Microsoft's partnership with OpenAI is undergoing a significant shift. With OpenAI announcing agreements with companies like SoftBank and Oracle under the Stargate protocol, Microsoft is no longer the exclusive provider of data center infrastructure for OpenAI. Under the new agreement, Microsoft has gained a "right of first refusal" on OpenAI's additional cloud computing capacity, meaning that if Microsoft cannot meet demand, OpenAI can seek support from other cloud service providers.
Microsoft confirmed this change in a blog post, stating that OpenAI has made new important commitments to Azure and will continue to support its product and training needs. Additionally, Microsoft has granted OpenAI the authority to independently build additional computing capacity, primarily for model research and training.
This shift stems from OpenAI's challenges with a shortage of computing resources. The company had previously delayed product launches, attributing them to insufficient computing power, which reportedly became a source of tension between it and its major investor, Microsoft. Under pressure from shareholders, Microsoft allowed OpenAI to collaborate with Oracle in June of this year to increase computing resources.
Nonetheless, Microsoft emphasizes that the core partnership will continue until 2030, including access to OpenAI's intellectual property, revenue-sharing arrangements, and exclusive rights to operate the OpenAI API. Notably, this agreement comes with a condition: if OpenAI achieves general artificial intelligence (AGI) capable of generating $100 billion in profit before that time, Microsoft will lose its rights to use that technology. Reports indicate that OpenAI is considering removing this clause to secure more funding from Microsoft.
Microsoft particularly emphasizes that the OpenAI API will continue to operate exclusively on the Azure platform, allowing customers to access leading models through the Azure OpenAI service or directly from OpenAI.
