AWS–OpenAI Deal

KM Ashmil

— 2 min read
AWS–OpenAI Deal
aws-openai
aws-openai

OpenAI has signed a major seven-year, $38 billion cloud partnership with Amazon Web Services (AWS), making AWS a crucial provider for OpenAI's AI workloads from November 2025 onward. This signifies a major strategic shift from its previously exclusive reliance on Microsoft Azure and demonstrates OpenAI’s move to diversify its cloud infrastructure for resilience and scalability.​

Details of the AWS–OpenAI Deal

  • The partnership, announced in November 2025, grants OpenAI direct access to state-of-the-art AWS infrastructure, including hundreds of thousands of Nvidia GPUs and ultra-high-performance EC2 UltraServers designed for AI model training and inference.​
  • AWS will deliver dedicated data center capacity for OpenAI’s needs, with all initial infrastructure deployed by the end of 2026 and further expansion possible into 2027 and beyond.​
  • The deal follows a recent organizational restructuring at OpenAI that allowed for greater operational and financial freedom, thereby enabling the shift to AWS.​

OpenAI’s Current Cloud Setup

  • As of November 2025, OpenAI is actively running critical workloads on AWS and leveraging its massive GPU clusters for language model training and real-time inference (e.g., powering ChatGPT).​
  • OpenAI is also engaged in partnerships with other providers: in 2025, it signed significant agreements with Oracle ($30 billion for data center power via the Stargate initiative) and Google Cloud (for access to TPU hardware and workload flexibility).​
  • Microsoft no longer holds exclusive hosting rights, though it remains an important financial and technological partner with some API products remaining Azure-exclusive.​

Strategic Implications

  • OpenAI aims to avoid vendor lock-in by creating a multi-cloud and multi-partner ecosystem, spreading its workloads across AWS, Oracle, Google Cloud, and Microsoft.​
  • This approach ensures scalability, reliability, and negotiating leverage while supporting the increasingly massive computational requirements for advanced AI models.​

OpenAI’s AI infrastructure is now powered primarily by AWS, supplemented by significant investments in Oracle and Google Cloud capabilities, marking the end of its Azure-only era and representing a fundamental shift towards diversified cloud computing for large-scale AI development.​

Read more