Microsoft No Longer OpenAI's Exclusive Cloud Provider

Shifting Dynamics in OpenAI's Cloud Infrastructure

Previously, Microsoft held the position of sole infrastructure provider for OpenAI, facilitating the training and operation of its artificial intelligence models. This arrangement has now evolved.

The announcement of Stargate, a substantial new AI infrastructure agreement involving SoftBank, Oracle, and other entities, coincides with Microsoft’s declaration of a revised agreement with OpenAI. This new arrangement grants Microsoft the “right of first refusal” regarding future OpenAI cloud computing capacity.

New Agreement Details

Essentially, Microsoft will be given the initial opportunity to host OpenAI’s AI workloads within its cloud environment. However, should Microsoft be unable to fulfill OpenAI’s requirements, the AI company is now free to explore alternative cloud providers.

Microsoft confirmed a significant new Azure commitment from OpenAI, designed to continue supporting all OpenAI products and their associated training processes. Furthermore, Microsoft has authorized OpenAI to develop additional capacity, primarily focused on research and model training.

Compute Capacity and Partnership Evolution

Compute capacity limitations have been cited by OpenAI as a factor in product delays. Reportedly, this scarcity has also created some tension between OpenAI and Microsoft, a close collaborator and significant investor.

Under shareholder influence, Microsoft allowed OpenAI to establish a deal with Oracle for supplementary capacity earlier in June. This move signaled a shift in the previously exclusive relationship.

Core Partnership Remains

Microsoft emphasized that the fundamental aspects of its long-term partnership with OpenAI, extending through 2030, remain intact. These include access to OpenAI’s intellectual property, revenue-sharing structures, and exclusive access to OpenAI’s APIs.

This continuity is contingent upon OpenAI not achieving artificial general intelligence (AGI), as defined by the two companies, before 2030. Should OpenAI develop AI systems generating profits exceeding $100 billion, Microsoft’s access to the company’s technology will be revoked, according to the reported agreement.

There are indications that OpenAI is contemplating a revision of the agreement, potentially to secure further funding from Microsoft.

API Exclusivity and Future Outlook

“The OpenAI API remains exclusive to Azure, operates on Azure infrastructure, and is also accessible through the Azure OpenAI Service,” Microsoft stated. “This arrangement ensures customers benefit from access to leading models on Microsoft platforms, as well as directly from OpenAI.”

Requests for further clarification have been directed to both OpenAI and Microsoft, and this article will be updated upon receiving a response.

TechCrunch offers a newsletter dedicated to AI! Sign up here to receive it weekly in your inbox every Wednesday.