OpenAI brings its models to Amazon's cloud after ending exclusivity with Microsoft

In a significant shift, OpenAI has announced that its models will now be available on Amazon Web Services (AWS), marking the end of its exclusive partnership with Microsoft for cloud services. This development comes just a day after OpenAI restructured its relationship with Microsoft, enabling the AI firm to utilize any cloud platform for its products. AWS users will soon have the chance to experiment with OpenAI's models and its Codex agent for code writing through Amazon Bedrock, with services expected to be generally available in the coming weeks. During a launch event in San Francisco, AWS CEO Matt Garman emphasized that this move responds to customer demand for more flexible cloud options. Previously, developers had access to certain open-weight models from OpenAI, which were introduced to AWS in August. OpenAI CEO Sam Altman, who was unable to attend the event due to ongoing legal proceedings, shared his enthusiasm through a recorded message, thanking Garman and the AWS team for their partnership. The collaboration will also see the introduction of Amazon Bedrock Managed Agents, a service powered by OpenAI that allows for the creation of advanced customized agents capable of retaining memory from previous interactions. While Microsoft has been a vital provider of computing power to OpenAI since before the launch of ChatGPT in 2022, the AI company recently indicated that this long-standing relationship had constrained its ability to cater to enterprise clients, many of whom prefer using AWS. In a notable update, OpenAI and Microsoft revealed a new aspect of their collaboration that permits the AI firm to limit revenue share payments while serving customers across various cloud providers. Amazon CEO Andy Jassy expressed enthusiasm about the partnership in a recent post, indicating that further details would be disclosed soon. OpenAI and Amazon have been strengthening their ties, with a $38 billion commitment from AWS announced last November. This move came shortly after OpenAI had initially designated Microsoft Azure as the exclusive cloud provider for third-party application programming interfaces. Additionally, OpenAI plans to utilize AWS's custom Trainium chips for training its AI models. These developments follow reports indicating that OpenAI had not met internal user and revenue goals, leading to a decline in shares for AI hardware companies like Nvidia and Broadcom. In response to these reports, Altman and OpenAI CFO Sarah Friar issued a statement clarifying their commitment to maximizing compute resources and collaborating closely every day.