SpaceX may need over 2 million Nvidia Rubin GPUs for compute buildout

SpaceX plans to expand its compute capacity more than fivefold by the end of 2027, betting exclusively on Nvidia's upcoming Vera Rubin platform, which could require well over a million new GPUs. The company's AI/compute segment posted $2.56 billion in Q2 revenue, driven mostly by leasing out its own server capacity.

Why it matters: This quantifies how large SpaceX's pivot into compute leasing has become, a scale of GPU commitment that rivals dedicated cloud providers, and it lines up with SpaceX's compute business already outearning its rocket business and Musk's stated plan to route Nvidia GPU purchases exclusively through SpaceX and xAI.