The Next Chapter of the Microsoft–OpenAI Partnership

Since 2019, Microsoft and OpenAI have shared a common vision: to advance artificial intelligence responsibly and make its benefits accessible to everyone. What began as an early investment in a research-driven organization has now evolved into one of the most influential collaborations in the technology industry.

In a significant move, both companies have signed a new definitive agreement that strengthens their long-standing relationship, reinforces mutual trust, and sets the stage for long-term growth in the AI landscape. This marks a pivotal moment in the ongoing story of the Microsoft–OpenAI partnership.

A Strengthened Partnership with a New Structure

Microsoft is supporting OpenAI’s decision to transition into a public benefit corporation (PBC) under the newly formed OpenAI Group PBC. Following the recapitalization, Microsoft now holds an investment valued at approximately $135 billion, representing about 27% ownership on an as-converted diluted basis. This includes shares across all stakeholders—employees, investors, and the OpenAI Foundation.

Before the recapitalization and recent funding rounds, Microsoft’s stake stood at 32.5% in OpenAI’s for-profit arm. The updated structure ensures both organizations continue to benefit from shared innovation while pursuing their individual goals within the AI ecosystem.

Preserving the Core of the Collaboration

Despite the structural changes, the essential elements that have made this partnership successful remain intact. OpenAI continues to serve as Microsoft’s frontier model partner, while Microsoft maintains exclusive intellectual property (IP) rights and Azure API exclusivity until the development of Artificial General Intelligence (AGI).

This framework has been central to the partnership’s success, enabling both companies to deliver cutting-edge AI technologies through the Azure platform, such as the integration of OpenAI’s models across Microsoft products including Copilot, Azure AI Studio, and Microsoft 365.

Key Updates in the New Agreement

The revised agreement introduces a number of changes that refine how Microsoft and OpenAI will collaborate and innovate moving forward.

1. Independent Verification of AGI

Once OpenAI declares that it has achieved AGI, this claim will be verified by an independent expert panel. This step ensures transparency and accountability as AI capabilities advance toward more generalized intelligence.

2. Extended IP Rights

Microsoft’s IP rights for models and products are now extended through 2032, including those developed after AGI—subject to strict safety guardrails.

In addition, Microsoft’s rights to research IP (confidential methods used in model development) will remain until either AGI is independently verified or through 2030, whichever comes first. However, these research IP rights exclude elements such as model architecture, model weights, inference code, and fine-tuning code.

Microsoft retains rights to non-research IP but will not have claims on OpenAI’s consumer hardware.

3. Expanded Collaboration Flexibility

OpenAI is now free to co-develop products with third parties. API-based products developed through these collaborations will remain exclusive to Microsoft Azure, while non-API products may be deployed on other cloud providers.

This change reflects a growing openness in the AI ecosystem while ensuring Azure remains a central part of OpenAI’s infrastructure.

4. Independent Pursuit of AGI

Microsoft now has the freedom to pursue AGI independently or collaborate with other partners. If it uses OpenAI’s IP in this pursuit before AGI is declared, certain compute thresholds will apply—these thresholds are much higher than those currently required for training leading AI models.

5. Revenue Share and Compute Agreements

The revenue-sharing arrangement between the two organizations will continue until AGI is independently verified. However, payments will now be made over a longer duration, allowing both companies to manage long-term investments effectively.

Additionally, OpenAI has agreed to purchase an additional $250 billion in Azure services, further deepening its reliance on Microsoft’s cloud infrastructure. Microsoft, in turn, will no longer hold the right of first refusal to be OpenAI’s compute provider.

6. New Operational Flexibility for OpenAI

OpenAI will now be able to:

• Offer API access to U.S. government national security clients across any cloud provider.

• Release open-weight models that meet specific capability and safety criteria.

These updates emphasize OpenAI’s growing role as both a research pioneer and a commercial platform provider.

Looking Ahead: A Shared Vision for Responsible AI

As Microsoft and OpenAI move into this new phase, the partnership stands as a model for how large-scale AI collaborations can evolve responsibly. Both organizations remain committed to developing technologies that enhance productivity, creativity, and problem-solving—while maintaining safety and ethical standards.

This agreement ensures that innovation continues to flow both ways: Microsoft gains extended access to OpenAI’s most advanced models, and OpenAI benefits from Microsoft’s unmatched cloud infrastructure and enterprise reach.

Together, they aim to accelerate progress toward responsible AI and expand the real-world impact of their technologies—ushering in a future where AI serves as a tool for empowerment, inclusion, and global progress.