Nvidia: SpaceX AI will deploy the Vera CPU to accelerate agentic AI, and expand the Vera Rubin platform to spaceborne computing.

Nvidia officially announced that SpaceXAI will deploy Nvidia’s Vera CPU to accelerate its next-generation agentic AI applications. Vera is the first CPU purpose-built for AI agents, boasting 88 of Nvidia’s in-house developed Olympus cores, and leveraging spatial multithreading technology and LPDDR5X high-bandwidth memory with a bandwidth of 1.2TB/s. In use cases including agentic AI, reinforcement learning, and data processing, it delivers up to 1.8x faster task completion than x86 CPUs. SpaceXAI President Mike Nicolls stated: “Vera’s CPU performance and memory bandwidth are sufficient to support large-scale task orchestration, code execution, and data processing, enabling GPUs to operate at peak efficiency, thereby generating more effective output per watt of computing power.” SpaceXAI also plans to expand the AI infrastructure powering Grok based on the NVIDIA Vera Rubin platform, advancing toward gigawatt-scale computing capacity. Vera Rubin integrates Nvidia accelerated computing, NVLink interconnect, Spectrum-X networking, and BlueField data processing technology, with the aim of reducing per-token costs and boosting the overall energy efficiency of AI factories. Most notably, SpaceXAI will send an optimized Vera Rubin NVL72 system into space; its first-generation Starmind AI satellite will adopt this rack-level architecture, marking Nvidia’s first landmark deployment extending accelerated computing from terrestrial data centers to orbital computing.