Nvidia Launches Line Of Processors For AI Data Centers In Space

Tech giants and their billionaire founders are getting serious about putting artificial intelligence data centers in space. Now, Nvidia is creating a supply chain to serve them. 

Nvidia will begin producing space computing modules, versions of its high-performance processing products intended specifically for use in orbital data centers and on spacecraft, the company announced this week. The new offerings include a space-oriented version of its most cutting-edge product line, known as Vera Rubin, along with two other types of processors. 

The chipmaker has launched its processors into space once before, placing one of its earlier H100 graphics processing units on a test satellite in November. But this week’s announcement marks a major step up, both in the capability of compute being sent to space and in Nvidia’s commitment to a strategy that targets space data centers as a distinct market.

“Space computing, the final frontier, has arrived,” Nvidia CEO Jensen Huang said in a statement. “AI processing across space and ground systems enables real-time sensing, decision-making and autonomy, transforming orbital data centers into instruments of discovery and spacecraft into self-navigating systems.”

While Huang primarily highlighted use cases involving processing data from satellites and spacecraft, Nvidia has made clear that it envisions terrestrial AI workloads ultimately being located in space, framing its modules as a product that “unlocks scalable, space-based AI infrastructure beyond Earth.”

Nvidia says it already has customers for its space-based processing modules, mostly satellite communications firms and orbital data center startups like Aetherflux and Starcloud. But the announcement comes as the idea of data centers in space is emerging from the speculative fringes of the digital infrastructure landscape and is gaining traction among tech’s biggest names.

Over the past six months, the world’s largest tech firms have begun steering resources toward making orbital data centers for AI computing a reality, driven by the promise of a limitless supply of free, uninterrupted solar power. 

In November, Google announced Project Suncatcher, a program that aims to create an AI data center in low Earth orbit, with computing spread across a tightly knit array of satellites. While Google is funding the projects as a long-term “moonshot,” the company has devoted capital and personnel toward the effort. A pair of prototype satellites carrying Google’s proprietary AI chips are scheduled to launch by early 2027.

Weeks later, Jeff Bezos’ Blue Origin and Elon Musk’s SpaceX revealed their ambitions to launch orbital data centers that will host AI computing payloads.

Blue Origin had reportedly been working for more than a year on the technology needed for orbital AI data centers. Musk, meanwhile, is merging SpaceX and xAI, with the development of space-based data centers a central focus of the business.

“Space-based AI is obviously the only way to scale,” Musk wrote on SpaceX’s website in February. “It’s always sunny in space!”