OpenAI will be free to sell its services across AWS and Google Cloud after striking a new deal with Microsoft that caps payments to the software giant at 20% through 2030. Microsoft remains OpenAI’s primary cloud partner and keeps a non‑exclusive licence to its models through 2032; the revised pact also removes the old AGI-trigger clause.
What changed in the partnership
The two firms rewrote key commercial terms of a partnership that has been central to both businesses since 2019. Under the new arrangement, OpenAI will pay Microsoft 20% of specified revenues; those payments are now subject to a total cap and will continue only through 2030. The agreement removes earlier language tied to commercial rights in the event of artificial general intelligence.
OpenAI can now distribute all of its products across other cloud providers, including Amazon Web Services and Google Cloud, while Microsoft remains the primary cloud partner. New OpenAI releases will ship first on Azure unless Microsoft cannot or will not support the necessary capabilities.
The revised terms also change how money flows. Historically, when customers bought access to OpenAI-powered services through Azure, Microsoft paid a revenue share back to OpenAI. Under the updated deal, Microsoft will no longer make those payments to OpenAI, while OpenAI will continue to pay Microsoft its share where the new mechanics apply.
Microsoft will keep a licence to OpenAI’s intellectual property for models and products through 2032, but that licence is non-exclusive. Microsoft has invested heavily in OpenAI since 2019 and the companies said the relationship remains strategic even as the terms evolve.
Why the change matters to cloud customers
The biggest practical shift is distribution. Enterprises that run most of their workloads on AWS or Google Cloud can now buy OpenAI-powered products from those platforms rather than being forced to port data and services to Azure, likely lowering integration costs and speeding adoption.
OpenAI has been building its own data-centre capacity and co-developing runtime technology with other providers. The amended terms remove a contractual obstacle that could have complicated OpenAI’s multi-cloud arrangements.
For developers and larger customers, multi-cloud availability can improve performance and compliance: products can be located closer to users or to customer data, cutting latency and simplifying local data rules. Availability may still be staggered, however, since Microsoft gets first access on Azure.
What Microsoft keeps — and what it gives up
Microsoft relinquishes exclusive rights it once held, including the special AGI-trigger clause. That clause has been removed to reduce long-term contractual uncertainty.
Microsoft retains several advantages: it remains the primary cloud partner so much of OpenAI’s cloud workload is likely to stay on Azure, it holds a licence to OpenAI’s IP through 2032, and it will continue to receive a portion of OpenAI revenue where the new mechanics apply. Market reaction was muted but slightly negative, with Microsoft shares dipping on the announcement.
Related Articles
- Microsoft simplifies Windows Insider to two channels — Experimental and Beta
- Ubuntu to add two AI feature tiers over the next year
- Letterboxd owner Tiny exploring sale after surge to 26 million users
Microsoft retains a non‑exclusive licence to OpenAI’s models through 2032, while OpenAI will pay Microsoft 20% of specified revenues under a capped arrangement running through 2030 as cloud exclusivity ends.
This article was created with AI assistance.
