Companies/Cerebras
Sunnyvale, CA, USA
Cerebras
Builds a processor the size of a wafer, abandoning the reticle limit that every other design works within. The architectural argument is that keeping a large model entirely on one piece of silicon removes the interconnect problem rather than solving it.
Systems this company appears in
- Systems
- 12
- Aggregate Rmax
- 0.0 GFlop/s
- Aggregate power
- 100.00 MW
- First appearance
- 2021-03
- Most recent
- 2026-09
- Concentration
- -
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 | 6 | 0.0 GFlop/s |
| Accelerator | 10 | 0.0 GFlop/s |
| Cooling | 1 | 0.0 GFlop/s |
Appearances
| System | Supplied as | Shipped under | Rmax | Rank |
|---|---|---|---|---|
| Cerebras Montreal Inference Site (MTL-3) Canada · 2025-10 | Integrator | Cerebras | - | unranked |
| Cerebras Oklahoma City USA · 2025-09 | Accelerator, Integrator, Cooling | Cerebras | - | unranked |
| Condor Galaxy 1 USA · 2023-07 | Accelerator | Cerebras | - | unranked |
| Cerebras Andromeda USA · 2022-11 | Accelerator, Integrator | Cerebras | - | unranked |
| EPCC Cerebras CS-2 Cluster United Kingdom · 2022-09 | Accelerator | Cerebras | - | unranked |
| HOLL-I USA · 2022-05 | Accelerator | Cerebras | - | unranked |
| LRZ Cerebras CS-2 Germany · 2022 | Accelerator | Cerebras | - | unranked |
| Neocortex USA · 2021-03 | Accelerator | Cerebras | - | unranked |
| Condor Galaxy 3 USA · | Accelerator, Integrator | Cerebras | - | unranked |
| Cerebras Guyana AI Data Center Guyana · | Accelerator, Integrator | Cerebras | - | unranked |
| Bell AI Fabric Saskatchewan Canada · | Accelerator | Cerebras | - | unranked |
| Cerebras Mikkeli AI Data Centre Finland · | Integrator | Cerebras | - | unranked |
Parts supplied
| Part | Category | Generation | Released |
|---|---|---|---|
| Cerebras WSE-3 | Accelerators | Wafer Scale Engine | 2024 |
| Cerebras CS-2 (WSE-2) | Accelerators | Wafer Scale Engine 2 | - |
AI datacenters operated
| System | Status | Accelerators | Stated power (MW) |
|---|---|---|---|
| Condor Galaxy 1 | operational | 64 | - |
| Cerebras Andromeda | operational | - | - |
| Condor Galaxy 3 | operational | 64 | - |
| Cerebras Guyana AI Data Center | planned | - | 100 |
| Cerebras Oklahoma City | operational | - | - |
| Cerebras Montreal Inference Site (MTL-3) | operational | - | 5 |
Figures come from the AI datacenter metrics; blank means none recorded.
Compute deals
Headline values are as announced and are not additive. See all compute deals.
Sites operated
| Site | Location | Cooling | Commissioned |
|---|---|---|---|
| Colovore (Santa Clara) | Santa Clara, CA, USA | - | 2023 |
| Condor Galaxy 3 site (Dallas) | Dallas, USA | - | 2024 |
| Cerebras AI Data Center (Wales, Guyana) | Wales, Guyana | - | - |
| Scale Datacenter (Oklahoma City) | Oklahoma City, OK, USA | Direct-to-chip water cooling (closed loop) | 2025 |
Related news
- Gimlet Cloud partners with Cerebras to deploy 100MW of wafer-scale compute · Data Center DynamicsA newly well-funded neocloud is committing to 100MW of Cerebras wafer-scale inference capacity and will be a launch partner for Cerebras' next CS-4 chip; a rare procurement data point for an accelerator vendor outside the Nvidia-dominated share this graph otherwise tracks.
Further reading
Inside a node
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.
Power and cooling
Why air ran out at 20 kW a rack, how warm-water cooling removes the chillers, and what PUE conceals.