Nvidia now counts Musk's SpaceX as a cloud provider, and the two are designing a server for space

What's going on? Nvidia CEO Jensen Huang says Elon Musk's business has moved from Nvidia's enterprise customer segment into its cloud-service-provider group, alongside the big cloud platforms. Speaking on the BG2 podcast, Huang said Musk is unusual in doing three things at once: training his own models, serving them (consumption of Grok "is really skyrocketing", he said) and renting compute out to others.

A server for space: Huang said Nvidia and SpaceX engineers are "working super closely together to create a new type of server, a new type of architecture" for data centres in space that "might be super good... in space and on Earth." His case for Nvidia hardware in orbit is economic: launches are expensive, so only the most valuable, most rentable compute makes sense up there, and Nvidia keeps updating its software for years after a chip ships.

How big? Musk wants 10 gigawatts of compute by the end of 2027. Replying to the clip on X, he wrote: "Maybe more than 10GW." Musk said in August that SpaceX's first Nvidia-powered AI satellites should launch in the fourth quarter of 2027 and reach significant scale in 2028.

Sources BG2 clip on X, Elon Musk on X, PJFP episode recap, Bloomberg (August)