Where Is The Compute?

Explore / chart

Chart

Largest single clusters

Largest single AI clusters, by GPU countColossus 2: 550,000 GB200/GB300 GPUs deployed; +440,000 announced (not deployed); Prometheus: “Hundreds of thousands” of GPUs deployed (no exact count); target ~1,000,000; Colossus 1: Over 220,000 NVIDIA GPUs (>300 MW)0240K480K720K960K1.2MGPUs (count, not H100e) · solid = deployed · hatched = announced / targetColossus 2SpaceXAI · Memphis550,000 deployed+440,000 announced550K + 440K ann.PrometheusMeta · deployed: “100Ks”Target ~1M GPUs~1M targetColossus 1SpaceXAI → Anthropicover 220,000 GPUs>220K
GPU counts as reported by operators — not an independent audit, and not H100e. Meta only says Prometheus has “hundreds of thousands” of GPUs today, so its deployed bar is left blank rather than guessed.

Note: The source doc converts Colossus 2’s 550K GPUs to ~1.4M H100e; Epoch describes ~1.4M H100e as xAI’s expansion target. We chart GPU counts only. D1

Show data table
Largest clusters (GPU count)
ClusterDeployedAnnounced / targetAs ofSource
Colossus 2 (SpaceXAI)550,000 GB200/GB300+440,000 announced2026-10-01State of AI
Prometheus (Meta)“hundreds of thousands” (exact count not disclosed)~1,000,000 GPUs2026-10-19Meta Engineering
Colossus 1 (SpaceXAI, rented to Anthropic)>220,000 NVIDIA GPUs (>300 MW)—2026-05-06Anthropic

Sources: State of AI Report Compute Index (Nathan Benaich, Air Street Capital), “Who has the largest GPU cluster in this index?” (2026-10-01); Meta Engineering, “Inside Prometheus: Meta's Journey to a Gigawatt-Scale Supercluster” (2026-10-19); Anthropic, “Higher usage limits for Claude and a compute deal with SpaceX” (2026-05-06)

Open in the explorer

Embed this chart

Free to embed with credit. The citation is baked into the image. Download SVG. Full-resolution PNGs, a web component and every chart in one place come with the Chart & embed kit.