2 min read
Originally, Microsoft was OpenAI’s exclusive provider of data center infrastructure for training and running its AI models. However, this arrangement has changed.
Microsoft has now entered a new agreement with OpenAI, granting it the “right of first refusal” on any new cloud computing capacity required by OpenAI. This means Microsoft will have the first opportunity to host OpenAI’s AI workloads in the cloud. If Microsoft chooses not to, OpenAI is free to partner with other providers.
This shift coincides with OpenAI’s announcement of “Stargate,” a major new AI infrastructure partnership involving SoftBank, Oracle, and others, signaling OpenAI’s move toward a more diversified cloud strategy.
In a recent blog post, Microsoft stated, “OpenAI recently made a new, large Azure commitment that will continue to support all OpenAI products as well as training.” The post also highlighted that Microsoft has approved OpenAI’s plan to develop additional capacity, “primarily for research and training of models,” to further bolster their collaboration.
Reportedly, OpenAI and Microsoft, its close partner and significant investor, have grown increasingly tense over compute capacity, which OpenAI has claimed is causing product delays. Under pressure from shareholders, Microsoft allowed OpenAI to sign a capacity deal with Oracle in June.
Microsoft reaffirmed in its blog post that “key elements” of its long-standing cooperation with OpenAI, such as access to OpenAI’s intellectual property, revenue-sharing agreements, and exclusivity on OpenAI’s APIs, will continue until 2030.
This suggests that OpenAI doesn’t reach artificial general intelligence (AGI) based on the definition agreed upon by the two companies before that time. Reportedly, the two companies have agreed that Microsoft will no longer have access to OpenAI’s technology once the latter creates artificial intelligence systems that may earn $100 billion or more in earnings.
Rumour has it that OpenAI is thinking about trying to get extra money from Microsoft by nullifying the deal.
In addition to being accessible through the Azure OpenAI Service, the blog post states that the OpenAI API is proprietary to Azure and operates on Azure. “This agreement means customers benefit from having access to leading models on Microsoft platforms and direct from OpenAI.”
