Microsoft, which once held an exclusive position as the sole provider of cloud infrastructure for OpenAI, is now adapting to a changing landscape.
Following the announcement of OpenAI’s ambitious new AI infrastructure project, dubbed ‘Stargate,’ in collaboration with SoftBank, Oracle, and others, Microsoft has revealed a fresh agreement with OpenAI. Under this deal, Microsoft secures a “right of first refusal” on any new cloud computing capacity that OpenAI may require. In essence, Microsoft retains the privilege of being OpenAI’s first choice for hosting its AI workloads. However, if Microsoft cannot fulfill the demand, OpenAI is free to seek alternatives from other cloud providers.
In an official blog post, Microsoft stated, “OpenAI has recently made a substantial commitment to Azure, which will continue to support all OpenAI products as well as AI model training.” To further bolster OpenAI’s growth, Microsoft has approved OpenAI’s efforts to build additional computing capacity, primarily for research and model training purposes.
OpenAI has previously cited insufficient computing resources as a major bottleneck, delaying product rollouts and causing friction in its collaboration with Microsoft. In June, mounting pressure from Microsoft’s shareholders led to OpenAI striking a deal with Oracle to secure additional cloud capacity.
Despite these shifts, Microsoft has assured that the “key elements” of its long-standing partnership with OpenAI will remain intact until 2030. These elements include access to OpenAI’s intellectual property, revenue-sharing agreements, and exclusive hosting of OpenAI’s API services on Azure.
However, the partnership’s stability is contingent upon a crucial factor: OpenAI’s progress towards artificial general intelligence (AGI). As per the agreement between the two companies, if OpenAI reaches the milestone of developing AI systems capable of generating $100 billion in profits, Microsoft could potentially lose its exclusive access to OpenAI’s cutting-edge technology.
Reports indicate that OpenAI is evaluating options to renegotiate or nullify aspects of the agreement, potentially seeking additional funding from Microsoft to fuel its ambitious goals.
In its blog post, Microsoft emphasized that the OpenAI API remains exclusive to Azure, running within its cloud ecosystem and available via the Azure OpenAI Service. “This agreement ensures that customers have access to leading AI models both on Microsoft platforms and directly from OpenAI,” the company concluded.
