Elon Musk says SpaceX intends to launch its first AI satellites carrying Nvidia chips in the fourth quarter of 2027, according to Times Now, with the orbital data center project expected to reach significant scale in 2028.
The effort is being called Starmind. According to Engadget, SpaceX plans to use Nvidia GPUs for the project, and Musk's pitch centers on harnessing solar energy in space to power AI computing.
The idea behind an orbital data center is straightforward even if the engineering is not. AI models are trained and run on enormous banks of power-hungry chips, and on Earth those chips need electricity, land, and cooling — all of which are getting harder and more expensive to secure. In orbit, sunlight is constant and unfiltered by atmosphere or weather, which is the appeal Musk is leaning on.
Engadget frames its coverage around the obstacles such hardware would face in space, a reminder that moving server-grade GPUs off the planet is not simply a matter of strapping them to a rocket.
Investors reacted to the announcement. Yahoo Finance, via Google News, reported that SpaceX stock climbed as Musk targeted the 2027 launch date for the Nvidia AI satellites.
A few caveats are worth holding onto. The available reporting gives a timeline and a chip partner, but not the number of satellites, the cost, the scale of computing power involved, or who the customers would be. Musk-announced dates have a long history of moving, and Q4 2027 is more than a year out.
Why it matters: if AI's electricity appetite really is outgrowing what the grid can comfortably supply, then whether computing can be moved into orbit — and whether one company can own both the rockets and the chips pipeline to do it — becomes a question about who controls the infrastructure the next decade of AI runs on.