Million-GPU class: 1 million and more
The plans and fleets of the hyperscalers and Stargate. 5 providers.
KindNeocloudInference platformSovereign cloudHyperscalerDeveloper cloudMarketplaceFormer minerChipmaker cloudSupercomputerResearch cluster
ChipNvidia BlackwellNvidia HopperNvidia AmpereOther Nvidia GPUsAMD InstinctGoogle TPUAWS Trainium and InferentiaIntel GaudiCerebras wafer-scaleGroq LPUOther and custom chips
馃嚭馃嚫 United StatesColossus
An AI supercomputer cluster in Tennessee and Mississippi built by xAI.Supercomputer 路 Nvidia Hopper 路 2024
馃嚚馃嚟 SwitzerlandAlps
Swiss national supercomputer using Nvidia GH200 chips for AI and scientific research.Supercomputer 路 Other Nvidia GPUs 路 2024
馃嚭馃嚫 United StatesSpaceX Super Intelligence
Musk's AI company behind Grok and the Colossus data centres, now part of SpaceX.Neocloud 路 Nvidia Hopper 路 2023
馃嚭馃嚫 United StatesSequoia
IBM Blue Gene/Q supercomputer built for US nuclear weapons simulation at LivermoreSupercomputer 路 Other and custom chips 路 2011
馃嚭馃嚫 United States and 4 moreGoogle data centers
Hyperscale campuses running Search, YouTube, Gemini and Google Cloud AI workloadsHyperscaler 路 Google TPU
馃嚚馃嚟 SwitzerlandAlps
Swiss national supercomputer using Nvidia GH200 chips for AI and scientific research.Supercomputer 路 Other Nvidia GPUs 路 2024
馃嚭馃嚫 United StatesSequoia
IBM Blue Gene/Q supercomputer built for US nuclear weapons simulation at LivermoreSupercomputer 路 Other and custom chips 路 2011
Photo credits on this page
The small pictures above come from Wikimedia Commons under free licences. Some are cropped to fit.
- Alps: Marco Abram, CSCS, CC BY-SA 3.0, source
- SpaceX Super Intelligence: BrokenSegue, CC BY-SA 4.0, source
- Sequoia: Lawrence Livermore National Laboratory, Public domain, source
- Google data centers: erwinboogert, CC BY-SA 3.0, source