Amazon Web Services (AWS) and OpenAI have officially entered into a massive $38 billion, multi-year strategic partnership to power and scale OpenAI’s next generation of artificial intelligence (AI) workloads. The collaboration will see AWS providing its extensive cloud and GPU infrastructure to accelerate OpenAI’s research and product development.

Under the seven-year agreement, OpenAI will gain immediate access to AWS’s advanced computing resources – including hundreds of thousands of NVIDIA GPUs and the ability to expand up to tens of millions of CPUs. This infrastructure will support agentic workloads, training, and large-scale inference operations for models like ChatGPT and future OpenAI systems.

The growing demand for AI compute power has become one of the defining challenges of the industry. By partnering with AWS, OpenAI gains access to a highly scalable, secure, and high-performance environment that can handle the increasing complexity of frontier models.

AWS has decades of expertise managing large-scale clusters – some reaching up to 500,000 chips – with optimized networking and reliability for mission-critical workloads.

The newly designed infrastructure will include Amazon EC2 UltraServers with NVIDIA GB200 and GB300 GPUs, unified via a low-latency architecture. This configuration is engineered to deliver maximum AI processing efficiency, enabling OpenAI to perform both training and inference workloads seamlessly.

OpenAI is expected to begin utilizing AWS resources immediately, with full deployment targeted before the end of 2026. The agreement also includes flexibility for additional capacity expansion through 2027 and beyond, ensuring long-term scalability as AI demands continue to grow.

This partnership builds upon previous collaborations between the two companies. Earlier in the year, OpenAI’s open-weight foundation models became available on Amazon Bedrock, expanding AI accessibility for AWS customers worldwide. The new deal represents a deepening of that relationship, uniting AWS’s cloud leadership with OpenAI’s innovation in generative AI.