Chipmaker Nvidia is taking its artificial intelligence computing ambitions into space through a new collaboration with SpaceX. The companies recently announced a joint effort to design the compute payload for the Starmind AI1 satellite, a spacecraft built to process heavy artificial intelligence workloads directly from orbit using Nvidia hardware. This initial launch represents the first step in a broader strategy, as SpaceX aims to eventually deploy up to one million AI satellites that would operate collectively as a massive orbital data center.

A significant element of the agreement is SpaceX’s exclusive commitment to Nvidia hardware. SpaceX CEO Elon Musk stated that the company chose Nvidia exclusively because “the Vera Rubin architecture is the best architecture.” This endorsement highlights Nvidia’s technical lead, particularly given that it comes from a partner with substantial resources to develop custom silicon or seek alternative vendors.

The hardware driving this initiative is Nvidia’s Vera Rubin platform, which the company states delivers up to 25 times the AI performance of its H100 chips in orbit. While terrestrial data centers face growing constraints related to power consumption, land availability, and cooling capacity, orbital systems benefit from continuous solar energy and natural cooling. Industry estimates from Futurum Research project that orbital computing could reach a market value of $1 trillion by 2030, accelerated by declining launch costs and terrestrial power bottlenecks.

Although immediate financial returns for Nvidia investors will remain modest as space chips begin shipping later this year, the partnership underscores the company’s early advantage in emerging technological frontiers. By securing a central role in space computing, Nvidia is establishing its presence in an entirely new market before its full infrastructure is built.