Companies/Amazon
Seattle, WA, USA
Amazon
AWS designs Trainium and Inferentia accelerators and operates them at cluster scale, almost entirely outside public HPL reporting.
Systems this company appears in
- Systems
- 11
- Aggregate Rmax
- 724.27 TFlop/s
- Aggregate power
- 609.41 kW
- First appearance
- 2011-11
- Most recent
- 2026-05
- Concentration
- 67%
systems it appears in
where disclosed
largest is Amazon EC2 C3 Instance 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 | 10 | 724.27 TFlop/s |
| Accelerator | 5 | none published |
| Interconnect | 5 | 724.27 TFlop/s |
Appearances
| System | Supplied as | Shipped under | Rmax | Rank |
|---|---|---|---|---|
| AWS Ridgeland Campus USA · 2026-05 | Accelerator, Integrator | Amazon | - | unranked |
| Project Rainier USA · 2025-10 | Accelerator, Interconnect | Amazon | - | unranked |
| AWS Madison Mega Site (Canton) USA · 2025-06 | Accelerator, Integrator | Amazon | - | unranked |
| Double Helix United Kingdom · 2024-05 | Integrator | Amazon | - | unranked |
| Amazon EC2 C3 Instance Cluster USA · 2013-11 | Interconnect, Integrator | Amazon | 484.18 TFlop/s | #65 November 2013 → |
| Amazon EC2 Cluster Compute Instances USA · 2011-11 | Interconnect, Integrator | Amazon | 240.09 TFlop/s | #42 November 2011 → |
| Project Ceiba USA · | Interconnect, Integrator | Amazon | - | unranked |
| SK Ulsan AI Data Center South Korea · | Integrator | Amazon | - | unranked |
| AWS Trainium2 UltraServer USA · | Accelerator, Interconnect, Integrator | Amazon | - | unranked |
| AWS Trainium3 UltraServer USA · | Accelerator, Integrator | Amazon | - | unranked |
| AWS Susquehanna Campus (Salem Township) USA · | Integrator | Amazon | - | unranked |
Parts supplied
| Part | Category | Generation | Released |
|---|---|---|---|
| AWS Trainium3 | Accelerators | Trainium | 2025 |
| AWS Trainium2 | Accelerators | Trainium | 2024 |
| AWS Graviton4 | CPUs | Neoverse V2 | 2024 |
| AWS Elastic Fabric Adapter | Interconnect | EFA | 2019 |
AI datacenters operated
| System | Status | Accelerators | Stated power (MW) |
|---|---|---|---|
| Project Rainier | operational | 500,000 | 1,092 |
| Project Ceiba | under construction | 20,736 | - |
| AWS Madison Mega Site (Canton) | operational | 326,600 | 284 |
| AWS Ridgeland Campus | operational | 261,200 | 228 |
| AWS Trainium2 UltraServer | operational | 64 | - |
Figures come from the AI datacenter metrics; blank means none recorded.
Compute deals
| Date | Role | Counterparty | Type | MW | Value ($bn) | Source |
|---|---|---|---|---|---|---|
| 2026-04-20 | Seller | Anthropic | Cloud contract | 5,000 | 100 | anthropic.comsec.gov |
| 2026-02-27 | Seller | OpenAI | Cloud contract | - | 100 | sec.gov |
| 2025-11-03 | Buyer | Cipher Mining | Datacenter lease | 300 | 5.5 | sec.gov |
| 2025-11 | Seller | OpenAI | Cloud contract | - | 38 | aboutamazon.comsec.gov |
| 2025-10-29 | Seller | Anthropic | Cloud contract | - | - | aboutamazon.com |
| 2025-06-20 | Buyer | SK Telecom | Joint venture | 100 | - | lightreading.comtechblog.comsoc.org |
| 2025-05-13 | Seller | HUMAIN | Other | - | 5 | press.aboutamazon.com |
Headline values are as announced and are not additive. See all compute deals.
Quantum computers operated
- AWS Ocelot (Superconducting, operational, 9 qubits)
Sites operated
| Site | Location | Cooling | Commissioned |
|---|---|---|---|
| AWS Project Rainier (St. Joseph County, Indiana) | St. Joseph County, IN, USA | - | 2025 |
| AWS Canton Madison Mega Site | Canton, MS, USA | - | 2025 |
| AWS Ridgeland Data Center Campus | Ridgeland, MS, USA | - | - |
| Cumulus Data Campus (Salem Township) | Salem Township, PA, USA | - | - |
Related news
- Amazon to acquire two data centers in Hong Kong · Data Center DynamicsAmazon is converting two data-center buildings it already leases in Hong Kong -- iTech Tower 3.1 and 3.2 -- into owned assets, paying landlord Grand Ming a minimum of about US$278m (up to US$312m) as Grand Ming sells to deleverage. It's a small but checkable data point on a hyperscaler shifting from leasing to owning facility real estate in a supply-constrained market, distinct from new capacity coming online.
- Prysmian, Rio Tinto to supply Amazon data center in Ohio with low-carbon aluminum · Data Center DynamicsA materials-supply link in Amazon's Ohio buildout: Prysmian will manufacture electrical cabling from Rio Tinto and Alcoa's Elysis low-carbon aluminum for an unnamed AWS data center, the second Amazon-Rio Tinto low-carbon materials deal after a copper agreement in January. It's a small but checkable data point on hyperscalers extending decarbonization commitments into power-delivery hardware, not just chips.
- AWS activates Project Rainier, a nearly 500,000-chip Trainium2 cluster for Anthropic · About AmazonRainier is the largest non-NVIDIA training deployment in this dataset by chip count and the clearest evidence yet that a hyperscaler's own silicon, not merchant GPUs, can anchor a frontier model training run. AWS says Anthropic is headed toward more than 1 million Trainium2 chips by year end, which would make this entry's chip count stale within months rather than years.
Further reading
Inside a node
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.
Operating a supercomputer
Acceptance testing, why "installed" and "in production" are six months apart, and the machine lifecycle.