Industry news live: the latest news from Nvidia, Intel, and AMD
What's going on in the world of computing
New AMD Ryzen Threadripper smashes PassMark record
Benchmark results for AMD’s 64-core Ryzen Threadripper 9980X show record-breaking multi-threaded performance, surpassing even the 96-core Threadripper Pro, but relatively modest single-threaded scores.
Nvidia's newest top-tier AI supercomputers deployed for the first time
CoreWeave deploys first Dell GB300 cluster with Switch: Up to 1.1 ExaFLOPS of AI inference performance per rack.
Nvidia RTX 50 Series appears in Steam Hardware Survey for the first time
Despite being out for a few months at this point, Nvidia's notorious RTX 50-series has failed to show up in Steam's internal stats — up until now, when they've finally appeared with a 3.69% share.
Nvidia's Arm chips rapidly gain share in server market as AI booms
The global server market soared 134% to $95.2 billion in Q1 2025 as hyperscalers ramped up AI infrastructure, with Arm-based Nvidia GB200 systems driving growth and annual spending projected to exceed $366 billion.
Nvidia's incendiary 12VHPWR connector consumes its latest victim
Wherever the 12VHPWR connector goes, it brings death with it. The latest in the series of murders is a measly RTX 3060 Ti from China.
Intel might axe the 18A process node for foundry customers, essentially leaving TSMC with no rival
Intel CEO Lip-Bu Tan is weighing whether to stop offering the company's 18A technology to external clients and instead focus on the next-generation 14A process, a move that could withdraw Intel from the broad foundry market for several years.
Intel jumps to HBM4 with Jaguar Shores, 2nd Gen MRDIMMs with Diamond Rapids
SK hynix has revealed that Intel's upcoming Xeon 'Diamond Rapids' CPUs will adopt 2nd Gen MRDIMMs and its next-generation Gaudi AI accelerator 'Jaguar Shores' will feature SK hynix HBM4 memory.
AMD Krackan Point 2 APU emerges on Geekbench
An AMD Krackan Point 2 APU Geekbench test result has appeared, featuring a single Zen 5 core and three Zen 5c cores.
DG1 defaults
Better Late Than Never: Linux 6.17 To Enable @IntelGraphics DG1 Graphics By Defaulthttps://t.co/mGC4lUVbNHJuly 2, 2025
Linux 6.17 will now enable Intel DG1 graphics by default, nice!
Nvidia's Upcoming 'Double D' gaming GPU becomes the '5090D V2' in latest leak
The RTX 5090DD, intended for release in China, has been renamed to the RTX 5090D V2 while the specs remain untouched.
Nvidia's 16GB RTX 5060 Ti reportedly 16x more popular than its 8GB variant
The 16GB variant of the RTX 5060 Ti reigns supreme against its 8GB peer with a whopping 16x difference in sales figures.
Nvidia RTX 5050 benchmarks reveal close contest with RTX 4060
Benchmarks by Inno3D show the RTX 5050 performing nearly on par with the older RTX 4060.
Nvidia to axe Maxwell, Pascal, and Volta GPUs with end of driver support
An official Nvidia Unix graphics feature deprecation schedule says that driver branch 580 will be the last to include support for the Maxwell and Pascal architectures.
RTX 5000 graphics cards get deep discounts at Newegg
Newegg is hosting a blowout sale on Nvidia GPUs, with the RTX 5070, 5060 Ti, and 5060 all falling below MSRP.
AMD's massive GPU VRAM on its Instinct cards has broken Linux's hibernation feature
AMD AI Linux-powered servers are failing to hibernate due to excessive VRAM and a high number of AMD Instinct accelerators per system.
Nvidia is giving away free Adobe Creative Cloud subscriptions with access to all software
If you've got an RTX 50 series GPU from Nvidia, you can now access two months of free Adobe Creative Cloud, including all of its most popular tools. Got a 30 or 40 Series? You can still get one month.
Not sure what AMD was thinking giving it Dual-channel RAM, but leaving it with 2CU, and 4x PCIe 3.0 lanes, and then selling it along side new old stock 5700U/5825U that end up costing roughly the same for the laptop/mini-PC.
Intel has the right idea with Gracemont/Twin Lake.
Single channel RAM, but 9x PCIe 3.0 lanes to play around with in a variety of configurations.
The problem is that AMD is allergic to selling budget products at budget prices, especially mobile-focused APUs. They also probably didn't have, or weren't willing, to make Mendocino in the wafer volume it would take to match Alder Lake-N's rock bottom prices.
I would rather see dual-channel in Atom chips, as it had been for Jasper Lake and prior gens. Maybe that will come back with Wildcat Lake.
A Wildcat Lake leak says it will have PCIe 4.0 x6. That's faster than PCIe 3.0 x9, but with less lanes, so it could limit the number of devices that can be used. However, it's possible that there are PCIe 3.0 lanes in addition to PCIe 4.0 that the leaker failed to specify.