NVIDIA released an updated version of its open-source vk_lod_clusters sample featuring a refreshed 1.6-billion-triangle Zorah scene rendered with path tracing and clustered raytracing. The new glTF scene data totals about 32 GB, down slightly from 36 GB last year, despite adding a full attribute set and textures, largely because processing now relies on meshoptimizer's clusterlod.h module. The author, who previously documented processing the original Zorah scene, has since contributed improvements to meshoptimizer's hierarchical clustered LOD support and is writing about the updated pipeline's other aspects beyond processing speed.
Nvidia stock climbed near its all-time high of $236.54 set in May, having risen 15% over the past three months even as other AI-related stocks lagged. CoreWeave separately announced plans to deploy Nvidia's upcoming Vera Rubin chip platform in its data centers. Micron shares also rose ahead of its Wednesday evening earnings report, with investors anticipating strong results driven by tight memory supply.
Former EVGA product manager Brendon Ray Hedrick published an account describing how Nvidia's 2016 Founders Edition cards put it in direct competition with board partners like EVGA. He also says Nvidia required partners to sell at least one model at the announced MSRP even at a loss, threatening reduced chip allocations for those who didn't comply, forcing EVGA to offer loss-leading cards alongside pricier premium versions.
A new open-source project called OpenDLSS reimplements the neural network behind Nvidia's DLSS 5, reproducing the same 71-block transformer architecture used in DLSS-NR build 310.8.0, down to matching intermediate outputs at all 75 block boundaries. The project runs on Vulkan using FP8 tensor-core computation, and includes a separate WebGPU browser port that reproduces identical results without tensor cores or FP8 support. Users must supply their own copy of Nvidia's trained model weights to run it.
President Donald Trump convened a meeting of leading US technology executives at the White House, describing it as a gathering aimed at advancing 'America's Golden Age' in AI. Attendees included leaders from OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google. The meeting took place amid growing calls from some industry figures and experts for stronger regulation of AI's potential risks.
AMD and World Labs announced AMD will acquire the AI startup by the end of the year, pending regulatory approval, in a deal valued at $8.2 billion. World Labs, founded in 2024 by Fei-Fei Li and colleagues with $230 million in funding including from AMD, builds world models such as Marble and Atlas that generate 3D scenes from video data.
Eric Gullichsen, who served on Nvidia's Technical Advisory Board from 1993 to 1996, says he was granted 25,000 stock options but only 15,625 vested before he left, due to a discrepancy over whether his options were meant to vest quarterly or annually. He argues that after 30 years of stock splits, the missing 9,375 options would now be worth roughly $1 billion, and he is disputing Nvidia's account of the vesting terms.
Nvidia launched the Open Agent Safety Platform, a consortium of over 100 companies aimed at preventing rogue AI agent behavior, with Anthropic as a supporter but OpenAI, Amazon, Google and Apple notably absent from the list. An OpenAI spokesperson told TechCrunch the company supports Nvidia's effort and is already collaborating with Nvidia on agent security tools, including OpenShell, an open-source sandboxing system meant to contain agents.
President Trump and House Speaker Mike Johnson hosted a lunch at the White House with leading technology executives to discuss artificial intelligence. A seating chart shared by Trump showed Nvidia's Jensen Huang and Elon Musk seated beside the president, with Meta's Mark Zuckerberg and Google's Sundar Pichai nearby, while Jeff Bezos sat opposite alongside Vice President JD Vance. Palantir CEO Alex Karp told reporters outside that tech leaders must take responsibility for known AI risks and that Americans don't want separate rules for the industry versus the public.
Nvidia announced a new $150 billion share repurchase authorization on Monday, adding to the $80 billion buyback plan it revealed in May alongside a dividend increase to 25 cents per share from 1 cent. The move comes as Nvidia's forward price-to-earnings ratio has fallen to 14.5, well below its five-year average of 62.9, even as analysts project net income near $385 billion in fiscal 2028.
Eric Gullichsen, a former Nvidia Technical Advisory Board member, says a 1993 stock option grant he re-examined in 2024 shows a one-year vesting schedule rather than the four years he was told, meaning 9,375 additional shares should have vested. After accounting for Nvidia's stock splits, he calculates this discrepancy would now be worth roughly $1 billion. He says he and his lawyer concluded the statute of limitations likely barred any legal claim.
An excerpt from Kevin Roose's upcoming book The AGI Chronicles describes a contentious May 2022 dinner between Nvidia CEO Jensen Huang and Anthropic CEO Dario Amodei over Anthropic's chip needs. The book alleges Huang called Anthropic's compute chief a 'bean counter' after being shown data favoring Google's TPUs, and that Amodei later described Huang as 'kind of Trump-like' for repeating himself. Nvidia disputes the 'bean counter' remark, and despite the rocky start, the two companies eventually formed a partnership.