Microsoft and OpenAI have rewritten key terms of their landmark AI partnership, giving the ChatGPT maker more freedom to distribute its products through rival cloud providers while preserving Microsoft’s long-term access to OpenAI technology.

The amended agreement, announced Monday by both companies, ends Microsoft’s exclusive license to OpenAI’s models and products. OpenAI can now serve customers across any cloud provider, a change that opens the door to broader use of its technology on platforms such as Amazon Web Services and Google Cloud. Microsoft, however, remains OpenAI’s primary cloud partner, and OpenAI products will continue to launch first on Azure unless Microsoft cannot or chooses not to support the necessary capabilities.

The shift is one of the most consequential changes yet to the partnership that helped define the current AI boom. Microsoft invested heavily in OpenAI beginning in 2019 and integrated OpenAI models into products including Copilot, Azure AI services and Microsoft 365. OpenAI, in turn, relied on Microsoft’s infrastructure and capital to scale the models behind ChatGPT and its enterprise AI offerings.

Under the revised terms, Microsoft will continue to hold a license to OpenAI intellectual property for models and products through 2032, but that license is now non-exclusive. OpenAI will continue making revenue-share payments to Microsoft through 2030 at the same percentage, but those payments will be subject to an undisclosed cap and will no longer depend on whether OpenAI reaches specific technology milestones. Microsoft will no longer pay OpenAI a revenue share.

The companies framed the agreement as a way to simplify the relationship while keeping the core partnership intact. In near-identical statements, Microsoft and OpenAI said the amended deal provides “flexibility” and “certainty” as both companies continue work on large-scale AI platforms, data centers, next-generation silicon and cybersecurity applications.

The practical effect is that OpenAI gains more room to court enterprise customers that already run large workloads on Amazon, Google or other cloud providers. That matters because many large companies prefer to buy AI services through the cloud platforms they already use for data storage, security, compliance and application development. A more open distribution model could make it easier for OpenAI to compete against Anthropic, Google, Meta and other model providers in corporate AI deployments.

The timing is also important. Demand for AI compute has become one of the defining constraints in the industry. Frontier model developers require massive data-center capacity, specialized chips and long-term energy commitments. By loosening exclusivity, OpenAI can seek infrastructure from multiple providers rather than relying primarily on Microsoft’s ability to supply capacity through Azure.

Reuters reported that Microsoft’s earlier investment in OpenAI totaled $13 billion since 2019 and that tensions had risen as OpenAI sought more freedom to strike cloud deals with Microsoft rivals. Reuters also reported that Microsoft’s right to make OpenAI products available first on Azure remains in place, along with OpenAI’s commitment to use at least $250 billion in Azure services by 2032.

Amazon quickly signaled interest in the new opening. Reuters reported that Amazon CEO Andy Jassy said OpenAI models would become available directly to developers on AWS in the coming weeks, with more details expected at a San Francisco event. The Associated Press also reported that Jassy described the announcement as “very interesting” and said OpenAI models would soon be available through Amazon Bedrock, AWS’s platform for accessing AI models.

For Microsoft, the deal reduces ambiguity around its financial relationship with OpenAI while allowing the Redmond-based company to continue developing a broader AI portfolio. Microsoft has been building its own models, expanding Copilot across business software and working with other AI companies, including Anthropic. A non-exclusive arrangement may help Microsoft argue that its OpenAI relationship does not lock up the enterprise AI market.

The agreement may also matter for regulators. Microsoft’s OpenAI partnership has drawn scrutiny in the United States, the United Kingdom and Europe over whether the alliance gives Microsoft too much influence over a leading AI developer or an unfair advantage in cloud and enterprise AI markets. Ending exclusivity does not remove all regulatory questions, but it gives both companies a clearer argument that OpenAI’s products can reach customers through multiple infrastructure providers.

For OpenAI, the revised terms could support a longer-term push toward public-market readiness. The Associated Press reported that analysts saw the change as reducing barriers from OpenAI’s original Microsoft partnership and strengthening the company’s path toward a possible initial public offering. OpenAI has been moving from its nonprofit origins toward a more capital-intensive structure designed to finance large-scale AI development.

The new agreement does not mean the Microsoft-OpenAI alliance is ending. Microsoft remains a major OpenAI shareholder, keeps long-term access to OpenAI technology and continues to serve as the company’s primary cloud partner. OpenAI still benefits from Azure’s global infrastructure and Microsoft’s enterprise reach. But the balance of power has changed: OpenAI now has more commercial flexibility, and Microsoft has more freedom to pursue its own AI roadmap without being as tightly bound to a single model supplier.

The result is a looser but still strategically important partnership. For enterprise customers, it could mean more ways to buy and deploy OpenAI products. For cloud providers, it raises the stakes in the race to host AI workloads. For the AI industry, it signals that the biggest model developers are moving toward multi-cloud, multi-partner strategies as the cost and scale of frontier AI continue to rise.