SpaceX Plans 2027 Launch of Orbital AI System With Nvidia
The first Starmind satellite will use an optimized Vera Rubin architecture, with a larger deployment targeted for 2028.
Topics
News
- Apple Tightens Mac Access Rules as AI Agents Gain Reach
- Trump Puts Intelligence Chief Jay Clayton in Charge of AI Task Force
- IBM Widens AI, Quantum Research With IIT Bombay, IISc
- OpenAI, Synopsys Build AI Model That Could Slash Chip Design Time
- Pentagon Plans Four-Star Command for AI, Drone Warfare
- California Draws the Line on AI Firings, Layoffs and Workplace Surveillance
SpaceX plans to put Nvidia’s next-generation AI computing hardware into orbit in late 2027, taking an early step toward running data-center infrastructure in space.
Chief Executive Elon Musk said SpaceX and Nvidia have designed a space-optimized Vera Rubin NVL72 system for launch in the fourth quarter of 2027, with a larger deployment planned for 2028.
Nvidia separately said SpaceXAI’s first-generation Starmind AI satellite will be based on an optimized version of its Vera Rubin rack-scale architecture.
A standard terrestrial NVL72 system combines 72 Rubin GPUs and 36 Vera CPUs with networking and data-processing hardware, according to Nvidia.
Musk said the orbital design would be simpler, cheaper, denser and lighter than a conventional rack but provided no specifications.
The companies are adapting the platform for the power, thermal, bandwidth, reliability and physical constraints of orbit.
SpaceXAI plans to use Vera CPUs for tasks such as AI-agent orchestration, code execution, data processing and simulation while reserving GPUs for heavier model workloads.
The announcement did not disclose the launch vehicle, orbital configuration, cooling system or protections against radiation and launch stresses. SpaceX has not yet operated a Starmind satellite.
