Key Takeaways:
SpaceXAI is betting its agentic AI workloads on Nvidia's Vera CPU, extending the Vera Rubin platform from gigawatt-scale data centers to a planned orbital satellite.
SpaceXAI will deploy Nvidia's Vera CPU to accelerate agentic AI workloads, scaling its Grok infrastructure toward gigawatts of compute and extending the Vera Rubin platform into orbit with a planned Starmind satellite. The announcement, made Aug. 24, marks the first CPU built specifically for AI agents.
"Vera gives us the CPU performance and memory bandwidth to run enormous amounts of orchestration, code and data processing while keeping GPUs doing what they do best," Mike Nicolls, president of SpaceXAI, said. "That means higher-performance AI agents and more useful work from every watt of compute."
Vera features 88 Nvidia-designed Olympus cores, Spatial Multithreading technology and LPDDR5X memory delivering up to 1.2TB/s of bandwidth. Nvidia claims up to 1.8x faster task completion versus x86 CPUs across agentic AI, reinforcement learning and data processing workloads, though it did not disclose the comparison processors or test conditions. SpaceXAI plans to expand its Grok infrastructure on Vera Rubin and base its first-generation Starmind satellite on an optimized Vera Rubin NVL72 rack-scale system.
The deal extends Nvidia's reach from terrestrial AI factories into orbital computing, where power, thermal management, bandwidth, reliability and physical integration impose constraints far beyond conventional data centers. Nvidia separately said the Groq 3 LPX inference accelerator, an extension of the Vera Rubin platform, has entered full production, delivering 3,400 output tokens per second running Gemma 4 31B with a 100,000-token context length.
Agentic AI systems break tasks into multiple steps, access external tools, execute code and run simulations between model calls — work that falls on CPUs rather than GPUs. Vera is designed to accelerate those stages, keeping accelerators continuously fed. The 88 Olympus cores and Spatial Multithreading target the orchestration bottleneck that emerges as agents chain together dozens of tool calls per task.
The competitive stakes are clear. Nvidia's Vera Rubin platform faces AMD's Instinct line and custom silicon from hyperscalers building in-house accelerators. But Vera's positioning as a dedicated agentic-AI CPU gives Nvidia a distinct angle: x86 rivals such as Intel and AMD have yet to ship a processor purpose-built for the orchestration-heavy workloads that define agentic systems.
SpaceXAI plans to expand its Grok infrastructure on the Vera Rubin platform, which combines Nvidia accelerated computing, NVLink interconnect, Spectrum-X Ethernet, BlueField data processing and Nvidia software in an integrated architecture designed to drive down cost per token. The company is scaling toward gigawatts of computing capacity.
The most ambitious element is orbital. SpaceXAI's first-generation Starmind AI satellite will be based on an optimized Vera Rubin NVL72 rack-scale system, adapted for space where power, thermal management, bandwidth, reliability and physical integration differ sharply from terrestrial data centers. Nvidia and SpaceXAI are working to preserve a common architecture and software ecosystem across both environments, though neither company has disclosed a deployment timeline or validated the NVL72 system under space conditions.
The partnership strengthens Nvidia's position in the agentic AI buildout, where demand for CPU-side orchestration is rising alongside GPU compute. Nvidia shares have historically reacted modestly to AI infrastructure announcements — AI-tagged events averaged a -1.15% 24-hour move across five recent reports, according to StockTitan data. The deal also deepens SpaceXAI's reliance on Nvidia hardware, following Elon Musk's commitment to Nvidia GPUs for the company's AI efforts. The orbital extension gives Nvidia a first-mover position in space-based AI computing, a market with no established incumbent, though the products remain available only on a when-and-if-available basis.
This article is for informational purposes only and does not constitute investment advice.
