Nvidia plans stronger fp64 performance for next gen high performance computing gpus
Nvidia is reaffirming its commitment to 64-bit floating point performance in high performance computing, signaling that upcoming architectures will restore and enhance fp64 capabilities after recent generations prioritized lower precision throughput.
OpenAI debuts GPT-5.2 on NVIDIA infrastructure for large scale training
OpenAI has launched GPT-5.2, describing it as its most capable model series yet for professional knowledge work, trained and deployed on NVIDIA’s full stack Artificial Intelligence infrastructure. The release underscores the growing importance of massive pretraining and post-training at scale using thousands of GPUs.
SMIC reaches 5 nm N+3 volume production without EUV tools
SMIC has moved into volume production of its 5 nm-class N+3 node using only deep ultraviolet lithography, a full generation ahead of its earlier N+2 process, but yield and cost issues remain significant.
OpenAI GPT-5.2 launch underscores NVIDIA’s role in large-scale Artificial Intelligence infrastructure
OpenAI’s new GPT-5.2 model for professional knowledge work was trained and deployed on NVIDIA’s latest accelerators, highlighting how frontier Artificial Intelligence systems increasingly depend on NVIDIA’s full-stack infrastructure and benchmark performance gains.
A decade of OpenAI and the road to superintelligence
Sam Altman reflects on OpenAI’s first ten years, from uncertain beginnings to systems that outperform top humans on difficult intellectual tasks, and lays out a confident path toward superintelligence over the next decade.
Global artificial intelligence safety rules tighten across eu, us and uk in 2025
Governments in the eu, us and uk are shifting from voluntary principles to binding rules for high-risk artificial intelligence systems, reshaping expectations for developers and experimental platforms alike.
Artificial intelligence hallucination benchmark compares popular language models
Researchers benchmarked 37 large language models on 60 fact-focused questions to measure hallucination rates and found that context window size and model scale are poor proxies for reliability.
Microsoft Foundry expands Azure model catalog with GPT-5.2, partner models, and granular regional options
Microsoft’s Foundry Models lineup in Azure now spans the latest Azure OpenAI families such as GPT-5.2 and GPT-4.1, as well as partner offerings from Black Forest Labs, Cohere, DeepSeek, Meta, Mistral, Moonshot AI, and xAI. The documentation details capabilities, deployment types, and region-by-region availability for standard, provisioned, batch, and data zone scenarios.
Time names architects of artificial intelligence as 2025 person of the year
Time magazine has named the architects of artificial intelligence as its 2025 Person of the Year, highlighting the technology’s sweeping cultural and economic influence and featuring interviews with several prominent industry and government figures.
NVIDIA designs location verification technology to track GPU smugglers
Reuters reports NVIDIA has developed a software-based location verification tool that identifies where and how its GPUs are deployed to curb GPU smuggling and enforce export restrictions. The company says the service will help data center operators monitor the health and inventory of their Artificial Intelligence GPU fleets.