Nvidia details Vera CPU as it expands into AI server chips

According to CNBC, Nvidia released new specifications for its data center CPU, Vera, and said the chips were delivered in June to clients including OpenAI, Anthropic and SpaceX. The company said Vera is its first server CPU designed from the core, rather than built on an off-the-shelf Arm design, and that it is intended to address bottlenecks in AI agents. Nvidia said Vera offers 50% better performance for AI agents than x86 chips and will be sold on its own, paired with its GPUs, in a liquid-cooled rack of 256 chips, and in a Vera Rubin system. The company also said OpenAI plans to deploy Vera chips in large quantities starting this quarter. Analysts cited in the article said the move could challenge Intel and Advanced Micro Devices in server CPUs, while Nvidia said the overall server CPU market could eventually reach $200 billion.