Microsoft and OpenAI Enter New Chapter with $135 Billion Partnership Overhaul

Rabat – A new era has dawned in the extensive association between Microsoft and OpenAI, with partners signing a definitive agreement that redefines one of the most significant collaborations in the technology industry.

This new contract reaffirms their dedication to responsibly advance AI while facilitating greater independence in innovation.

The two companies have shared a vision to unlock the benefits of AI technology for a wide range of users and businesses since they first partnered in 2019.

What began as an initial investment in a research group has evolved into an extensive, multi-tiered deal valued at approximately $135 billion.

Following OpenAI’s recapitalization and the formation of OpenAI Group PBC (Public Benefit Corporation), Microsoft now holds an equity stake of 27% on an as-converted diluted basis, including employees, investors, and OpenAI Foundation stakeholders. Its equity stake previously stood at 32.5% before the latest funding rounds in OpenAI.

The new deal retains the core strengths on which the alliance has succeeded. OpenAI remains Microsoft’s frontier model partner, while Microsoft continues to hold exclusive IP rights and Azure API exclusivity until the announcement of Artificial General Intelligence (AGI).

This new era, however, signals an implicit transition, one where both companies can move forward on their own terms.

Enhanced structure, extended rights, and freedoms

The new deal encompasses several crucial elements to ensure balance between collaboration and independence.

A key new provision states that to certify OpenAI’s achievement of AGI, an independent expert panel must validate the declaration—adding an oversight layer to what could be one of the most defining moments in AI history.

Microsoft’s IP rights have also been extended through 2032, covering post-AGI models subject to agreed safety guardrails.

In parallel, Microsoft retains rights to “research IP” until either AGI is verified or until 2030 — whichever comes first.

These rights exclude model architectures, weights, inference and fine-tuning code, as well as data center hardware and software IP, with Microsoft retaining the non-research IP rights related to these elements.

The deal also creates opportunities for OpenAI to operate with greater independence. The company can now jointly develop products. API products will be hosted exclusively on Azure, while non-API products may run on any cloud service.

OpenAI can also publish open-weight models that meet safety and capability requirements, and provide US national security organizations with API access on any cloud infrastructure.

Financially, OpenAI has contracted to purchase an incremental $250 billion in Azure services, while Microsoft’s former right of first refusal to provide compute resources has been revoked.

Both companies will continue their revenue-sharing agreement until AGI is verified, though payment terms will now be extended over a longer period.

Through the optimization of terms, Microsoft and OpenAI are paving the way for a more agile and resilient partnership capable of adapting to evolving AI technologies.

While Microsoft can now pursue AGI independently or in collaboration with other companies, any use of OpenAI’s IP before AGI is declared will remain subject to defined computational constraints.

As both companies gear up to enter an AI-driven future, the rebalanced deal is expected to strike the right equilibrium between collaborative innovation and autonomy, marking the beginning of a new era in strategic alliances across the field of artificial intelligence.