SpaceX and Nvidia have built a space-optimized version of Nvidia's Vera Rubin NVL72 AI system, with a launch into orbit planned for the fourth quarter of 2027, Elon Musk said in a post resharing Nvidia's announcement, as reported by MSN.
The timeline was echoed across outlets. Quartz reported that SpaceX is targeting late 2027 to launch Nvidia-powered AI satellites into orbit, while Outlook Business framed it as Musk setting a 2027 launch date for putting Nvidia AI chips in space. Engineering.com described the tie-up as a collaboration between Nvidia and SpaceX on agentic AI infrastructure.
Market reaction was mixed depending on who you read. Barron's reported that SpaceX stock rose as the Nvidia space chip paved the way for Musk's orbital AI ambitions. Yellow.com, by contrast, reported that both stocks fell even as Nvidia powered the 2027 orbital AI plan.
The basic idea behind orbital data centers is straightforward, even if the engineering is not. AI systems consume enormous amounts of electricity and generate heat that has to go somewhere. Space offers uninterrupted sunlight and no neighbors to object to a new power-hungry facility, but it also offers radiation, launch mass limits, and no technician who can walk over and swap a failed component — which is why a "space-optimized" variant of an existing rack-scale system is the notable detail here, rather than simply flying standard hardware.
The available reporting does not specify how many satellites are planned, what the systems would compute, how they would be powered or cooled, or what the project costs.
Why it matters: if AI computing starts moving off the planet, the fight over electricity, land, and water that data centers are currently having with communities on Earth changes shape entirely — and the companies that control both the rockets and the chips would be the ones setting the terms.