately, laptops with Intel Meteor Lake or AMD Ryzen 7000 CPUs aren’t powerful enough to make the cut so, if you bought one hoping it would benefit from future Windows updates, you wasted your money. The first Copilot+ PCs will only come with Qualcomm Snapdragon X series processors.

For more detail on the event, check out our blow-by-blow coverage of this event. This live blog is my primary source for images, quotes and information as presented in this story.

Us versus Them (People and PCs, That Is…)

When Microsoft CEO Satya Nadella got up early on to talk about the new Copilot initiative, he put emphasis on moving past the cloud and onto the device where AI is concerned. Said he: we need to learn “how to build computers that understand us, instead of us having to understand computers.” FWIW, I agree. He continued: “I think we’re real close to [that] breakthrough.” 

I completely endorse the idea that computers should help bear the load of making themselves more useful, friendly and productivity-enhancing. It’s interesting to see this however, as a play to get users of all stripes, from hobbyists to IT professionals, data scientists, and software developers to pony up for new and more expensive PCs in the name of attaining a brave new nerdvana. It’s just a shame that those who buy an Intel-powered  “AI PC” today are left out of the party.

During today’s event, new Surface Laptop and Pro PC models carried some of the burden of demonstrating what kinds of benefits a Copilot+ PCs can deliver. MS EVP and Consumer Chief Marketing Officer Yusuf Mehdi described them as the “fastest Windows PCs” ever built, thanks in no small part to a re-architected Windows 11 OS (presumably, some preview version of the upcoming 24H2 release expected later this year). 

Indeed, the afore-linked Tom’s coverage asserts that “Mehdi…claimed that the PCs are 58 [percent] faster than Apple’s new[] MacBook Air with M3 processor.” That’s significant. A new Windows 11 feature called Recall can help users “find anything [they’ve] ever seen or done on [their] computer[s].” Recall works with voice prompts, too: a demo showed a voice search that produced “a chart with purple text … created earlier in the week” from a PowerPoint deck. Good stuff!

Copilot+ PCs that are powered by Qualcomm Snapdragon X series CPUs are on the way from all the usual OEMs, including Acer, Asus, Dell, HP, Lenovo. And right now, those Snapdragon CPUs are the only ones which qualify a system for the Copilot+ PC branding and the added features that come with it. 

One big issue with using a PC that has a Snapdragon processor is that it requires Windows on Arm and there could be compatibility issues. However, a raft of enhancements to Windows on Arm will support Copilot+ PCs, including native Microsoft and 3rd-party apps for the Arm platform and a new Prism emulator that promises solid compatibility with x86 apps. Adobe is also bringing its Creative Suite to Windows on Arm.

These new Surface models are much faster than their predecessors, too: the Surface Laptop Copilot+ Pro is said to be 80% faster than the Surface Laptop 5). What’s more it can run three 4K monitors along with its built-in display (which has no notch) and promises the longest battery life for any Surface model, period. Similar stats apply to the Surface Pro Copilot+ PC model as well. Sounds pretty amazing, eh?

What About Existing “AI PCs?”

There’s at least one fly in the ointment, though. There are already so-called “AI PCs” in user’s hands. I’ve got one right next to me as I write this story: a pretty powerful Lenovo Yoga Pro 9i (Intel Core Ultra 9 185H with built-in ARC Meteor Lake GPU and built-in NPU, 32 GB RAM, NVIDIA GeForce RTC 4050,and more). The problem is that its built-in NPU fails to deliver the kind of performance (40+ TOPS or Trillion Operations Per Second) that Microsoft says is needed to meet Copilot+ PC hardware requirements. Intel’s Meteor lake chip, its current state-of-the-art for laptops, can only hit about 10 TOPS, but next-gen Lunar Lake processors could meet the requirement. 

When I checked this item on the Lenovo website at the end of April it came with an online price tag of just over $2,100. But when I tried to visit its link just now, it’s showing “out of stock.” My best guess is that Lenovo probably won’t build any more. Instead, it will probably wait to get the parts it needs to put valid Copilot+ PC units together, then resume shipping when those are in hand.

But gosh, people who’ve already paid good money for early models of AI PCs may feel left behind in the wake of these now-clarified hardware requirements. The real gotcha comes from the CPU and NPU requirements which exceed currently available laptops like the Lenovo Yoga Pro 9i. Qualcomm-powered laptops are probably coming soon, but Intel and AMD-powered competitors may take a while to hit store shelves. If history is any guide, that probably means October or November, once Windows 11 24H2 is in general distribution and suitable hardware to support Copilot+PC requirements is up for sale.

The real question has to be: Is it worth spending US $1,500 or more for a new Copilot+ PC to take advantage of its AI capabilities? Based on Microsoft’s discussion of its new Surface models with Copilot+PC capability, there’s likely to be plenty of pure performance to help justify the cost. The AI features that Copilot will bring to Windows 11 also appear to have at least great potential value. 

Thus, the answer for those already planning a hardware refresh, or buying a new PC in the right timeframe is probably an emphatic “Yes!” But for those who’ve just recently bought what they thought was a state-of-the-art PC that doesn’t meet Copilot+PC hardware requirements that answer is somewhere between “Ouch!” and “No!”

Note: As with all of our op-eds, the opinions expressed here belong to the writer alone and not Tom's Hardware as a team.

Qualcomm and Microsoft have tried to make Windows on Arm a serious platform for years, and this next big swing is the most promising shot yet. In mid-2024, the first devices with Qualcomm's new Snapdragon X Series processors — the Snapdragon X Elite and Snapdragon X Plus — are poised to ship.

This time, things may be different. Qualcomm's X Elite and X Plus are based on the company's new Oryon cores, which the company is betting on for more power and efficiency. Following Apple's success with its M-series chips, Qualcomm is looking to prove what Arm can do for Windows notebooks, aiming for thin, long-lasting, powerful PCs. 

And in keeping with the times, these processors are also being touted as artificial intelligence powerhouses. Each carries an NPU capable of 45 trillion operations per second (TOPS), which Qualcomm is calling "the world's fastest NPU for laptops." Intel's existing Core Ultra "Meteor Lake" processors have NPUs that support 10 TOPS, while AMD's 8040-series "Hawk Point" goes up to 16 TOPS. This means Qualcomm will offer a significant advantage (at least, until Intel's Lunar Lake launches). There have been many rumors about a version of Windows 11 that runs some AI elements locally, and the X Series chips may be some of the first to enable that.

Among Windows machines, the Snapdragon X Elite and X Plus are a chance to show off what Arm can do. If it succeeds, it will be incredibly disruptive to the x86-dominated mobile market.

Qualcomm Snapdragon X Elite and X Plus at a glance

• Qualcomm Oryon CPU: 12-core Snapdragon X Elite, 10-core Snapdragon X Plus
• Qualcomm Adreno GPU
• Qualcomm Hexagon NPU
• Qualcomm Sensing Hub
• Snapdragon X Elite: 75 TOPS across CPU, GPU, NPU, and micro NPU (45 TOPS from NPU)
• 4nm TSMC process
• Coming "mid-2024"
• Supports up to 64GB LPDDR5, with 136 GB/s memory bandwidth
• Wi-Fi 6, 6E, and 7 support, along with Snapdgargon X65 5G Modem
• Will be used by at least 9 OEM partners

Qualcomm Snapdragon X Elite and Plus specifications and features

Qualcomm's Snapdragon X Elite and X Plus are the first processors to use Qualcomm's new Oryon cores. The Oryon cores come from Nuvia, a startup that Qualcomm bought back in 2021 for $1.4 billion. 

The Snapdragon X Elite will have 12 cores running up to 3.8 GHz, or up to 4.3 GHz using single- and dual-core boost. The Snapdragon X Plus has a 10-core CPU, running up to 3.4 GHz on all cores.

The GPU is a Qualcomm Adreno that promises up to 4.6 TFLOPS. The X Elite includes support for up to a 4K/120 Hz display and up to triple 4K external monitors. (This beats some of Apple, the other big Arm competitor, on some chips. M3 laptops can handle two external monitors, but only with the lid closed. The m3 Pro can handle two external displays with the laptop screen, while the M3 Max can handle three external screens with the laptop open.) The NPU, dubbed Hexagon, is rated for 45 TOPS. 

Qualcomm hasn't listed a specific TDP for these chips, though we've previously been told that Qualcomm "dynamically scales performance across various TDPs[.]" That seems to suggest there will be some room for OEMs to decide how to use these systems, including the possibilities of clamshells, 2-in-1s, and both fanless and actively cooled machines. (Intel and AMD also offer configurable TDPs, though those companies publish more concrete numbers.)

Another, smaller NPU, the "Micro NPU," is on a part of the chip that Qualcomm calls the "sensing hub," which also includes the Qualcomm Spectra image signal processor (ISP) and a DSP.

Qualcomm Snapdragon X Series Platform Specs

(A note on the memory: Qualcomm previously listed the Elite's transfer speed at 8533 MT/s, but has since changed it to 8448 MT/s.)

Qualcomm Snapdragon X Series Processor Specs

We know there will be multiple configurations of the Snapdragon X Elite chips, but there has only been one Plus chip announced so far.

There are three Snapdragon X Elite chips:

• Snapdragon X Elite X1E-84-100
• Snapdragon X Elite X1E-80-100
• Snapdragon X Elite X1E-78-100

The sole Snapdragon X Plus is the X1P-64-100.

All three Elite chips use 12 cores, but their speeds vary. The top chip, the X1E-84-100, goes up to 3.8 GHz on multi-threaded workloads, and can hit 4.2 GHz with dual-core boost. The X1E-80-100 goes up to 3.4 GHz and 4.0 GHz, respectively, while the X1E-78-100 also goes up to 3.4 GHz on multi-threaded tasks, but doesn't offer dual-core boost.

The name Snapdragon X Plus suggests it's above at least one more chip, but Qualcomm hasn't made any other announcements yet, so we can only speculate.

The Snapdragon X Plus X1P-64-100 has 10 cores, and the same 3.4 GHz for multi-threaded performance.

Each of the chips has the same 45 TOPS of performance from the Hexagon NPU. The top Elite chip offers 4.6 TFLOPS of GPU performance, while the entire rest of the lineup drops to 3.8 TFLOPS.

These names are lengthy, to say the least. Qualcomm already has a decoder for it.

After Snapdragon X, the first number (in all of these instances, 1), marks the generation of the chip. E stands for Elite, while P is Plus. Easy enough so far.

The next number serves as the SKU level. In other words, on a platform like Elite, where there are three options, the higher numbers mean more powerful chips. The 84 is the top chip, followed by the 80, and 78. The Snapdragon X Plus tops out at 64. It's unclear if these are on some sort of scale out of 100, or why 84 is the current top number. It seems kind of arbitrary, but it does leave room for more powerful chips.

The last three numbers, which are all "100," are marked as "variants." It's unclear what exactly Qualcomm means by that, and the company didn't tell us when we asked.

"We are not disclosing any further details at this time," Qualcomm spokespeople wrote in an email.  "These variant numbers can be used for future products."

Qualcomm Snapdragon X Series benchmarks

When we got to see the Qualcomm Snapdragon X Elite in a controlled hands-on, the company had published the following benchmarks:

Snapdragon X Elite X1E80100 (3.4 GHz), paired with 16GB of RAM:

Snapdragon X Elite X1E84100 (3.8 GHz), paired with 64GB of RAM:

As we wrote in our hands-on, we found that these benchmarks (albeit very controlled benchmarks that were pre-installed), appeared within Qualcomm's range. One run of Geekbench on the 3.4 GHz processor was slightly lower, but everything else was either in range or even better than what Qualcomm suggested. Of course, we'll need to see these chips in retail laptops to get an idea of what the real experience will be like.

There have also been a handful of leaked Geekbench results, but it's hard to verify if they're real. Some of them do fall in line with Qualcomm's suggestions, while others fall below, though it's hard to know if new drivers have changed things, or if there are more configurations we don't know about. 

We've also been hands-on briefly with a reference design using the Snapdragon X Plus:

Again, we saw numbers that matched the claims (and sometimes the claims were conservative), in Qualcomm's controlled situation.

Qualcomm has made a number of claims that we'll only be able to verify once we have laptops to test. 

The only test that Qualcomm has shared against Apple is multi-threaded CPU performance in Geekbench 6.2 (the M3 was tested in a MacBook Pro), where the X Elite was 28% faster and the X Plus was 10% faster than M3. Qualcomm should beat Apple here: M3 has 8 cores, while Snapdragon X Plus has 10 cores and X Plus has 12 cores. Single-core performance was notably compared, and Apple's 12-core M3 Pro and 14-core or 16-core M3 Max were left out.

In many of Qualcomm's tests, it claims that the Snapdragon X Elite and X Plus are faster than the Intel Core Ultra 7 155H and AMD Ryzen 9 7940HS and can match peak performance at less power:

Granted, now we know that Qualcomm has various X1 Elite chips, so we'll have to see in our own testing how each one stacks up against the competition — and against each other.

Qualcomm Snapdragon X Series laptops

There aren't any Qualcomm Snapdragon X Elite or Plus laptops available yet, nor have any been officially announced. But at the 2023 Qualcomm Snapdragon summit, the company said it was partnering with Acer, Asus, Dell, HP, Honor, Lenovo, Microsoft Surface, Samsung, and Xiaomi. Otherwise, it has just been showing off its bright red reference notebooks.

Qualcomm's mockups in presentations have suggested that we'll see a mix of clamshells, foldable 2-in-1s, and detachable 2-in-1s.

There have been a number of rumors about laptops using the new Snapdragon X chips. Earlier this year, Microsoft announced business versions of the Surface Pro 10 and Surface Laptop 6, but not consumer version; rumors suggest those will have these new chips. We've also seen reports of a new Samsung laptop, the Galaxy Book 4 Edge to use these processors, as well as leaks of Lenovo's ThinkPad T14s Snapdragon Edition and Yoga Slim 7 14 Snapdragon Edition.

Qualcomm Snapdragon X Series pricing and availability

Qualcomm says its Snapdragon X Elite and Plus chips will show up in laptops in "mid-2024," which means we should see them in the next few months.

We don't yet know pricing, but the impression that we get is that these chips — especially the ones running the X Elite processors — are meant to be for high-end systems. We'll get a better idea when the actual laptops are announced.

Will Qualcomm Snapdragon X Series work with more Windows apps?

If you remember Qualcomm's previous Snapdragon for Windows chips, you'll know that Microsoft needed to do emulation work to get everything running. Back then, you couldn't run most games, antivirus, and a number of other applications.

That should be better this time around. Snapdragon has shown off games such as Control and Baldur's Gate 3 running in emulation.

Qualcomm has also stated that there are more native apps. Most notably, there's now an ARM64 version of Chrome. Microsoft has been pushing for more native Arm apps since the release of its developer kit, previously known as Project Volterra. Either way, through emulation or native apps, we're hoping to see better performance and fewer issues.

Microsoft Build is coming up, and we expect to hear a lot about the future of Windows 11 on Arm and AI on the platform there. We’ll of course update this article when we learn more.

We have just witnessed a 3% year-on-year growth in global PC shipments, following two years of decline, with the industry returning to pre-pandemic levels. The change -- confirmed in a report by Counterpoint Research -- is credited to the onset of the AI PC era, shipment recoveries in respective sectors, and end-user upgrade cycles. 

Counterpoint's report broadly agrees with the recent findings of IDC, which detailed 8% growth during Q1 2024. Both research groups reckon Lenovo is enjoying the lion's share of the global PC market with 23-24%, followed by HP and Dell. Apple also enjoyed some growth, which may be credited to its M3-based systems. These companies are consistent and resilient in maintaining their top four placings in the global market.

This year, we should see many AI PCs and notebooks sold globally. Qualcomm also wants a piece of the personal computing market, as it intends to compete against Intel and AMD with its highly anticipated Snapdragon X-powered notebooks, teased to debut in just a few days. That said, even AMD and Intel have been preparing to launch processors and SoCs with artificial intelligence processing boosts. Industry players hope AI adoption will exponentially grow the market, but we are still waiting for the killer app(s) that integrate AI on PCs.

Progress despite trade sanctions in China and Russia

It's also noteworthy that domestic brands in Russia and China are expected to commandeer large portions of their PC market, largely due to sanctions. With mainstream brands being displaced, one might have expected market shares to be more disrupted. Remember, China is regarded as one of the most important and largest PC markets in the world. 

While China has made progress with its chip-making business as it continues to import new tools, Russia has seen less success. It is thought that China has been supplying its chips to Russia, creating another revenue flow. Regardless, both countries have their own solutions for operating systems and software to replace Windows and cloud computing apps like Office 365. It is too early to say if research groups will be able to assess these country's growth levels, but it will be interesting to keep track of regardless. 

In conclusion, despite the disruptive conditions and headwinds, the AI-fuelled growth and reinvigoration of the global PC market looks like it already has some momentum in early 2024.

Intel CEO Pat Gelsinger flashed a Lunar Lake processor, the company's next-gen laptop chip, at its Vision 2024 event, saying the chip provides 100+ TOPS of performance in AI workloads, with 45 of those TOPS coming from the NPU alone. That's 3X the AI performance of Intel's current-gen chips and meets the bar of 45 TOPS from the NPU that the company recently discussed during its AI Summit in Taipei as the requirement for next-gen AI PCs. At that event, Intel executives, in a question-and-answer session with Tom's Hardware, also said that elements of Microsoft's Copilot will soon run locally on Windows PCs. 

Intel's current-gen Meteor Lake chips only provide 10 TOPS of performance from the NPU, falling below the bar for next-gen AI PCs. Lunar Lake's 45 TOPS from the NPU meets that bar exactly. Gelsinger didn't elaborate on how much of the remaining 55+ TOPS comes from the CPU and GPU, but it's reasonable to expect in the range of 50 TOPS from the GPU and 5 to 10 FLOPS from the CPU cores.

AMD's current-gen Ryzen Hawk Point platform has an NPU with 16 TOPS of performance, which also falls below the bar for next-gen AI PCs. However, the company recently sent us a statement regarding its next-gen products: 

“We believe an AI PC requires strong CPU, GPU and NPU engines, which is what AMD has been delivering for more than a year with our Ryzen 7040 and now 8040 Series. At our December Advancing AI event, we disclosed our next-gen “Strix Point” mobile processors with XDNA 2 architecture would have up to 3x the generative AI performance of the current generation. We believe this performance will position us to remain the leading choice for next-gen AI PCs." — AMD representative to Tom's Hardware.

AMD wouldn't elaborate further and break out the amount of performance delivered from each unit, but simple math puts AMD's Strix Point NPU at 48 TOPS if NPU performance triples.

Qualcomm's Snapdragon X Elite chips are the elephant in the room — these Arm chips will debut with 45 TOPS of performance from its NPU (75 TOPS total) in the "middle of the year," setting up a tight race between Intel, AMD, and Qualcomm for the lead in the next generation of AI PCs. 

Intel appears to have a running start with its Core Ultra Meteor Lake chips. Gelsinger says the company has already shipped five million AI PCs to date and plans to ship 40 million units by the end of the year. That's a solid start towards the company's goal of shipping 100 million AI PC chips by the end of 2025. 

We've previously reported on industry rumors that Microsoft's Copilot AI service will soon run locally on PCs instead of in the cloud and that Microsoft would impose a requirement for 40 TOPS of performance on the Neural Processing Unit (NPU), but we had been unable to get an on-the-record verification of those rumors. That changed today at Intel's AI Summit in Taipei, where Intel executives, in a question-and-answer session with Tom's Hardware, said that Copilot elements will soon run locally on PCs. Company representatives also mentioned a 40 TOPS requirement for NPUs on next-gen AI PCs.

Microsoft has been largely silent about its plans for AI PCs and even allowed Intel to officially announce Microsoft's new definition of an AI PC. Microsoft’s and Intel’s new co-developed definition states that an AI PC will have an NPU, CPU, GPU, Microsoft’s Copilot, and a physical Copilot key directly on the keyboard. 

PCs meeting those requirements are already shipping, but that is just the first wave of the AI PC initiative. Intel divulged future AI PC requirements in response to my questions about potential memory criteria. 

"But to your point, there's going to be a continuum or an evolution, where then we're going to go to the next-gen AI PC with a 40 TOPS requirement in the NPU," said Todd Lewellen, the Vice President of Intel's Client Computing Group. "We have our next-gen product that's coming that will be in that category." 

"[..]And as we go to that next gen, it's just going to enable us to run more things locally, just like they will run Copilot with more elements of Copilot running locally on the client. That may not mean that everything in Copilot is running local, but you'll get a lot of key capabilities that will show up running on the NPU."

Currently, Copilot computation occurs in the cloud, but executing the workload locally will provide latency, performance, and privacy benefits. Notably, Intel's shipping Meteor Lake NPU offers up to 10  TOPS for the NPU, while AMD's competing Ryzen Hawk Point platform has an NPU with 16 TOPS, both of which fall shy of the 40 TOPS requirement. Qualcomm will have its oft-delayed X Elite chips with 45 TOPS of performance in the market later this year. 

Lewellen explained that Microsoft is focused on the customer experience with the new platforms. As such, Microsoft insists that Copilot runs on the NPU instead of the GPU to minimize the impact on battery life. 

"We had a lot of discussions over the course of the last year[with Microsoft], and we asked, 'Why can't we just run it on the GPU?' They said they want to make sure that the GPU and the CPU are freed up to do all this other work. But also, we want to make sure it's a great battery life experience. If we started running Copilot and some of those workloads on the GPU, suddenly you're going to see a huge hit on the battery life side," Lewellen explained. 

While AMD holds a slight lead in NPU TOPS performance, and Qualcomm claims a much bigger advantage in its not-yet-shipped chips, Intel says its roadmap includes next-gen processors to address every segment of the AI market.

"We have our product roadmap and plan on where we're at in mobile with premium and mainstream, but then you also go down into entry. And so we have plans around entry. From a desktop standpoint, we have plans on the desktop side, what we would say [is an] AI PC. And then there's also the next-gen AI PC, the 40 TOPS requirements; we have all of our different steps in our roadmap on how we cover all the different segments."

Intel's Lunar Lake processors will come to market later this year with three times more AI performance on both the GPU and the NPU than the existing Meteor Lake chips. Intel is already sampling those chips to its partners in preparation for a launch later in the year. Those chips will face off with Qualcomm's X Elite and AMD's next-gen processors. 

In the meantime, Intel is working to expand the number of AI features available on its silicon. As we covered in depth earlier today, the company plans to support 300 new AI-enabled features on its Meteor Lake processors this year. 

Many of those features will be optimized specifically for Intel's silicon. The company told me that roughly 65% of the developers it engages with use Intel's OpenVino, which means those applications are optimized specifically for Intel's NPU. The remainder of the developers use a 'mix of splits' between ONNX, DirectML, and WebNN, and Intel says it is happy to work with developers using any framework. 

However, the work with OpenVino could provide Intel with a nice runway of Intel-specific AI features as it heads into the Lunar Lake generation. Those are the types of advantages the company is obviously looking to enable through its AI PC Accelerator Program. The company says it has seen plenty of enthusiasm from the developer community, particularly in light of Intel's stated goal of selling 100 million AI PCs by 2025, which represents a big market opportunity for new AI software.

However, Microsoft's Copilot will run on NPUs from all vendors through DirectML, and having more TOPS will undoubtedly result in higher performance. That means we can expect a TOPS war to unfold, both in silicon and in marketing, over the coming years. 

Update 3/27/2024 6:50am PT: corrected Intel Meteor Lake and Ryzen Hawk Point NPU TOPS specifications. 

Intel announced two new extensions to its AI PC Acceleration Program in Taipei, Taiwan, with a new PC Developer Program that’s designed to attract smaller Independent Software Vendors (ISVs) and even individual developers, and an Independent Hardware Vendors (IHV) program that assists partners developing AI-centric hardware. Intel is also launching a new Core Ultra Meteor Lake NUC development kit, and introducing Microsoft’s new definition of just what constitutes an AI PC. We were given a glimpse of how AI PCs will deliver better battery life, higher performance, and new features.

Intel launched its AI Developer Program in October of last year, but it's kicking off its new programs here at a developer event in Taipei that includes hands-on lab time with the new dev kits. The program aims to arm developers with the tools needed to develop new AI applications and hardware, which we’ll cover more in-depth below.

Intel plans to deliver over 100 million PCs with AI accelerators by the end of 2025. The company is already engaging with 100+ AI ISVs for PC platforms and plans to have over 300 AI-accelerated applications in the market by the end of 2024. To further those efforts, Intel is planning a series of local developer events around the globe at key locations, like the recent summit it held in India. Intel plans to have up to ten more events this year as it works to build out the developer ecosystem.

The battle for control of the AI PC market will intensify over the coming years — Canalys predicts that 19% of PCs shipped in 2024 will be AI-capable, but that number will increase to 60% by 2027, highlighting a tremendous growth rate that isn’t lost on the big players in the industry. In fact, AMD recently held its own AI PC Innovation Summit in Beijing, China, to expand its own ecosystem. The battle for share in the expanding AI PC market begins with silicon that enables the features, but it ends with the developers that turn those capabilities into tangible software and hardware benefits for end users. Here’s how Intel is tackling the challenges.

What is an AI PC?

The advent of AI presents tremendous opportunities to introduce new hardware and software features to the tried-and-true PC platform, but the definition of an AI PC has been a bit elusive. Numerous companies, including Intel, AMD, Apple, and soon Qualcomm with its X Elite chips, have developed silicon with purpose-built AI accelerators residing on-chip alongside the standard CPU and GPU cores. However, each has its own take on what constitutes an AI PC.

Microsoft’s and Intel’s new co-developed definition states that an AI PC will come with a Neural Processing Unit (NPU), CPU, and GPU that support Microsoft’s Copilot and come with a physical Copilot key directly on the keyboard that replaces the second Windows key on the right side of the keyboard. Copilot is an AI chatbot powered by an LLM that is currently being rolled into newer versions of Windows 11. It is currently powered by cloud-based services, but the company reportedly plans to enable local processing to boost performance and responsiveness. This definition means that the existing Meteor Lake and Ryzen laptops that have shipped without a Copilot key actually don't meet Microsoft's official criteria, though we expect Microsoft's new definition to spur nearly universal adoption of the key.

While Intel and Microsoft are now promoting this jointly developed definition of an AI PC, Intel itself has a simpler definition that says it requires a CPU, GPU, and NPU, each with its own AI-specific acceleration capabilities. Intel envisions shuffling AI workloads between these three units based on the type of compute needed, with the NPU providing exceptional power efficiency for lower-intensity AI workloads like photo, audio, and video processing while delivering faster response times than cloud-based services, thus boosting battery life and performance while ensuring data privacy by keeping data on the local machine. This also frees the CPU and GPU for other tasks. The GPU and CPU will step in for heavier AI tasks, a must as having multiple AI models running concurrently could overwhelm the comparatively limited NPU. If needed, the NPU and GPU can even run an LLM in tandem.

AI models also have a voracious appetite for memory capacity and speed, with the former enabling larger, more accurate models while the latter delivers more performance. AI models come in all shapes and sizes, and Intel says that memory capacity will become a key challenge when running LLMs, with 16GB being required in some workloads, and even 32GB may be necessary depending on the types of models used.

Naturally, that could add quite a bit of cost, particularly in laptops, but Microsoft has stopped short of defining a minimum memory requirement yet. Naturally, it will continue to work through different configuration options with OEMs. The goalposts will be different for consumer-class hardware as opposed to workstations and enterprise gear, but we should expect to see more DRAM on entry-level AI PCs than the standard fare — we may finally bid adieu to 8GB laptops.

Intel says that AI will enable a host of new features, but many of the new use cases are undefined because we remain in the early days of AI adoption. Chatbots and personal assistants trained locally on users' data are a logical starting point, and Nvidia's Chat with RTX and AMD's chatbot alternative are already out there, but AI models running on the NPU can also make better use of the existing hardware and sensors present on the PC.

For instance, coupling gaze detection with power-saving features in OLED panels can enable lower refresh rates when acceptable, or switch the screen off when the user leaves the PC, thus saving battery life. Video conferencing also benefits from techniques like background segmentation, and moving that workload from the CPU to the NPU can save up to 2.5W. That doesn’t sound like much, but Intel says it can result in an extra hour of battery life in some cases.

Other uses include eye gaze correction, auto-framing, background blurring, background noise reduction, audio transcription, and meeting notes, some of which are being built to run on the NPU with direct support from companies like Zoom, Webex, and Google Meet, among others. Companies are already working on coding assistants that learn from your own code base, and others are developing Retrieval-Augmented Generation (RAG) LLMs that can be trained on the users’ data, which is then used as a database to answer search queries, thus providing more specific and accurate information.

Other workloads include image generation along with audio and video editing, such as the features being worked into the Adobe Creative Cloud software suite. Security is also a big focus, with AI-powered anti-phishing software already in the works, for instance. Intel’s own engineers have also developed a sign-language-to-text application that uses video detection to translate sign language, showing that there are many unthought-of applications that can deliver incredible benefits to users.

The Core Ultra Meteor Lake Dev Kit

The Intel dev kit consists of an ASUS NUC Pro 14 with a Core Ultra Meteor Lake processor, but Intel hasn’t shared the detailed specs yet. We do know that the systems will come in varying form factors. Every system will also come with a pre-loaded software stack, programming tools, compilers, and the drivers needed to get up and running.

Installed tools include Cmake, Python, and Open Vino, among others. Intel also supports ONNX, DirectML, and WebNN, with more coming. Intel’s OpenVino model zoo currently has over 280 open-source and optimized pre-trained models. It also has 173 for ONNX and 150 models on Hugging Face, and the most popular models have over 300,000 downloads per month.

Expanding the Ecosystem

Intel is already working with its top 100 ISVs, like Zoom, Adobe and Autodesk, to integrate AI acceleration into their applications. Now it wants to broaden its developer base to smaller software and hardware developers — even those who work independently.

To that end, Intel will provide developers with its new dev kit at the conferences it has scheduled around the globe, with the first round of developer kits being handed out here in Taipei. Intel will also make dev kits available for those who can’t attend the events, but it hasn’t yet started that part of the program due to varying restrictions in different countries and other logistical challenges.

These kits will be available at a subsidized cost, meaning Intel will provide a deep discount, but the company hasn’t shared details on pricing yet. There are also plans to give developers access to dev kits based on Intel’s future platforms.

Aside from providing hardware to the larger dev houses, Intel is also planning to seed dev kits to universities to engage with computer science departments. Intel has a knowledge center with training videos, documentation, collateral, and even sample code on its website to support the dev community.

Intel is engaging with Independent Hardware Vendors (IHVs) that will develop the next wave of devices for AI PCs. The company offers 24/7 access to Intel’s testing and process resources, along with early reference hardware, through its Open Labs initiative in the US, China, and Taiwan. Intel already has 100+ IHVs that have developed 200 components during the pilot phase.

ISVs and IHVs interested in joining Intel’s PC Acceleration Program can join via the webpage. We’re here at the event and will follow up with updates as needed.

Microsoft announced that Pavan Davuluri will be the new Windows and Surface team lead, following the unification of the previously-separate Windows and Surface teams under the new "Windows + (Surface) Devices" umbrella [h/t The Verge].

Pavan Davuluri has worked at Microsoft since 2001 and has been Corporate VP for the past three years, so the position seems a natural fit— especially since he's been managing the Surface team since 2015. Davuluri was responsible for Microsoft's prior work with AMD and Qualcomm for custom Surface CPUs, which makes him an easy MVP of the established Surface team.

Earlier this week, Microsoft announced two new Surface laptops, both with integrated Copilot AI assistant keys. These Surface laptops, Surface Pro 10 and Surface Laptop 6, had their announcement posted by Davuluri on Twitter, just days before he was revealed as the new Windows + Devices team lead at Microsoft.

Part of what reportedly drove this reorganization at Microsoft was its hiring of Google DeepMind co-founder Mustafa Suleyman for the new Microsoft AI team. The assembly of a dedicated AI team also aligns with previous reports of Microsoft developing its own networking gear for AI datacenters.

A memo posted by Rajesh Jha, Executive VP at Microsoft, states, "We are bringing together the Windows Experiences + Devices (E+D) division. This will enable us to take a holistic approach to building silicon, systems, experiences, and devices that span Windows client and cloud for this AI era. Pavan Davaluri will lead this time and continue to report to me. [...] The Windows team will continue to work closely with the Microsoft AI team on AI, silicon, and experiences."

It seems as though every major tech company is now busy talking about AI and how it will reshape their company, products, and the world around us. Nvidia has raked in billions over the past year selling its Hopper H100 and now H200 GPUs. It recently revealed the Blackwell B200 family, which ups the ante for AI training and inference yet again. The AI PC aims to democratize access to AI-powered tools and applications, supplemented by processing at the data center level.

Intel is working closely with Microsoft and others to push the AI PC narrative, with developer workshops and a renewed emphasis on building customized experiences. The hope is to sell more hardware and drive new PC experiences, powered by a combination of CPUs, GPUs, and NPUs — neural processing units.

As Microsoft, Intel, and others continue trying to push forward the era of "The AI PC", it seems paradigm shifts are necessary at the big tech companies. The unification of the Windows and (Surface) Devices teams at Microsoft will hopefully benefit both Windows as an operating system and future Surface devices as the pro-ready machines they're meant to be — but only time can determine Microsoft's success, here.

At an event in China earlier today AMD shared its AI PC ecosystem plans. The Beijing AI PC Innovation Summit saw the introduction of the Ryzen 'Hawk Point' 8040 series and 8000G desktop solutions to the Chinese market and, perhaps more interestingly to us, new AMD roadmaps including the next-gen consumer Ryzen chips. These codenamed 'Strix Point' chips will combine the charms of Zen 5, RDNA 3+, and XDNA 2 processing.

AMD CEO Lisa Su began her presentation on the now-familiar theme of artificial intelligence driving a revolution touching every corner of the tech industry. AMD has a wide range of AI accelerating processors covering the data center to AI PCs to the edge, but as this event was meant to introduce the latest Ryzen chips to China (and with partners like Asus and Lenovo taking part in the summit) we got a pleasingly large chunk of discussion about PCs.

Above you can see two of the key moments of Su's presentation. She had talked earlier at some length regarding Hawk Point (second slide) but above you can see that she revealed the "Next-Gen AMD Ryzen" chips codenamed Strix Point. It is good to see the CPU, GPU, and NPU cores are all upgraded beyond the current state-of-the-art Ryzen 8040 family

AMD's SVP of GPU Technology and Engineering R&D, David Wang, was next on stage to dive a little deeper into the newly official RDNA 3+ and XDNA 2 architectures. It really was just a little deeper, as he mainly talked about the current-gen Hawk Point. For example, perhaps the biggest thing we learned about RDNA 3+ was its name... Previously we have reported on sightings of RDNA 3.5 drivers, but now it looks like the GPU series also sometimes identified as GFX115X will be dubbed RDNA 3+. Without a major version number update, we expect only slight tweaks to the now-familiar RDNA 3 in the upcoming Hawk Point APUs.

Moving along to XDNA 2 information, we learned that this new NPU will offer triple the performance of a current-gen XDNA NPU. For reference, we know that the AMD Ryzen 8040 series delivers 16 NPU TOPS for 39 total TOPS. A quick calculation indicates that the next-gen Ryzen APUs should deliver over 70 total TOPS. Intel's Core Ultra series "delivers up to 34 TOPS," and Qualcomm's upcoming Snapdragon X Elite chips are supposedly capable of 45 TOPS.

AMD's Strix Point processors are set for launch and release later this year. Intel's Arrow Lake and Lunar Lake CPUs will also arrive in 2024 – packing three times more AI performance via both GPU and NPU enhancements.

Samsung has announced that its first AI PC series will become available son Feb. 26. The Samsung Galaxy Book 4 series consists of three base designs: Galaxy Book 4 Ultra, Book 4 Pro, and Book 4 Pro 360. All feature embedded Intel NPUs courtesy of one of Core Ultra 9/7 processors.

Other common yet desirable qualities of the Galaxy Book 4 family are a Dynamic AMOLED 2X touch screen with 120% DCI-P3, AKG quad speakers with Dolby Atmos, lengthy battery life, a new discrete Samsung Knox security chip, and enhanced connectivity as standard.

In the video embeded above, Samsung details ways the NPU will be useful to accelerate AI on the laptop. For example, we see AI in action in Adobe creative software and using Microsoft Copilot assistance for various smart and labor-saving tasks in Windows. (though at the moment, that's cloud-based). Another major focus of the brief video is on smart connectivity, which is particularly tightly integrated for owners of Samsung’s Android phones.

Though Samsung has great claims for the audio-visual experience on offer from all the above laptops and boasts about the utility delivered by its onboard AI, readers should, as always, check for reviews on sites like Tom's Hardware. We reviewed the Galaxy Book 3 Ultra last April, and it was surprisingly good though. Of course, the AI PC experience delivered by Samsung will also rely on partner software, and we are at a stage where developers are still establishing compelling use cases for PCs with onboard NPUs.

Samsung US hasn’t listed or detailed any of the new Galaxy Book 4 devices at the time of writing, though it is selling off Book 3 models in an event that ends on February 25. We note Samsung Canada has all the models available for pre-order, though. The Canadian page is selling multiple configurations of all three designs with configuration options for GPU, CPU, and storage (depending upon the model) at prices ranging from CAD$999 (USD$740) to CAD$2,999 (USD$2,222). Pricing also seems to be shown in South Korean and Indian markets.

Microsoft is seemingly lining up a number of new minimum specification levels for AI PCs that it hasn’t yet broadcast via official channels. We have heard from our own sources that AI PCs will move up the bar concerning minimum RAM configuration, and TrendForce appears to have heard the same thing and says that 16GB will be the minimum RAM configuration for Windows AI PCs. Meanwhile, both our own sources and TrendForce agree that new Windows PCs will require at least 40 TOPS of compute power to make the grade for labeling as an AI PC.

“Microsoft has set the baseline for DRAM in AI PCs at 16GB,” stated TrendForce in a press release about Microsoft Copilot on Wednesday. Thus, Windows will again be instrumental in driving growth for the minimum memory capacity acceptable in new PCs. Desktop users with easily accessible upgrade options might shrug, but those buying laptops and discovering they aren’t upgradeable due to soldered RAM should no longer have to filter out memory-starved systems - simply look for AI PCs.  

Memory makers should be happy with a boost in the number of PCs sold with more memory as standard. Last year we reported on some of the biggest players in the industry slowing production to constrain supplies and achieve better prices. That seems to have hurt the revenue generation of Samsung and SK hynix during 2023, but the damage was partially self-inflicted.

As mentioned in the intro, this won’t be a one-dimensional change of PC system requirements. The expectation that a new PC will run Microsoft Copilot AI assistance in a slick and responsive manner also relies on adequate local acceleration. A minimum of 40 TOPS of computational power has been decided upon by Microsoft HQ. That might be provided by a discrete GPU, but PC processors are almost all now up to speed in building-in efficient NPUs that can meet or exceed that compute performance target.

We said 'almost' above, and that is an important caveat, as the combined CPU, GPU, and NPU power within Intel’s current Meteor Lake chips are said to reach 34 TOPS at best. TrendForce speculates that Intel Lunar Lake will address this baseline underperformance for AI PCs. Intel itself has said that Lunar will have three times the AI performance of its predecessor, Meteor Lake. 

Other Windows PC processor makers like AMD and Qualcomm aren’t quite as far behind. The AMD Ryzen 8000 series (Strix Point) is expected to be capable of 45 TOPS. Qualcomm’s Snapdragon X Elite platform is also thought to deliver around 45 TOPS. It will be interesting to see if the Arm architecture processors from Qualcomm are as competitive using other performance metrics.

There has been quite a lot of speculation about the upcoming wave of AI PCs, as the industry seems to be quite excited by the hardware refresh cycle they think it will inspire. When the Windows Copilot key was unveiled, we wondered about any minimum spec a device might need. However, it turned out that even entry-level modern PCs without onboard NPUs were given the green light to equip this key. Perhaps, more stringent AI PC labeling and minimum specs will come in the summer with Windows 12, we mused. Now, perhaps, we have at least a partial answer to AI PC and Windows 12 minimum specs: a system must have at least 16GB of RAM and a processor that can achieve at least 40 TOPS of AI compute.

Intel's Core Ultra chips (formerly Meteor Lake) launched on December 14th, and are already appearing in laptops you can buy. At Intel's official launch here in New York City, I sat down with Dan Rogers, the head of Intel's performance marketing lab in its Client Computing Group.

Rogers, the former product manager for Alder Lake and Raptor Lake, has a unique job as Intel pushes out its first chips based around AI: figuring out how to tell the story of what these can do, and also measuring their performance for Intel's marketing to the general public (Intel is in the PC business, after all).

Because right now, while Intel (and, let's be frank - AMD, Qualcomm, and Apple) have systems with neural processing units, or NPUs, lots of potential customers don't quite know what they could do with them, or why it would matter. I asked him about how people are supposed to know what an AI PC can do, when they can expect software that will affect their lives, and also learned more about how Intel is benchmarking AI applications internally.

Selling the AI PC

Intel's debut mobile processor with an on-board NPU is coming as the company is making a major brand name change. Intel Core i is old hat. Intel Core Ultra is your AI-equipped chip. Rogers suggested to me that the name change is part of the "multiple levels of education" that Intel will go through to inform customers about what they're getting.

"All of our processors code named 'Meteor Lake' will be Core Ultra and all Core Ultra processors will be Meteor Lake," Rogers said. "So, it's a fairly simple method of communicating to the end user, right, especially at the point of sale and retail and so forth that this is a PC that has built-in AI."

How people will actually use an NPU or other AI support, however, is a bit more nuanced. Rogers suggested to me that while AI has been foreign to a lot of people until now, he expects that won't be the case going forward as Zoom, Adobe, and other name-brand PC software start shipping features. After all, while you can download an OpenVINO plugin for GIMP or Audacity, but let's be real — most people won't do that.

"Definitely there are more open source community tech enthusiast-oriented usages for sure, early adopters and folks that want to really dig in," he said. "But, you know, also for just the general user, much of what they do today will now be made better through AI.

Otherwise, Rogers said that Intel is relying on three major points to convince people to make the leap. The first point, of course, is the NPU. The second is the GPU, which is seeing Arc-branded graphics for the first time. The third is battery life — Intel is promising a way more efficient machine, especially with certain tasks from the CPU and GPU offloaded to the NPU. (I'm very curious to test this.)

Is Intel still a CPU company in 2024?

So Intel has a three-pronged message: NPU, GPU, battery. Huh.

Perhaps I was being facetious, but I couldn't help but notice one thing as Intel introduced the new processors: a lack of emphasis on the CPUs. The CPU felt like it took a backseat. Is Intel, I asked, still a CPU company?

"[The CPU is] extremely important to us… We still believe we have a really strong position in the core technology, but absolutely, Meteor Lake kind of builds on what we were able to do with

the Core and performance hybrid architecture and Alder and Raptor for performance," Rogers said. "Now we have a third level of hybrid architecture for power efficiency, and then we're working on GPU and NPU as well.

"So, you know, as the kind of former Alder and Raptor guy, for me, it's a cool product because,

you know, it sort of compliments those advances that we had in the prior gen now working across all the other vectors of compute and efficiency."

For now, that's starting in Intel's mobile chips. It sure sounds like some of Meteor Lake's tech is coming to desktop soon, though. Rogers said that Intel will "share more later in the year" regarding moving the architecture into other computing segments.

The chicken and the egg

So you have an AI PC. Now what? 

That's the "chicken and the egg" problem of NPUs. They're out there, and now developers need to make software work with them.

I gave Rogers the case of my brother, who doesn't have a very complicated use case for work: he does emails; he does spreadsheets; he's got documents. He's still on Windows 10. Would he even notice?

"For sure, I would guess your brother is watching YouTube and Netflix," Rogers suggested. (He's right on YouTube. He doesn't have Netflix.) "Quite likely, he's making a family call or a business call with Zoom or Teams or WebEx, in these popular applications."

And these, Rogers tells me, are "much better" on Meteor Lake laptops. Video effects in Zoom, for instance, will be offloaded from the GPU to the NPU, helping to save power, leading to longer battery life. That, along with the "low-power island", including efficiency cores directly incorporated into the SoC tile, should save power.

"That-low power island is a really important technology for us, and we spent a ton of time in software tuning that to make that seamless, so that your brother doesn't necessarily need to know how all the bits are actually working there, but in the end, web browsing and video playback both locally streaming, just longer battery life, just better battery life, right? And that's all with similar performance to what we had in Raptor Lake."

What I got from this was that to start, the best AI features might not be ones you actually notice. Do you care if you know that your applications are using AI or an NPU as long as you're getting some sort of tangible benefit? Maybe not!

But Intel is promising that over 100 software companies are working with it to put out features that are optimized for the NPU. So when will we see those features come that really push people to new laptops with new CPUs?

Well, Intel has demos of Llama 2-based applications — I saw one at Intel's showcase — and Rogers expects to see LLMs as one major way app developers will use the NPU locally.

"My expectation is we're going to see not just one, but tens and hundreds of these from various ISVs, preloaded with various OEMs," he said. "And there's a ton of energy going on on this and people looking to run these both on the GPU and the NPU."

But Intel suggests it has another benefit on its side: scale. Rogers said that a "giant developer ramp" is required to get coders working on these types of projects, and that "in many cases, we have developers coming to us" for guidance with tools and frameworks.

How do you benchmark PCs on AI?

I test hardware for a living. So I had to ask, how did Intel decide how to test Meteor Lake on AI? The claims are bold, but there are only so many workloads to pick from right now.

"We're talking about the PC here right? So of course we want to benchmark it, right? That's just how we handle understanding PC technologies."

Rogers said his team went in looking to explain how the NPU compares to say, running a workload on the GPU. But it ends up that much of the benchmark, when it comes to AI applications, is really about something else: software.

"What we realized really quickly is we're not really benchmarking the hardware — we're actually benchmarking the code paths," he said. "We're benchmarking the software, which is really interesting, right? Like that's very different than what we see today in the CPU and the GPU."

That, Rogers said, is why Intel shared a chart showing that Meteor Lake runs all sorts of AI workloads, and does so "performantly." (AMD and Qualcomm's offerings were also listed, with lesser degrees of success.)

And benchmarking AI tools isn't cut and dry in other ways, Rogers said. "What's happening right now is okay, how big of a model can I run? And that's a question of computing power, how do I quantize that model, how do I minimize not just the compute but also the memory footprint and the memory bandwidth requirements which are enormous for many of these."

This isn't a hardware story, not entirely, Rogers said. You can come back to an improved model in weeks, he said, and see massive gains. This is also about software — or at least software optimization.

Goodbye, P-Series! Hello, H-Series?

Intel has simplified its naming scheme a bit with Meteor Lake, removing the "P series," which has only been around for a couple of generations, and wrapping that into the long-existing H-series. Now, there's just H and U.

H series is the larger chip, with a 6+8 die (6 P-cores, 8 E-cores, though there are processors with 4 P-cores), while the U series is designed for 2+8. 

Perhaps the biggest difference is that the H series is running at 28W, which the P series ran at. On Raptor Lake, the H chips ran at 45W. When the Intel Core Ultra 9 185H launches in early 2024, that will be a 45W part. So Intel still has a range of TDPs. It's simpler in naming, but not inherently in power.

"In [the] Meteor Lake generation it will be largely what was our P series, so that thin and light performance — thin and light concept — packing a lot of power into an integrated graphics system," Rogers said.

But that means that people could, in theory, go into a store and think they're getting a different amount of power. How, I asked, will Intel make it clear what you're getting?

"Part of improving communication is simplifying, so we definitely got the feedback and we wanted to just simplify the number of choices," he said. So that's why we made it from a three-letter structure to a two-letter structure which we think is helpful."

Rogers said that the decision will really be made when a customer goes to a store and looks for what they want to buy. That messaging, he said, will come from the company that makes the laptop and what they offer in their lineup.

"And then as a user, you're just able to kind of see, 'OK, what are my needs,' right? Do I want lighter, do I want a little bit more performance? So we have a letter that gives that choice, but really that's a choice that you're going to see in the exact laptops that you're interested in purchasing. So if you know that you're looking for a more powerful [PC], lots of those laptops generally will have H-series."

9W parts and "very different" form factors

Intel didn't have a low-wattage part available in 13th Gen mobile, which led to 12th Gen chips in foldables released this year. In Intel's list of chips coming in Q1 of 2024, there's the Core Ultra 7 164U and Core Ultra 5 134U, both of which are 9W processors. So what new form factors could Intel be trying to enable by going back to low-power parts?

"Without pre-announcing from my friends, colleagues, and partners I think you will see some, notable, very different form factors for Core Ultra, not necessarily in the PC space," Rogers said, pausing.

"It's a PC, but in a different category, let's say," he resumed to finish his thought. I couldn't help but think that every major gaming handheld so far has used AMD Ryzen processors. There's also Lenovo's rollable prototype from Mobile World Congress. At the very least, we could get more modern foldables.

But most Meteor Lake designs will be two-in-ones, clamshells, and other more traditional shapes, he said. "Most of the focus of our partners is actually driving more compute into thinner, lighter envelopes. Thermal technologies have gotten a lot better." Rogers suggested that the increased thermal capacity means you can put a 28W part in a system you used to put a 15W CPU into.

And because AI can be computationally demanding, Rogers predicted that you'll see more computers with the higher TDP choice (the 28W H-series) rather than the lower-power 15W U-series chips, though both will be available.

"At the end of the day, OEMs make the right decision based on their form factor, but we'd like to provide choice," he said. "And I think the biggest thing that I'm most excited about in Core Ultra, for me, is the battery life improvements." All of the data Intel showed at launch, he pointed out, was on the 28W part.

We'll see what Intel and laptop companies come up with over the next year. While a handful of devices from MSI, Lenovo, Dell, HP, Asus, Acer, Samsung, Gigabyte, and more launched this week, CES 2024 is right around the corner. And what would a new year be without a slew of new laptops? We'll see then.

Microsoft will launch Windows 12 in June 2024, claims a report published by the Commercial Times today. The business-focused Taiwanese newspaper highlighted this date as very important to the island’s PC industry sector, sparking significant sales of a new wave of AI PCs. Leading industry figures from companies like Acer, Quanta, MSI, and Gigabyte were all said to be very excited about “the first year of AI PC” and the business opportunities it will herald.

Though the Commercial Times report leads with its statement about Windows 12 and the purported June release date, no direct quotes from industry leaders are given in support. It implies the information emanated from recent statements by Acer Chairman and CEO Jason Chen and/or Quanta Chairman Barry Lam.

Both of these important executives from the world of PCs were at the Taiwan Medical Technology Exhibition in Taipei on Thursday. Acer’s Chen was reportedly cautiously optimistic about AI PCs contributing to "continuously accelerate” the industry through a virtuous cycle of AI PCs, new AI app development, new AI PCs, and so on.

Quanta’s Lam was more bullish. He highlighted an expected upturn across three segments for Quanta - AI PC, AI server, and AI automotive electronics. The Commercial Times adds that Lam expects “next summer, when Microsoft launches a new generation of Windows operating systems, AI PCs will also be launched one after another.” However, that doesn't appear to be a direct quote from Mr Lam.

If the June 2024 launch of Windows 12 happens, it will be earlier than a few previous clues have indicated. In March, we reported on an Intel Meteor Lake desktop processor (MTL-S) leak, where the as-yet-unreleased Windows version was apparently listed as a supported operating system. At the time, we pondered over Windows 12 being likely to launch in H2 2024, given its recently adopted three-year cycle. Also, in May, we reported that Microsoft was preparing a new own-branded (Arm?) CPU for Windows 12 devices.

Intel Meteor Lake chips will be important to the mass adoption of accelerated AI due to dedicated processing hardware on all SoCs in this family. Meanwhile, hardware AI acceleration on the latest AMD Ryzen 7000 family offers is less clear-cut. Only Ryzen mobile 7040 ‘Phoenix’ chips have the XDNA architecture hardware to accelerate local AI processing. This will hopefully change, as there are signs of Phoenix coming to desktop, and of the Ryzen 8000 family chips on the way to mobile in the coming months. All this raises the specter that hardware AI processing support could be one of the minimum spec requirements for Windows 12.

Just ahead of the weekend, Asus co-CEO Samson Hu talked about his company’s plans to roll out the first AI PCs in 2024. The executive told attendees at the annual technology forum at National Cheng Kung University in Tainan, Taiwan, that AI PCs should carry a price premium over standard PCs. However, Hu was vague about what would differentiate an AI PC – hardware, software, or possibly a mix of both. Today Asus also released its 2023 Q3 financials, and the investor slide deck suggests it already has big plans for AI PCs over the coming months.

We think that two major tech events are going to collide in 2024, and give birth to high-performance AI PCs. Microsoft has already started rolling out Copilot AI features for Windows users, and adding AI to several of its popular applications. Meanwhile, early 2024 should see the first availability of Intel Meteor Lake systems, with a dedicated NPU tile as part of the processor – accelerating local AI processing. Let us also not forget AMD is expected to release desktop APUs with AI acceleration in January 2024, too.

The Asus co-CEO told forum attendees that there are about 300 Windows developers working on AI apps that will benefit from local AI acceleration. This will be another important driver of sales of AI PCs, which are expected to initially be targeted at enterprises and content creators.

However, Hu wasn’t firm in providing a definition of an AI PC. The first wave of this new class of PCs could feature processors that feature on-board AI acceleration. However, some might offer AI PCs that are simply capable of running Microsoft365 Copilot, for example. Different PC vendors, many of which now have AI PC plans in place, will outline different definitions, reckons Hu. Nevertheless, the Asus co-CEO reckons an AI PC should feature more advanced processor families and bigger memory quotas.

Asus sees 2024 as a year that sets a foundation for AI PCs, with a mere single-digit percentage market penetration until 2025 or later. Despite these modest expectations, a large slice of today’s investor presentation stoked enthusiasm for AI PCs. One slide said Asus was “redefining the PC with AI,” and another that it was looking forward to “rapid technological advancements in the era of AI PC.” Moreover, in its business highlights it confirms it is “targeting to introduce the world’s first AI PC.” Outside of consumer and creator AI PCs, we can also see Asus reckons AIoT, AI Server, and AI HPC are all drivers for continued growth.

Qualcomm is taking another stab at an Arm PC system-on-a chip, this time based on the new Oryon cores that come from its purchase of Nuvia. The new chip, the Qualcomm Snapdragon X Elite, will launch in laptops in mid-2024 and was unveiled today at the company's annual conference in Hawaii.

The chips are the latest volley in the attempt to successfully move more Windows users onto Arm. While the operating system support is there, there have been performance and compatibility issues with the Snapdragon 8cx chips. Now, Qualcomm says that Snapdragon X Elite will deliver far better efficiency with equivalent or better performance than competitors (a claim we'll have to test next year).

"We believe that this is like a pivotal moment for us," Kedar Kondap, senior vice president and general manager of compute and gaming at Qualcomm told Tom's Hardware in an interview. "Because what you will see with the Snapdragon X Elite platform is something you haven't seen before in terms of leadership, in terms of technology, performance per watt. You'll see the Qualcomm Oryon CPU distinguish itself in terms of its capabilities and the efficiency that it will drive."

Qualcomm’s chip, using TSMC's 4-nanometer process on a single die, will have 12 "high-performance" Oryon cores on each chip going up to 3.8 GHz. However, of those 12 chips, the company says single- or dual-core boost up to 4.3 GHz is possible. The cores are batched into three clusters of four on the die. 

The chip will have up to 64GB of LPDDR5x RAM, with up to 136 GB/s of memory bandwidth, and 42MB of total cache.

Qualcomm is pushing the chip as having "best-in-class" multi-threaded CPU performance. The company showed journalists charts of performance compared to a competitor's 10-core and 12-core laptop chips, showing up to twice the performance in Geekbench 6.1 at one-third of the power. (Qualcomm didn't name the chips in its presentation, but later told me the 10-core chip is a Intel Core i7-1355U and the 12-core chip is a Core i7-1360P). Compared to a 14-core chip on the same test (later clarified as the Intel Core i7-13800H), Qualcomm showed significant improvements in performance per watt, and it suggested the Snapdragon X Elite gets up to a 60% performance improvement.

The company also claims a "50% better peak multi-thread performance" against an "ARM-based competitor." Though it didn't say which competitor, in a briefing Qualcomm suggested it was the "latest M2," presumably the standard Apple M2. That slide doesn't even have an accompanying chart, making Apple's briefings looks deeply technical in comparison.

All of this is to say that while it sounds exciting, we'll need to test a laptop with one of these chipset to see how it really performs.

The integrated graphics consists of Qualcomm's Adreno GPU, which it says delivers up to 4.6 Teraflops of performance. It supports a laptop screen up to 4K at 120 Hz with HDR 10, with external support for three 4K monitors or dual 5K displays. (Apple's M2 Pro only supports two external displays, while the M2 Max can handle four, depending on resolution and refresh rate).

Again, Qualcomm used unnamed competitors in its comparirons, but later clarified when asked specifically that "Competitor A" is the Intel Core i7-13800H, while "Competitor B" is the AMD Ryzen 9 7940HS. It claims the Snapdragon X Elite's GPU offers approximately twice the performance of the competition in 3D Mark WildLife Extreme at one-fourth the power, while it claimed 80% improvement at one-fifth the power against a different integrated GPU.

Qualcomm isn't listing a specific TDP for the Snapdragon X Elite. I've been told that it "dynamically scales performance across various TDPs," which will seemingly depend on what laptop manufacturers choose to do with it. The company suggested that the chip will be in several form factors, including clamshells, 2-in-1s, and both fanless and actively cooled machines, the latter of which would almost certainly provide the strongest performance.

All About AI

Possibly because everything is about AI now, or possibly because Qualcomm has been building AI features for smartphones for a bit now, it's no surprise that Qualcomm is pushing the Snapdragon Elite X as a chip meant to handle generative AI tasks.

"You will see the NPU, or the intelligence in the PC, get elevated to a level that no one has ever seen before…" Kondap said. "What we're focused on is taking all of the models that we all hear about, like Llama 2 and stuff, making sure they drive the best experience, and then enabling this whole ecosystem of applications as well as partnerships that ISVs [independent software vendors] and partners can take advantage of this NPU to drive a much better experience."

Qualcomm claims you can run 13 billion parameters on device (it specifically called out Meta's Llama 2 open source language model). It also suggested it could run a 7-billion parameter model at 30 tokens per second. Qualcomm says its AI Engine has a 2.5 times faster tensor accelerator than previous laptop chips, with twice as much shared memory. Qualcomm is suggesting the Snapdragon X Elite's NPU is operating 45 trillion operations per second (TOPS).

Kondap predicts that enterprise may be among the first to embrace AI, and that models running on device, as opposed to in the cloud, will be important for security.

Another smaller NPU, the "Micro NPU" is part of another part of the chip that Qualcomm calls the "sensing hub," which includes… well, the sensors, including cameras, microphones, and also connectivity. This is the part of the chip controlling the Wi-Fi module (including Wi-Fi 7 support), and camera features (Qualcomm says it's an "always-sensing ISP").

Qualcomm states that the new chip supports Windows Studio Effects, but that's table stakes. Heck, Qualcomm already did that on Surface devices. The company also promises support for "many other AI-accelerated applications and experiences." The bet, then, is that many more AI apps are going to come with a focus on local performance.

Efficiency, we're told

For Apple, switching from x86 to Arm brought huge gains in efficiency. That's been the promise for Windows on Arm, but it hasn't materialized yet. Qualcomm is calling the Snapdragon X Elite's efficiency "game-changing."

In a press release, Qualcomm claimed that the chip is efficient enough to support "multiple days of battery life." I suspect that includes time with the device asleep. The company hasn't ball-parked any numbers. On the one hand, that makes sense: Battery life depends on the device, including the size of the battery, as well as other parts like the display. On the other hand, we'll have to wait quite a while to see how long these chips last on our battery test.

Over the summer, reports popped up that Qualcomm was requiring the use of its power management integrated circuits (PMICs) with the Oryon-based processors. Kondap said that Qualcomm's PMICs are "an option for OEMs to choose," and that the company has been providing them for "many years now."

"We come from a legacy where, for us, power and performance per watt is very important in how we architect our platforms…" he said, pointing out all of the Snapdragon Elite X's features. "We control the power within our SoC very tightly, right? So we have a lot of, for example, dedicated power rails on how we manage power across different use cases and how this is managed. And there's a reason why we spend a lot of time thinking about this first, because it obviously matters, right? If you go back to any consumer and ask what's the most important thing to them, they will say battery life."

Snapdragon X Elite laptops and Windows compatibility

The Snapdragon X Elite will show up in computers from "leading OEMs" in mid-2024, according to Qualcom''s press release. As of this writing, there haven't been any announcements about who Qualcomm's partners are and what kinds of devices they plan to build. I'm told we'll hear more about partners during the Snapdragon Summit, though product specifics may be further out.

"You'll start to see devices drive a good, premium experience," Kondap said. "So you'll see us in multiple form factors, right from whether they're thin and light fanless designs all the way to detachable laptops to clamshell. So you'll see us in a variety of form factors over time with the Snapdragon X Elite." He notes, however, that Qualcomm is "obviously not targeting high end gaming as a category with X Elite right now."

Kondap suggests, that with partnerships with Adobe, Microsoft, and Da Vinci, that "casual content creation" will be a focus for the laptops, and perhaps "casual gaming."

Perhaps the biggest worry I've had with Snapdragon-based laptops before is app compatibility. Some software has never been made native, and I've seen issues with games and anti-virus. Kondap tells me that in partnership with Microsoft, "significant" progress has been made.

Part of this, Kondap suggests, is due to the Windows on Arm Dev Kit (previously known as Project Volterra). "We actually have lots of developers working on these platforms, optimizing specifically to porting their applications on Snapdragon," he said.

That is the second part of a three-step process, he suggests. The first is that Qualcomm and Microsoft get the emulation up to snuff so that any app you open works. The second is getting developers to make native apps for Arm and Snapdragon. Third is getting developers to use the NPU, Oryon cores and other parts of the new chip to take full advantage of it.  He said the emulation is looking "very promising," and that Qualcomm is going after a long list of applications - perhaps the top 200 to 300 apps. Of course, he points out that Microsoft has made its apps native, and that more anti-virus and VPNs, previously problematic apps, are becoming native.

We'll see how well it all works when these laptops are released in 2024.