Microsoft AI CEO Mustafa Suleyman has instructed his teams to prepare for a potential scenario where OpenAI restricts access to its intellectual property. The company is developing a contingency plan that focuses on two main strategies: the expansion of its own MAI model family and the pursuit of partnerships with external AI startups.

The internal MAI project includes both frontier models and specialized solutions. During the Ignite conference held on November 17-18, 2026, Microsoft introduced new additions to this lineup, specifically MAI Code 2 and MAI Cyber 2 Flash. The company positions these models as cost-effective alternatives that offer performance levels comparable to those of OpenAI and Anthropic.

Since 2019, Microsoft has invested over $13 billion into OpenAI. The current licensing agreement, last restructured in April 2026, remains in effect until 2032. While Azure continues to serve as the primary cloud partner for OpenAI, the developer now has the flexibility to provide services to customers across other cloud platforms.

This shift reflects Microsoft's broader multi-model strategy, which aims to reduce dependency on a single technology provider. By diversifying its offerings with proprietary models and startup collaborations, the company seeks to mitigate risks associated with potential changes in supplier terms, although no immediate termination of access by OpenAI has been reported.