Unsurprisingly, it appears that OpenAI’s relationship with Microsoft (NASDAQ:) is nearing exhaustion. According to the Wall Street Journal report on Monday, OpenAI is exploring a legal avenue out of its exclusive contract with Microsoft. Specifically, by seeking a federal regulatory review for potential antitrust law violations.
If such an antitrust complaint were to be deployed, it would serve as a nuclear option to terminate the Microsoft-OpenAI relationship.
However, a federal decision in that direction would be in contrast to the UK’s decision favoring Microsoft. In March 2025, the UK’s Competition and Markets Authority (CMA) decided that Microsoft’s partnership with OpenAI “does not qualify for investigation under the merger provisions of the Enterprise Act 2002.”
If this is the likely outcome of an antitrust investigation in the US as well, it may be the case that OpenAI is seeking to revise the terms of the agreement with Microsoft. Specifically, for Microsoft to allow OpenAI to transition into a public benefit corporation from its existing nonprofit status.
The Information previously reported that OpenAI is willing to give Microsoft a 33% stake in this new restructured entity, while foregoing rights to future profits. Additionally, Microsoft would no longer hold exclusive rights to OpenAI’s models on Azure.OpenAI’s evolving relationships with Microsoft and other companies suggest a growing drive for diversification and independence.
Evolving Entanglement: Microsoft, OpenAI, and Oracle
After OpenAI’s ChatGPT pushed the AI into the public spotlight, Microsoft was the first Big Tech company to go all in. But even before the AI hype ramped up, Microsoft invested $1 billion in OpenAI in 2019, followed by another round of funding in March 2021, and the largest commitment of $10 billion in January 2023.
In September 2024, Microsoft disclosed that total OpenAI investments accounted for $13 billion, with Microsoft getting 20% of OpenAI’s revenue. The arrangement has been mutually beneficial, as Microsoft offered ready-to-go Azure cloud infrastructure and seamless AI integration into a wide range of legacy software.
In November 2023, Microsoft CEO Satya Nadella noted that if OpenAI were to disappear, Microsoft would own all of the organization’s resources.
However, the exclusivity slowly started to shift out of Microsoft’s favor. This first became apparent in June 2024 when OpenAI inked a deal with Oracle (NYSE:) for its Oracle Cloud Infrastructure (OCI), while continuing the partnership with Microsoft.
In January 2025, the strategic relationship between Microsoft and OpenAI officially changed from the existing exclusive access to OpenAI’s IP through 2030. Under that arrangement, OpenAI API (application programming interface) is exclusive to Azure, as OpenAI rolls out its latest models.
Further, Microsoft’s cloud infrastructure would support all OpenAI products, including LLM training. By introducing the right of first refusal (ROFR), Microsoft enabled OpenAI to “build additional capacity, primarily for research and training of models.”
The January 2025 agreement opened the door for OpenAI’s compute capacity outside Microsoft. This expansion is likely within Oracle. Shortly after President Trump was inaugurated, Larry Ellison presented the Stargate Project at the White House, together with OpenAI CEO Sam Altman and SoftBank CEO Masayoshi Son.
From the entanglement with Oracle, Microsoft’s cloud infrastructure competitor, it was clear back then that OpenAI is not likely to settle for being under Microsoft’s umbrella and its Copilot AI suite. As recently as this month, OpenAI also made a deal with Alphabet’s Google Cloud Platform, although it directly competes with Google’s Gemini model.
At the same time, Microsoft is diversifying its AI stake as well.
Microsoft’s Road Ahead without OpenAI
While OpenAI sent a strong signal through WSJ that the relationship with Microsoft needs to change, Microsoft was not taken aback. In late May, the Big Tech giant opened up its Azure cloud infrastructure for Elon Musk’s xAI which is developing Grok models.
Microsoft further emphasized that it supports “an open, diverse AI ecosystem, rather than relying on a single model provider.” This is a clear play on the company’s strength to serve as an agnostic cloud provider.
Ending April for fiscal Q3 2025, Microsoft reported 21% year-over-year growth of its Intelligent Cloud division to $26.8 billion. Of the Big Three hyperscalers – Amazon (NASDAQ:) (AWS), Microsoft (Azure), Alphabet (NASDAQ:) (GCP) – Microsoft holds 22% global cloud market share. This is significantly above Google’s 12% and Oracle’s 3% but under Amazon’s 29% share as the dominant cloud provider.
In addition to Grok, in the new landscape of AI agents, Microsoft is also expanding Azure’s AI Foundry to Anthropic. Interestingly, both Google and Amazon poured billions into Anthropic, showing that Microsoft is willing to collaborate even with AI companies backed by its biggest cloud rivals in order to strengthen Azure’s position in the broader AI infrastructure race.
On top of this diversification, Microsoft has plenty of resources and human capital to launch its own model. Namely, the small language model (SML) Phi-4 with its 14 billion parameter range. In late 2024, Phi-4 gained top math scores against its LLM rivals.
Ultimately, Microsoft could end up integrating different AI models based on their most performant niches. But for end users, only robust results would matter, not the name of the AI model under the hood.
***
Neither the author, Tim Fries, nor this website, The Tokenist, provide financial advice. Please consult our website policy prior to making financial decisions.
This article was originally published on The Tokenist. Check out The Tokenist’s free newsletter, Five Minute Finance, for weekly analysis of the biggest trends in finance and technology.
