Musk targets 13GW AI compute by 2027 with 440,000-GPU Vera Rubin cluster
scaling01 · x · 2026-10-21
Musk laid out SpaceX's compute expansion: 200,000 NVIDIA GB300 GPUs online per month through the rest of 2026, reaching 2.5-3GW by end of 2026 and potentially 13GW total by 2027. He plans a single 440,000-GPU Vera Rubin training cluster by end of Q2 2027, requiring an additional 20GW of power and cooling infrastructure next year.
Related event: Musk Targets 13 GW Compute by 2027 with 440K-GPU Vera Rubin Cluster(4 posts)→
More from Infra
- NVIDIA says AI-assisted CUDA optimization delivered 2x in a month, vs 5x/year by experts — SumitGup · 2026-10-21
- Anthropic's cloud partnerships turn messy as rivals-turned-providers starve it of compute — TechCrunch AI · 2026-10-21
- SpaceX is deploying 220K Nvidia GB300s per month, and Musk says the pace will accelerate — firstadopter · 2026-10-21
- Musk: xAI to bring ~200K GB300s online per month, targeting 13GW by end of next year — XFreeze · 2026-10-21
- Why Google is putting AI chips in orbit: near-constant solar power, brutal engineering challenges — soumitrashukla9 · 2026-10-21
- Training linear-attention Laplace 3B on AWS Trainium2: 120K tokens/sec, fixed 29.5MB state — AWS ML Blog · 2026-10-21