Eight machines from seven makers went on sale on 16 October, each built on a single Grace-Blackwell superchip designed with MediaTek.
Company
NVIDIA
NVIDIA is an organisation working in artificial intelligence, founded in 1993 and based in Santa Clara, United States. This page collects every Aivio News story about NVIDIA, together with its models and recorded benchmark results.
- Founded
- 1993
- Headquarters
- Santa Clara, United States
- Official website
- nvidia.com
Models
All storiesMore from NVIDIA
The site was Nscale's only announced customer for that building. Nscale says its wider relationship with Microsoft is unaffected and that it is looking at other power sources.
Two-nanometer parts reached 8% of wafer revenue in the quarter. Chief executive C.C. Wei told analysts the company is working very hard to narrow the gap between demand and capacity.
Today an unauthorised shipment is an administrative violation, and prosecutors reach for forgery charges instead. Under the planned change it becomes a crime under the Foreign Trade Act.
No models, launch dates or financial terms were given. The announcement names Level 4 and robolimousines as destinations, not products.
The launch was announced on 15 October, months after the machine started work. Jensen Huang attended the event in Stockholm.
- Microsoft has finished pre-training a 2.8-trillion-parameter model, and the efficiency headline is a fitted estimate
- Perplexity says nine models failed to break out of its sandbox 118 times, and its own report counts 108 runs
- Black Forest Labs' new image model has 5.6 billion parameters and an 8GB build that runs on a consumer graphics card
- Nvidia's new open model activates 12 billion of its 120 billion parameters, and the small one is quantised down to 24GB
- Nvidia is giving UK agent-safety labs $250m of compute over five years, and the provider has not been chosen yet
- Nvidia's share of China's data-centre AI chip market has fallen to 34%, and Huawei is six points behind
- Four rivals have written down what an AI data centre must do when the grid wobbles: come back to near-full load in under two seconds
- Amazon is reported to be in talks to buy Decart for about $7bn, two months after Anthropic walked away
- Microsoft's third in-house AI chip carries 288GB of memory, and 96 of them share one 167kW rack
- A Seoul paper says Samsung's HBM4E passed Nvidia's quality tests, and Samsung says it cannot confirm that
- About 2,100 servers were publishing their GPU telemetry to anyone who asked, and none of them required a password
- Liquid AI's 3B classifier answers in 16 milliseconds on a Jetson, and its weights are downloadable
- The White House now treats the words "artificial intelligence" as a loyalty test, and no one has said what happens to people who fail it
- Investors would not pay three times August's price for an AI data centre operator, so Firmus pulled a $5bn listing on the day
- OpenAI's run rate is $50bn, not the $70bn that went round last month, and the gap is whose cloud sales you count
- Every AI infrastructure target at Pwn2Own was broken, and the bugs were the same injection flaws the web had twenty years ago
- TSMC's revenue is up 55% on a year ago and slightly down on last month, which is what a supply ceiling looks like
- An Nvidia-backed startup is selling a way to mix chip vendors in one data centre, and part of it runs on Nvidia's own networking
- TSMC is paying a rival $2bn to make the silicon slabs its AI packaging sits on, and none of it ships before 2028
- A 9.8-rated hole in the cache layer behind vLLM has no patch, and whether you are exposed comes down to one flag
- A three-year-old lab says its open agent drives 2.5% of the world's AI tokens, and investors have valued it at $1.5bn
- Broadcom, Oracle and SpaceX are all trying to borrow for AI chips from private credit rather than the bond market
- Nvidia's chip is now in a Windows laptop with 128GB of unified memory, and Microsoft will take your MacBook in part exchange
- Windows now ships a box to run AI agents in, and Codex, Copilot and LM Studio are already using it