SpaceXAI Adopts NVIDIA Vera CPU, Targets Gigawatt AI Expansion and 1st Starmind Satellite
Updated
Updated · NVIDIA Blog · Aug 24
SpaceXAI Adopts NVIDIA Vera CPU, Targets Gigawatt AI Expansion and 1st Starmind Satellite
3 articles · Updated · NVIDIA Blog · Aug 24
Summary
SpaceXAI said it will deploy NVIDIA Vera CPUs for next-generation agentic AI workloads and use the same architecture to support Grok on Earth and its first Starmind AI satellite in orbit.
Vera is aimed at the CPU-heavy work around inference—tool orchestration, code execution, data processing and simulation—and NVIDIA says it delivers up to 1.8x faster task completion than x86 CPUs.
88 Olympus cores and up to 1.2TB/s of LPDDR5X memory bandwidth are meant to keep GPUs fully utilized as SpaceXAI scales its AI infrastructure toward gigawatts of computing capacity.
The orbital push extends an optimized Vera Rubin NVL72 system into space, where SpaceXAI and NVIDIA are adapting for power, thermal, bandwidth and reliability constraints while keeping a common software stack.