Power & cluster atlas
Follow the watts: the named sites behind the numbers, with power, chips, status and how sure we are. Named sites only. Power and chip counts are what operators or reporters published; blanks mean no number was published in our sources, not zero. Grid positions are schematic, not geographic.
Hyperion
Meta · Louisiana
announced / under development
- Power
- 2,000–5,000 MW 2 GW by 2030, up to 5 GW over several years
- Chips
- not published no chip count published
- As of
- 2026-10-19
- Confidence
- announced target
- Sources
- TechCrunch, “Mark Zuckerberg says Meta is building a 5GW AI data center” (2025-07-14); Meta Careers blog, “Why networking infrastructure is critical for AI innovation” (2025)
Prometheus
Meta · New Albany, OH
- Power
- ~1,000 MW ~1 GW announced target ('gigawatt-scale'; see D7)
- Chips
- not published 'hundreds of thousands' deployed (no exact count); ~1,000,000 GPU target
- As of
- 2026-10-19
- Confidence
- operator-reported; deployed count not disclosed
- Sources
- Meta Engineering, “Inside Prometheus: Meta's Journey to a Gigawatt-Scale Supercluster” (2026-10-19); Meta Careers blog, “Why networking infrastructure is critical for AI innovation” (2025)
Colossus 1
SpaceXAI (xAI), capacity rented to Anthropic · Memphis, TN
- Power
- over 300 MW over 300 MW (floor)
- Chips
- over 220,000 over 220,000 NVIDIA GPUs (floor)
- As of
- 2026-05-06
- Confidence
- company-stated floor
- Sources
- Anthropic, “Higher usage limits for Claude and a compute deal with SpaceX” (2026-05-06)
Colossus 2
SpaceXAI (xAI) · Memphis, TN area
- Power
- not published no MW figure in our sources
- Chips
- 550,000 550,000 GB200/GB300 deployed; +440,000 announced
- As of
- 2026-10-01
- Confidence
- operator-reported count (not an independent audit)
- Sources
- State of AI Report Compute Index (Nathan Benaich, Air Street Capital), “Who has the largest GPU cluster in this index?” (2026-10-01)
Power by site
Show data table
| Site | Operator | Status | Power | Chips | As of | Confidence |
|---|---|---|---|---|---|---|
| Colossus 2 | SpaceXAI (xAI) | operating, expanding | not published | 550,000 | 2026-10-01 | operator-reported count (not an independent audit) |
| Colossus 1 | SpaceXAI (xAI), capacity rented to Anthropic | operating | over 300 MW | over 220,000 | 2026-05-06 | company-stated floor |
| Prometheus | Meta | phase 1 operating; scaling | ~1,000 MW | not published | 2026-10-19 | operator-reported; deployed count not disclosed |
| Hyperion | Meta | announced / under development | 2,000–5,000 MW | not published | 2026-10-19 | announced target |
Scale
Cluster scale: one square = 10,000 GPUs
Show data table
| Cluster | Reported | As of | Confidence | Source |
|---|---|---|---|---|
| Colossus 2 | 550,000 GPUs deployed + 440,000 announced | 2026-10-01 | operator-reported count (not an independent audit) | State of AI Report Compute Index (Nathan Benaich, Air Street Capital) — “Who has the largest GPU cluster in this index?” (2026-10-01) |
| Prometheus | “hundreds of thousands” deployed (no exact count); ~1,000,000 target | 2026-10-19 | operator-reported: 'hundreds of thousands' deployed (no exact count); ~1M GPU target | Meta Engineering — “Inside Prometheus: Meta's Journey to a Gigawatt-Scale Supercluster” (2026-10-19) |
| Colossus 1 | over 220,000 GPUs (floor) | 2026-05-06 | company-stated floor ('over 220,000 NVIDIA GPUs', >300 MW) | Anthropic — “Higher usage limits for Claude and a compute deal with SpaceX” (2026-05-06) |
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)
