SpaceXAI is adopting NVIDIA Vera CPUs to support next-generation agent-based AI workloads and expand the infrastructure behind Grok. The CPUs are designed for tasks that increasingly sit outside model inference, including tool control, code execution, data processing, orchestration and simulations. NVIDIA says Vera can complete relevant workloads up to 1.8 times faster than x86 CPUs.
Also Read: Chiyoda Construction Launches Shizuoka Consultation Branch
SpaceXAI plans to scale its AI infrastructure toward gigawatt-level computing using the NVIDIA Vera Rubin platform. The company also plans to take the architecture beyond Earth, with its first-generation Starmind AI satellites expected to use an optimized Vera Rubin NVL72 system. The move points to a broader shift toward AI infrastructure designed specifically for agent workloads and large-scale computing.

