As part of the Amazon-OpenAI deal, the companies plan a “stateful runtime environment” for AWS, letting agents carry context forward to handle ongoing projects
The later expansion of OpenAI model availability through AWS reinforces the arc: AWS is positioning to provide both the models and the operating layer needed to put them into sustained enterprise work.
First-order effects
AWS gains a planned agent runtime designed to preserve project context across sessions, rather than treating each model interaction as isolated.
OpenAI’s agents would have a purpose-built AWS environment for ongoing work, deepening the companies’ technical integration alongside their compute arrangement.
Second-order effects
Enterprise buyers evaluating agents will increasingly compare cloud offerings on persistent context and runtime controls, not only on model selection or raw inference access.
Rival cloud and agent platforms face pressure to package memory, project continuity, and governance capabilities as core infrastructure for long-running agents.
Third-order effects
If these runtime layers become standard, differentiation in enterprise AI may shift from access to individual models toward control of the state, workflows, and infrastructure surrounding them.
The arrangement points toward a more vertically integrated agent stack, though adoption will depend on whether organizations trust persistent agent context for production work.
The trend: Cloud providers are moving from hosting AI models to owning the stateful operating environments that make agents usable in recurring business workflows.
This is smart from AWS Azure had exclusivity for the OpenAI API, but the Stateful Runtime Environment seems like a product API not model API, so AWS can host/offer it without interfering with the Azure model API exclusivity
Early positioning beats late capital every time. Amazon just invested $50 billion in OpenAI for 6 percent ownership and distribution rights. Microsoft invested $13.8 billion starting in 2019 and owns 23.5 percent with exclusive IP, revenue sharing, and API control locked
Shame doesn't help. — You could say the same of anyone that still uses, works at a place that uses, and/or works for Microsoft, Google/Gmail, Instagram, Facebook, AWS, etc.
This is not an extension of the November deal. It is a fundamentally different relationship. The November 2025 agreement was straightforward. OpenAI committed $38 billion over seven years to rent NVIDIA GPU capacity on AWS. EC2 UltraServers, GB200s and GB300s, pure infrastruc…
Massive: “Amazon will also invest $50 billion in OpenAI, starting with an initial $15 billion investment and followed by another $35 billion in the coming months when certain conditions are met.”
Excited to partner with AMZN. We're co-building a new Stateful Runtime product for agentic applications on Bedrock powered by OpenAI models, expanding Frontier via AWS, scaling with 2GW of Trainium compute, and creating custom models for Amazon apps. https://openai.com/...
Excited about our new strategic partnership with @OpenAI. Developers and companies of all kinds are eager to run services powered by OpenAI models on AWS, and our unique collaboration will provide a stateful runtime environment for them that's powered by OpenAI's frontier intell…
We are excited to partner with Amazon to bring a new generation of products to market, especially around new enterprise products like the stateful runtime environment. We are also very excited to make great use of Tranium.
The underlying story of OpenAI's jaw-dropping $110 billion raise is really about Amazon, and how as a “cloud leader” it slipped when it came to AI, and recognizing the opportunity early compared to Azure. Now it's costing the company billions. om.co/2026/02/27/a...