COMPUTE ATLAS Supercomputer supply-chain graph
1540 systems 545 sites Sourced data

Companies/Meta Platforms

Menlo Park, CA, USA

Meta Platforms

Operates among the largest GPU fleets in the world for internal training; disclosure is via engineering blog rather than HPL submission.

Systems this company appears in

Systems
12
Aggregate Rmax
3.31 PFlop/s
systems it appears in
Aggregate power
349.5 kW
where disclosed
First appearance
2017
Most recent
2026-07-08
Concentration
100%
largest is Facebook DGX-1 Cluster

Aggregate Rmax is the sum of the systems this company supplies any part of. It is not revenue and not attributable share: an integrator and a memory vendor in the same machine both count its full FLOPS here. Attribution weighted by role is Phase 4.

Role breakdown

Role Systems Rmax in those systems
Integrator 12 3.31 PFlop/s
Storage 2 none published
Cooling 2 none published

Appearances

System Supplied as Shipped under Rmax Rank
Meta GenAI cluster (RoCE)
USA · 2024-03
Integrator, Storage, Cooling Meta Platforms - unranked
Meta GenAI cluster (InfiniBand)
USA · 2024-03
Integrator, Storage, Cooling Meta Platforms - unranked
Facebook DGX-1 Cluster
USA · 2017
Integrator Meta Platforms 3.31 PFlop/s #32 June 2017 →
Meta Prometheus
USA ·
Integrator Meta Platforms - unranked
Meta Hyperion
USA ·
Integrator Meta Platforms - unranked
Meta Lebanon (Indiana)
USA ·
Integrator Meta Platforms - unranked
Meta El Paso (Texas)
USA ·
Integrator Meta Platforms - unranked
Meta Rosemount (Minnesota)
USA ·
Integrator Meta Platforms - unranked
Meta Montgomery (Alabama)
USA ·
Integrator Meta Platforms - unranked
Meta Cheyenne (Wyoming)
USA ·
Integrator Meta Platforms - unranked
Meta Sturgeon Data Centre
Canada ·
Integrator Meta Platforms - unranked
Meta Temple (Texas)
USA ·
Integrator Meta Platforms - unranked

Parts supplied

Part Category Generation Released
Meta MTIA v2 Accelerators MTIA 2024
Meta Grand Teton Racks Grand Teton 2022
Meta Tectonic Storage Tectonic 2021

AI datacenters operated

SystemStatusAcceleratorsStated power (MW)
Meta GenAI cluster (RoCE) operational 24,576 -
Meta GenAI cluster (InfiniBand) operational 24,576 -
Meta Research SuperCluster (RSC) operational 6,080 -
Meta Prometheus operational 166,015 612
Meta Hyperion under construction - 2,260
Meta Lebanon (Indiana) under construction - 1,000
Meta El Paso (Texas) under construction - 1,000
Meta Rosemount (Minnesota) operational 85,400 178
Meta Montgomery (Alabama) operational 69,100 153
Meta Cheyenne (Wyoming) operational 68,400 152

Figures come from the AI datacenter metrics; blank means none recorded.

Compute deals

DateRoleCounterpartyTypeMWValue ($bn)Source
2026-03-31 Buyer CoreWeave Cloud contract - 21 sec.govsec.gov
2026-03-13 Buyer Nebius Cloud contract - 27 sec.gov
2026-02-23 Buyer AMD Equity for compute 6,000 - sec.gov
2026-02-17 Buyer NVIDIA Chip supply - - nvidianews.nvidia.com
2025-11 Buyer Nebius Cloud contract - 2.9 sec.gov
2025-10 Buyer Blue Owl Capital Joint venture - 12.3 sec.gov
2025-09-30 Buyer CoreWeave Cloud contract - 14.2 sec.govsec.gov
2025-08-21 Buyer Alphabet Cloud contract - 10 cnbc.comdatacentrereview.com

Headline values are as announced and are not additive. See all compute deals.

Sites operated

Site Location Cooling Commissioned
Meta (undisclosed US site) USA - 2024
Meta New Albany AI Campus (Prometheus) New Albany, OH, USA - 2026
Meta Richland Parish AI Campus (Hyperion) Richland Parish, LA, USA - -
Meta Lebanon AI Campus Lebanon, IN, USA Closed-loop liquid cooling -
Meta El Paso AI Campus El Paso, TX, USA Closed-loop liquid cooling -
Meta Rosemount Data Center Rosemount, MN, USA - -
Meta Montgomery Data Center Montgomery, AL, USA - -
Meta Cheyenne Data Center Cheyenne, WY, USA - -
Sturgeon County Data Centre Site Sturgeon County, AB, Canada Closed-loop liquid cooling with dry cooling -
Meta Temple Data Center Site Temple, TX, USA - -

Related news

  • Meta to expand Eagle Mountain data center campus in Utah · Data Center Dynamics
    Meta's fourth expansion of its Eagle Mountain, Utah campus pushes its investment there past $3 billion, adding self-built capacity outside the HPL-measured world this site otherwise tracks -- the same disclosed-versus-undisclosed distinction that motivates the shadow-list page. The new buildings will use closed-loop liquid cooling with dry cooling, a facility detail worth watching as more hyperscaler capacity shifts away from air.
  • Panmnesia and Meta propose extending cache-coherent CXL beyond the rack to datacenter scale · HPCwire
    Every interconnect row in this dataset describes a fabric that either stops at the rack or treats cross-rack traffic as a network hop; the proposal here is to hold one cache-coherent domain across a whole facility, which would turn that boundary from a given into a design choice. It is an invited review article with components claimed as validated silicon but no deployed system behind it, so nothing in it belongs in the parts layer yet.
  • Louisiana regulators fast-track review of seven more gas plants for Meta's Hyperion campus · Union of Concerned Scientists
    Ten gas fired units totaling roughly 7.5 gigawatts, up from three approved in August 2025, would power a single Meta data center at a scale that a critic notes exceeds six times New Orleans' peak demand. It is a sharp illustration of the recurring finding elsewhere in this dataset that grid power, not chip supply, has become the binding constraint on gigawatt class AI campuses like Hyperion.
  • Meta signs nuclear power deals with Oklo, Vistra and TerraPower for up to 6.6GW · Utility Dive
    Vistra's contribution includes uprates at its existing Perry and Davis-Besse reactors, an operating asset route to new capacity that is faster than the Oklo and TerraPower agreements, which depend on advanced reactor designs not expected online before 2030 and 2032 respectively. Meta's Prometheus campus in Ohio, already in this dataset, is named as one beneficiary, though only a fraction of the contracted capacity is available on any timeline that matches Prometheus's 2026 target.

All news →

Further reading