For the first time, OpenAI’s newly released models are now available on Amazon Web Services (AWS), through Bedrock and SageMaker.
The models are comparable in performance to OpenAI’s o-series models, and though users can already download them from Hugging Face, the AWS integration brings enterprise-ready access.
OpenAI has fully approved this offering, according to Dmitry Pimenov, OpenAI’s product lead. Amazon also confirmed this partnership in a statement shared with TechCrunch.
AWS’s Plan
AWS offers these models as part of its existing AI infrastructure. Bedrock allows developers to build generative AI applications with minimal effort.
And SageMaker supports those looking to train or build custom AI models, mainly for analytics and research.
This integration puts OpenAI’s capabilities directly into the hands of AWS customers. It allows companies to experiment, deploy, and scale OpenAI models without leaving the AWS ecosystem.
Amazon described it as consistent with how it previously offered other open models. One example is DeepSeek-R1, which joined the Bedrock catalog earlier this year.
Implications
This collaboration is OpenAI’s first technical partnership with AWS. More importantly, it places Amazon in closer competition with Microsoft, OpenAI’s primary cloud partner until now.
Microsoft’s Azure platform still holds the deepest integration with OpenAI. It offers access to ChatGPT and other models, including enterprise-ready tools optimized for Windows.
However, the AWS announcement is a significant change. It gives OpenAI a wider reach and reduces its dependence on a single provider.
This is especially relevant now as OpenAI and Microsoft are reportedly renegotiating the terms of their relationship.
By teaming up with AWS, even in a limited way, OpenAI gains some leverage.
Mounting Pressure
CEO Andy Jassy has faced increasing pressure from investors and analysts. During Amazon’s Q2 2025 earnings call, Jassy was repeatedly asked about the company’s AI strategy.
Questions focused on whether AWS was falling behind Microsoft and Google. Doug Anmuth of JPMorgan asked why AWS was growing more slowly than its closest competitors.
Brian Nowak of Morgan Stanley echoed those concerns; he noted a rising belief on Wall Street that AWS is losing ground in the generative AI space.
Jassy defended AWS; he stated that Microsoft, often considered the “second player,” remains only 65% the size of AWS.
However, many saw his response as insufficient, as market perception continued to trend in Microsoft’s favor.
This new partnership may be a stronger answer. By teaming up with OpenAI, AWS adds a marquee name to its model lineup.
Oracle
Interestingly, the AWS-OpenAI announcement comes just after Oracle revealed a separate deal.
Oracle signed a $30 billion annual agreement with OpenAI. Under this deal, Oracle will provide data center services for OpenAI’s infrastructure.
The size of the deal significantly exceeds what OpenAI pays all other cloud providers combined.
Until now, AWS had been notably absent from such large-scale OpenAI contracts. This new technical partnership, while smaller in scope, allows AWS to reenter the conversation.
OpenAI
OpenAI also benefits from the AWS collaboration: it gains access to thousands of enterprise users already working inside AWS.
Existing customers can now experiment with OpenAI’s models using familiar tools. This could accelerate adoption.
It could also expand OpenAI’s commercial reach without compromising its independence.
Additionally, the models were launched under the Apache 2.0 license. This permissive license allows both personal and commercial use.
It promotes open collaboration while enabling large-scale enterprise deployment.
Meta
The announcement also affects Meta. Meta’s Llama models gained popularity due to their open-access approach.
But Meta recently admitted that it may stop open-sourcing some of its upcoming “superintelligence” models.
In contrast, OpenAI’s decision to release its models openly and distribute them through AWS reinforces its appeal among developers.
It also positions OpenAI as a more consistent supporter of open access and may prompt developers to rethink their platform choices.
If Meta continues to limit access, OpenAI could gain favor with engineers, researchers, and startups seeking flexibility.
