Microsoft and OpenAI sign amended, less-exclusive partnership

Microsoft and OpenAI signed an amended agreement that unwound key exclusivity terms from their original 2019 partnership, following the companies’ October 2025 recapitalisation of OpenAI into a public-benefit corporation. Microsoft said it would remain OpenAI’s “primary” cloud provider, with new OpenAI products launching first on Azure, but OpenAI gained the right to serve products through any cloud provider — including competitors such as Google and Amazon — a change from an earlier arrangement that had given Microsoft exclusive access to OpenAI’s models and technology until the achievement of artificial general intelligence.

The revenue arrangement between the two companies also changed. Microsoft’s payments to OpenAI tied to Azure OpenAI Service usage ended. OpenAI’s own payments to Microsoft continue on the same percentage basis through 2030, but the total amount owed is now capped rather than open-ended, meaning the payments will stop once a fixed ceiling is reached regardless of OpenAI’s future revenue growth. Microsoft’s licence to OpenAI’s intellectual property — its models and products — was extended to run through 2032, but converted from exclusive to non-exclusive, so OpenAI can license the same technology elsewhere.

Both companies described the change as providing long-term clarity while giving each more room to pursue independent commercial relationships — Microsoft with rival model developers, OpenAI with rival cloud providers. The amendment followed years in which the relationship, once presented as tightly aligned, had shown growing friction as OpenAI expanded its own compute deals with Oracle, Amazon and others and pursued a corporate structure change that reduced Microsoft’s formal claim on OpenAI’s future profits. The announcement came the same day OpenAI’s founding dispute with Elon Musk went to trial, a case OpenAI had separately identified as a business risk.