OpenAI and Microsoft are recutting their partnership so the AI company can sell its technology across any cloud platform rather than remaining tightly bound to a single provider. The refreshed agreement limits how much revenue OpenAI must share with Microsoft through 2030, introducing a hard cap instead of an open-ended commitment. At the same time, Microsoft keeps access to OpenAI’s core technology, but only under a non-exclusive licence that now runs until 2032.
Under the new structure, Microsoft continues as OpenAI’s primary cloud partner and retains the right to bring new OpenAI products to market before rivals. OpenAI can now deliver its full product suite to customers on any major cloud, opening doors to competing infrastructure providers such as Amazon. Microsoft stops paying revenue share back to OpenAI, reversing part of the original economic flow between the companies. OpenAI’s payments to Microsoft continue at the same % rate until 2030, but those outflows are now subject to a clearly defined ceiling.
Microsoft’s backing helped propel OpenAI to the centre of the AI boom. That support gave OpenAI the compute capacity and distribution muscle needed to scale large models quickly, but also came with restrictive commercial boundaries. Tensions reportedly escalated so sharply in 2025 that OpenAI explored involving competition authorities to escape elements of the contract.
