OpenAI has partnered with Amazon Web Services (AWS) to make its AI models available on the cloud platform for the first time. The announcement coincides with OpenAI’s release of two open-weight reasoning models with performance similar to its o-series models. These models can now be accessed through AWS’s Bedrock and SageMaker AI services, enabling enterprises to integrate OpenAI’s technology into hosted applications or develop custom AI solutions.
Previously, AWS was best known as a key backer of Anthropic’s Claude models and offered AI services from Cohere, DeepSeek, Meta, Mistral, and its own in-house models. The addition of OpenAI places AWS on par with competitors offering leading AI models. While OpenAI models have been available via Microsoft Azure, this marks a significant shift, especially as AWS has faced criticism from Wall Street analysts for lagging in generative AI adoption compared to Microsoft and Google.
Amazon CEO Andy Jassy has recently faced questions about losing ground in AI. The AWS–OpenAI deal comes as Microsoft continues to deepen its integration with OpenAI, and Oracle reportedly signs a massive $30 billion annual agreement with OpenAI for data center services. For OpenAI, this partnership diversifies its cloud presence amid a reportedly strained relationship with Microsoft, giving it access to AWS’s vast enterprise customer base.
The move also aligns with OpenAI’s open-source approach, as the new models are licensed under Apache 2.0, enabling businesses to experiment freely. It also counters Meta, which has scaled back its plans to open-source future AI systems.
The collaboration strengthens OpenAI’s competitive positioning while giving AWS customers direct access to cutting-edge AI models, marking a pivotal moment in cloud and AI industry competition.
