Skip to content
Advertisement
Microsoft prepares teams for a possible loss of OpenAI access

Microsoft prepares teams for a possible loss of OpenAI access

Microsoft official logo (public domain, Wikimedia Commons) — CryptoBriefing brand treatment

Microsoft AI CEO Mustafa Suleyman has reportedly readied staff for a scenario where OpenAI cuts off its IP, with in-house MAI models and startup deals as the backup plan

Microsoft has put over $13 billion into OpenAI since 2019. It is now reportedly planning for a world where it can no longer use the technology that money helped fund.

According to a report from The Information dated November 19, 2026, Microsoft AI CEO Mustafa Suleiman has prepared his teams for a scenario in which OpenAI could potentially cut off access to its intellectual property. The contingency rests on two pillars: Microsoft building its own AI models and pursuing deals with startups.

What Microsoft is reportedly planning for

The Information’s report describes preparation, not a breakup. Suleyman’s teams have reportedly been readied for the possibility that OpenAI might pull access to its IP. In that case, Microsoft would be left without the OpenAI technology it currently licenses.

The backup plan runs on two tracks. The first is homegrown: Microsoft’s own family of models, branded MAI. The second looks outward, with Microsoft seeking startup deals.

The report surfaced one day after Microsoft wrapped its Ignite conference, held November 17-18, 2026. At that event, the company rolled out several new MAI models.

Advertisement

Two of the headliners were MAI Code 2 and MAI Cyber 2 Flash. Microsoft is pitching the new MAI lineup as cheaper options that deliver performance close to what OpenAI and Anthropic offer.

The MAI brand covers both frontier models, meaning the most capable general-purpose systems, and specialized models built for narrower jobs. The names of the Ignite releases hint at the specialized side, with one aimed at code and the other at cybersecurity.

A partnership that keeps getting rewritten

The Microsoft-OpenAI relationship has gone through several rounds of restructuring. The most recent overhaul came in April 2026.

That revision converted Microsoft’s license to OpenAI’s IP into a non-exclusive arrangement that runs through 2032. It also capped the revenue-share payments OpenAI sends to Microsoft, which continue through 2030.

The cloud side shifted as well. Azure is still OpenAI’s primary cloud partner. OpenAI, however, can now serve customers on any cloud platform.

What this means for Microsoft, OpenAI and AI buyers

For Microsoft, the core logic is risk management. Leaning on a single outside supplier for a core technology leaves a company exposed if that supplier changes the terms. Owning capable in-house models, and keeping startup options open, reduces that exposure.

The move also slots neatly into Microsoft’s broader “multi-model” strategy. Instead of betting everything on one provider, Microsoft wants to offer a menu that mixes its own MAI models with outside options.

Microsoft remains a major OpenAI backer and its primary cloud partner. It is also now marketing models positioned as lower-cost alternatives to OpenAI’s offerings.

There are real caveats. The report describes preparation for a scenario, not an action OpenAI has taken. Microsoft’s license still runs through 2032, and nothing in the reporting indicates OpenAI has moved to cut off access.

Performance claims also need testing outside the keynote stage. “Near parity” is a vendor’s description, and customers will judge MAI Code 2 and MAI Cyber 2 Flash by how they hold up in day-to-day work.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.