The tensions between OpenAI and Microsoft, long presented as strategic allies in artificial intelligence, are no longer limited to internal disagreements. They are escalating to the point where the California start-up is reportedly considering a high-profile legal response against its main investor.
Behind the usual communication and conciliatory statements of intent, negotiations are stalling. The reason, OpenAI’s desire to restructure itself as a for-profit company, in the form of a public benefit corporation. Such a change, according to the terms of the contract, requires Microsoft’s approval. However, this green light is slow in coming. And for good reason: the disagreement concerns the future stake Microsoft would hold in the new entity. OpenAI would be willing to cede up to 33% of the new legal vehicle to its historic partner, but no more. Microsoft, for its part, wants more.
From the outside, the alliance continues to produce powerful artificial intelligence tools, integrated into products like Copilot or hosted on the Azure infrastructure. In practice, the balance has become fragile.
OpenAI can no longer tolerate the constraints of exclusivity, nor the technical dependence on a single cloud provider, even though Microsoft has since lost exclusive hosting rights. The now-considered use of Google Cloud signals a change in direction.
But the most explosive element remains the option that some OpenAI executives would describe as “nuclear.” According to several sources close to the matter, the company has considered the possibility of filing an antitrust complaint against Microsoft with U.S. regulators. The argument would be based on the alleged abusive use of Microsoft’s dominant position in cloud services and on contractual clauses that would hinder competition.
Such a move could trigger a full federal review of the entire partnership. The content of these confidential discussions has leaked as time is running out. OpenAI must complete its legal conversion by the end of the year or risk losing up to $20 billion in potential funding. Yet, in a climate of regulatory uncertainty, legal obstacles are piling up. Elon Musk has filed a lawsuit to challenge the legality of OpenAI’s transformation, which he considers contrary to its original purpose.
Meta has also expressed its reservations, requesting the intervention of the California Attorney General. The split has not yet occurred, but the relationship is under severe strain. The Federal Trade Commission has also begun to take a closer look at this atypical partnership, whose hybrid structure raises questions of transparency and governance. The scale of Microsoft’s investment, estimated at $13 billion, has already drawn criticism in the sector, particularly from direct competitors like Google.
Behind the scenes, discussions are ongoing. Both companies insist they remain optimistic about continuing their collaboration until 2030, in accordance with the current contract terms. Statements that barely conceal a bitter negotiation, where the future of one of the most influential players in artificial intelligence could hang in the balance.