Anthropic Strategic Investments
| Investor / Partner | Investment Amount ($) | Date Announced | Anthropic’s Key Commitments / Use of Funds |
| Microsoft | $5 Billion | Nov 2025 | Committed to purchasing $30 Billion of Microsoft Azure compute capacity over several years. Claude will be available on Azure. |
| NVIDIA | $10 Billion | Nov 2025 | Collaboration on design and engineering to optimize Claude models for NVIDIA’s Grace Blackwell and Vera Rubin systems (up to 1 gigawatt of compute capacity). |
| Amazon | Up to $4 Billion ($1.25B initially, with a commitment for an additional $2.75B) | Sep 2023 / Mar 2024 | Anthropic committed to making Amazon AWS its primary cloud provider, including utilizing AWS Trainium chips for model training. |
| Google (Alphabet) | Up to $2 Billion ($500M initially, with a commitment for an additional $1.5B) | Oct 2023 | Anthropic committed to building its AI systems on Google Cloud Platform (GCP), including leveraging Google’s TPUs (Tensor Processing Units). |
Anthropic – Total Compute Secured Across Cloud Providers
| Announcement Date | Cloud / Infrastructure Partner | Compute Secured (GW) | Equity Swap / Investment Component? | Notes |
|---|---|---|---|---|
| Apr 21, 2026 | Amazon / AWS | Up to 5.0 GW | Yes – Amazon investing additional $5B now + up to $20B later | Anthropic commits $100B+ AWS spend over 10 years; includes Trainium / Graviton capacity |
| Apr 6, 2026 | Google + Broadcom | Multiple GW (market reports ~3.5 GW) | No disclosed equity swap | Next-gen TPU capacity starting 2027 |
| Nov 2025 | Microsoft Azure + NVIDIA systems | Up to 1.0 GW | Yes – Microsoft + NVIDIA strategic investment reported | Anthropic to buy $30B Azure compute |
| Oct 2025 | 1.0+ GW by 2026 | Yes – prior Google equity investment rounds | Access to up to 1 million TPUs (reported) | |
| Oct 2025 / ongoing | Amazon Project Rainier | Capacity tied to ~1M+ Trainium2 chips | Yes (part of broader Amazon stake) | Large dedicated cluster for Claude workloads |
Invest Alpha 