NVIDIA has delivered the first units of its Vera CPU to three leading AI labs and a major cloud provider. The chips arrived at Anthropic in San Francisco, OpenAI in Mission Bay, and SpaceXAI in Palo Alto on Friday, followed by a delivery to Oracle Cloud Infrastructure in Santa Clara on Monday. NVIDIA Vice President of Hyperscale and High-Performance Computing Ian Buck personally hand-delivered the processors.
Vera represents NVIDIA's first standalone CPU, designed specifically to support increasingly complex AI workloads. By targeting elite research organizations first, the company signals that Vera is built for the most demanding agent-based AI systems — not general-purpose computing. This move also tightens NVIDIA's grip on the AI hardware ecosystem, which has been dominated by its GPUs.
The specific numbers around performance or pricing were not disclosed by NVIDIA. However, the fact that Vera CPUs were directly delivered to labs training frontier models suggests they are being positioned as critical infrastructure for next-generation AI research. Oracle's inclusion indicates cloud availability may follow soon.
For AI labs currently reliant on GPUs for both training and inference, Vera could shift how workloads are partitioned across hardware. If the CPU delivers meaningful performance gains for agent-based tasks, it may accelerate the development of autonomous AI systems that require rapid reasoning and decision-making.
Neither NVIDIA nor the receiving labs have commented on testing timelines or Vera's specifications. The company's strategy of engaging top-tier labs first mirrors its approach to earlier hardware launches, which often preceded broader commercial availability.