SpaceX plans to scale its AI infrastructure to approximately 2 GW of compute capacity by late 2026.
The company targets nearly 10 GW of capacity by the end of 2027.
This expansion utilizes Nvidia’s next-generation Vera Rubin GPU architecture.
SpaceX will allocate most of this capacity to external customers.
The strategy positions the company to rival cloud giants like Amazon, Google, and Microsoft.
SemiAnalysis and CEO Elon Musk confirmed the strategic move into AI cloud services.