Nvidia Corp has unveiled new computing platforms designed for orbital data centers at its GTC 2026 conference, addressing growing artificial intelligence demand and Earth’s energy constraints by exploring data processing solutions in space.
The development is centered around Nvidia’s Vera Rubin Space-1 Module, which incorporates the IGX Thor and Jetson Orin, and is designed to support space missions. The company specified that these chips are “engineered for size-, weight- and power-constrained environments.”
Nvidia has partnered with several companies to deploy its technology, including Axiom Space, Starcloud, and Planet. According to Nvidia CEO Jensen Huang, “Space computing, the final frontier, has arrived.” Huang emphasized that as satellite constellations are deployed, intelligence must reside where data is generated.
Huang also acknowledged the engineering challenges that remain, particularly regarding thermal management. “In space, there’s no convection, there’s just radiation,” Huang stated, highlighting the need to develop new cooling solutions.
The concept of orbital data centers is seen as a potential solution to mitigate the rising electricity costs associated with AI infrastructure. However, high costs and limited access to rocket launches currently present significant barriers.
Despite these challenges, AI companies are actively pursuing space-based solutions. For instance, Google initiated ‘Project Suncatcher’ in November to explore space computing concepts. Additionally, last month, SpaceX acquired Elon Musk’s xAI in a $1.25 trillion deal, with plans to develop space data centers. Notably, SpaceX is a significant customer for Nvidia.
Furthermore, SpaceX sought Federal Communications Commission approval in January to launch one million satellites for AI data centers. This plan has drawn opposition from scientists who cite concerns over light pollution and orbital debris.
