OpenAI has signed a huge $38 billion deal with Amazon Web Services that will change the way artificial intelligence works. This strategic partnership, announced on November 3, 2025, is one of the largest cloud computing deals ever and marks a major step forward in OpenAI’s infrastructure strategy beyond its long-standing partnership with Microsoft. With this partnership between OpenAI and Amazon AWS, OpenAI can now quickly access “hundreds of thousands” of advanced Nvidia GPUs in Amazon’s data centers.
The seven-year agreement is part of OpenAI’s efforts to get the huge computing power it needs to build more advanced AI models faster. The timing of this partnership between OpenAI and Amazon AWS is very important. Last week, OpenAI restructured to generate revenue and relinquished Microsoft’s exclusive right to negotiate cloud deals. This new freedom has allowed the AI pioneer to expand its infrastructure partnerships to achieve its ambitious computing goals.
How the OpenAI Amazon AWS Partnership Reshapes Cloud AI
The OpenAI that just started OpenAI can use Amazon’s EC2 UltraServer infrastructure right away, thanks to its partnership with Amazon AWS. This infrastructure has large clusters of Nvidia’s newest Blackwell GPUs. These resources will help train new AI models and keep services like ChatGPT running in real time. AWS will set up dedicated resources for OpenAI, with full deployment planned by the end of 2026 and possible growth through 2027 and beyond.
This partnership between OpenAI and Amazon Web Services is a major change for both companies. For OpenAI, it means they won’t have to rely on Microsoft Azure as much and will have more resources to continue their AI research. Sam Altman, the CEO of OpenAI, said, “Scaling frontier AI needs a lot of reliable compute.” “Our partnership with AWS makes the broad compute ecosystem that will power this next era and bring advanced AI to everyone even stronger.” Getting OpenAI as a direct customer is a big win for Amazon in the hyperscale cloud market.
The computing power this OpenAI-Amazon AWS partnership provides you is mind-boggling. AWS says the infrastructure can grow to “tens of millions of CPUs,” giving OpenAI more power than ever to train larger models and handle more user queries. The architectural design groups Nvidia’s advanced GB200 and GB300 GPUs together using Amazon’s specialized UltraServers. This results in networks best suited for AI workloads with the lowest latency.
OpenAI is making several major infrastructure commitments right now, including a $300 billion agreement with Oracle and a $250 billion deal with Microsoft. This partnership with Amazon AWS is one of them. The table below shows how these partnerships all work together to make up OpenAI’s big plan for its computing infrastructure:
| Provider | Commitment Value | Key Resources | Timeline |
|---|---|---|---|
| Amazon AWS | $38 billion | Hundreds of thousands of Nvidia GPUs, EC2 UltraServers | Multi-year capacity through 2026-2027 |
| Microsoft | $250 billion | Azure cloud infrastructure | Continuing partnership post-restructuring |
| Oracle | $300 billion | Cloud computing capacity | Multi-year agreement |
| Nvidia | $100 billion | AI chips and systems | Part of broader spending commitments |
Even though this new OpenAI-Amazon AWS partnership is big, industry experts say there could be problems. The huge financial commitments have raised questions about whether AI infrastructure spending can sustain itself, especially given that OpenAI’s annual revenue is reportedly around $13 billion.
Some people who watch the market think these huge investments could signal an AI bubble, since so much money is going into technology that hasn’t yet shown clear returns. This partnership between OpenAI and Amazon AWS will be very good for AWS. The cloud provider has been trying to catch up to Microsoft in generative AI, and adding OpenAI to its client list is a big step forward in that direction. The announcement sent Amazon’s stock soaring 4%, a record high. This shows how excited the market was about the deal.
OpenAI’s models are already available on Amazon Bedrock, AWS’s managed service for accessing foundation models, as part of the OpenAI-Amazon AWS partnership. Bedrock lets thousands of AWS customers, including well-known companies such as Peloton, Thomson Reuters, and Comscore, use OpenAI models for a wide range of tasks, from coding and solving math problems to conducting scientific research.
