Archive, page 4
DLSS is Nvidia's AI upscaling and frame generation technology for games. GameGPU reports that natively integrating DLSS 5 into game engines could raise performance by up to 60%.
RTX Mega Geometry is an Nvidia graphics technology for rendering highly detailed geometry in real-time 3D games and applications. Nvidia's page describes its features.
AMD will raise GDDR6 memory prices for board partners starting October 1, while Nvidia reportedly holds off on GDDR7 price changes, according to Wccftech.
The Tech Buzz reports that more than 300 neocloud companies, which rent out GPU computing capacity, have emerged to meet surging demand for AI compute.
A Futu report says Nvidia ranks first in China's smart driving chip market, with Horizon Robotics holding second place as the market share gap narrows.
A CNBC explainer outlines the options companies have for accessing Nvidia GPUs, including buying hardware directly and renting compute from cloud providers.
Microsoft and Nvidia are joining forces to reshape the PC, with Jensen Huang saying AI agents will become the main actors in computing, BigGo Finance reports.
SpaceX reportedly seeks $40 billion to buy Nvidia GPUs for orbit-based AI, after Elon Musk backed Nvidia and said SpaceX will exclusively adopt its GPUs. Reports are unconfirmed.
Reports say Nvidia will end RTX 5090 production and reserve GB202 silicon for its more lucrative RTX PRO workstation cards, possibly replacing the RTX 5080 16GB with a 24GB version.
BGNES reports that the price of renting Nvidia H100 GPU computing power has risen 60% over the past year.
Wccftech and eTeknix report Nvidia's Vera CPU brings 2,000 sandboxes online faster than AMD EPYC Zen 5 can boot 1,000 for agentic AI workloads.
Reports say Chinese authorities may allow Alibaba, ByteDance and DeepSeek to buy Nvidia H200 chips, but supply may be tightly capped. Report, not confirmed policy.
Seismora Inc. is developing networking software that coordinates AI workloads across devices, edge and cloud providers, deciding where each task runs. Still in development.
Review of GMKtec's EVO-X3, a Ryzen AI Max+ 395 mini-PC with 128GB memory and an OCuLink port, positioned as a local AI workstation.
TechCrunch profiles Danu Robotics and founder Amy Ma, who has spent six years building a better robot for sorting recyclable waste.
CoreWeave is focusing on AI inference, the workload that will decide AI economics, by adding storage, networking and software to its GPU cloud to serve models faster and cheaper.
Nvidia has committed $1 billion to US scientific AI supercomputing, as the US government prepares AI-optimized supercomputers for national labs.
Nvidia describes how developers combine frontier AI models with Omniverse libraries to assemble assets, connect physics and rendering, and build simulation applications.
CoreWeave argues that the AI cloud market is about tooling, partners and expertise as well as GPUs, for improving models and agents in production.
Nvidia has released a fix for a high-severity flaw, tracked as CVE-2026-47483, that could crash GPU monitoring on exposed servers.
TSMC picked GlobalFoundries to produce silicon interposers for its CoWoS advanced packaging in the US, in a $2 billion deal. The packaging is used in AI datacenter chips.
Finnish authorities ordered Google to stop work on two AI data centers over alleged deforestation, saying they cannot conduct a proper environmental assessment after the site was already altered.
Ukrainian drones struck a Yandex data center in Sasovo that houses two Nvidia-powered Russian supercomputers. The fate of the systems and AI training in Russia is unknown.
Samsung is eyeing an $80 billion quarterly profit as AI demand lifts chipmaker margins, while cheaper PCs and phones become harder to produce profitably.
AMD's EPYC Verano AI host CPU will reportedly use a special SB1 socket. Dynatron also unveiled an air cooler for Verano CPUs aimed at AI servers.