Parts/Racks/Meta Grand Teton
Racks · Grand Teton · 2022
Meta Grand Teton
Meta's open-sourced GPU platform, contributed to the Open Compute Project and deployed on OpenRack. The operator designing its own node platform is the pattern that makes hyperscale clusters hard to read from vendor announcements.
Why this part
Meta designed its own GPU server platform and open-sourced the design through the Open Compute Project rather than buying one from an integrator, which puts the rack itself in the same category as Graviton, Trainium and EFA: infrastructure a hyperscaler builds because it is cheaper than paying someone else's margin at its scale, and infrastructure invisible to any market analysis that counts units a vendor shipped. Grand Teton is what the Meta GenAI clusters on this site are built from, and open-sourcing it does not make it purchasable; it makes the design public while the manufacturing still runs through Meta's own supply chain.
Systems using this part
- Systems
- 2
- Rmax underneath
- 0.0 GFlop/s
- Units recorded
- -
| System | Role | Country | Units | System Rmax | Supplied by | Rank |
|---|---|---|---|---|---|---|
| Meta GenAI cluster (RoCE) 2024-03 | Integrator | USA | - | - | Meta Platforms | - |
| Meta GenAI cluster (InfiniBand) 2024-03 | Integrator | USA | - | - | Meta Platforms | - |
“Units recorded” sums only the edges where a public source states a quantity. It is a floor, not a total, and should never be read as installed base.