SpaceXAI has adopted NVIDIA Vera CPUs for its Grok models and plans to deploy a Vera Rubin NVL72 system in orbit via the Starmind mission.
Alongside its Groq 3 LPX announcements at Hot Chips 2026, NVIDIA revealed that SpaceXAI will deploy NVIDIA Vera CPUs to accelerate its next generation of agentic AI applications. The company also plans to expand the infrastructure behind Grok on the Vera Rubin platform as it scales toward gigawatts of capacity, while extending that architecture into orbit via a first-generation Starmind AI satellite built on an optimized Vera Rubin NVL72 system.
This framing stands out in a week dominated by accelerator announcements. NVIDIA argues that agentic applications rely more heavily on the CPU than conventional inference does, since agents spend their time orchestrating tools, executing code, processing data, and running simulations between model calls. Under-provisioning that workload slows the agent loop and leaves GPUs idle.
NVIDIA describes Vera as the first CPU built specifically for AI agents, targeting the CPU-intensive operations surrounding model inference. The chip features 88 NVIDIA-designed Olympus cores, NVIDIA Spatial Multithreading, and high-bandwidth LPDDR5X memory delivering up to 1.2TB/s. NVIDIA claims up to 1.8x faster task completion than x86 CPUs across agentic AI, reinforcement learning, and data processing workloads—a vendor figure that warrants independent validation. These specifications align with earlier platform reveals detailing 176 threads, 1.8TB/s of NVLink-C2C bandwidth, 1.5TB of system memory, and 227 billion transistors.
“Agentic AI requires a new kind of computing system, one built not only to generate answers, but to take action,” said Ian Buck, vice president of hyperscale and high-performance computing at NVIDIA. “Vera gives AI agents the CPU performance to act in real time, executing code, processing data and coordinating complex tasks. SpaceXAI is taking this architecture from massive AI factories to the next frontier of computing in orbit.”
SpaceXAI plans to build the infrastructure behind Grok on Vera Rubin, which NVIDIA codesigns holistically across compute, networking, and software rather than shipping as discrete components. That integrated stack combines NVLink, Spectrum-X Ethernet, and BlueField data processing, with NVIDIA positioning the unified approach as the pathway to lowering cost per token at scale.
“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,” said Mike Nicolls, president of SpaceXAI. “That means higher-performance AI agents and more useful work from every watt of compute.”
The orbital component represents the more speculative half of the announcement. SpaceXAI is developing AI computing infrastructure for space, where constraints around power, thermal management, bandwidth, reliability, and physical integration differ drastically from terrestrial data centers. Its planned first-generation Starmind satellite will be based on an optimized Vera Rubin NVL72 rack-scale system, with NVIDIA and SpaceXAI adapting the design to those conditions while preserving a common architecture and software ecosystem.
NVIDIA has been building toward this capability for some time: at GTC 2026, it showcased a Space-1 Vera Rubin module aimed at orbital deployment. However, no timeline, launch date, or capacity figure accompanies the Starmind plan, and NVIDIA has filed the entire payload under forward-looking statements.