Until recently, the relationship between Microsoft and OpenAI was an example of how a large corporation and an ambitious startup could operate in perfect harmony. Now, however, it appears that the companies could end up in court. The alleged reason is a breach of the rules by OpenAI.
Back to the Beginning
When the startup OpenAI was getting started, Microsoft invested one billion dollars in it, and in 2023 it added another ten billion. As an investor, however, the company wanted a commitment that Azure would become the exclusive cloud provider for its AI models.
And OpenAI agreed. After all, why not? OpenAI would gain capital and infrastructure, while Microsoft would gain access to the world's most advanced AI models. But thanks to GPT, the situation soon began to change. OpenAI gradually left behind its position as Microsoft's protégé and grew into a huge company valued at hundreds of billions of dollars. And so it began making its own plans and looking for new partners.
A Third Party Joins In: Amazon and a $50 Billion Deal
The entire situation came to a head in February 2026, when OpenAI announced a deal with Amazon worth 50 billion dollars. OpenAI simply decided to run its new OpenAI Frontier service (an enterprise platform that allows companies to build and deploy teams of AI agents across their systems) on Amazon's servers.
And that is precisely what Microsoft took issue with. The agreement with Microsoft stipulates that access to OpenAI's models must be provided exclusively through Azure. According to Microsoft, using Amazon Web Services for Frontier directly violates this principle.
A Play on Words: Stateless vs. Stateful
OpenAI naturally refuses to admit that this constitutes a breach of the agreement. That is precisely why the dispute technically revolves around one small detail: the difference between the words “stateless” and “stateful.” And what does that mean in this context?
Although it may seem that your regular chatbot knows you, it operates as a stateless environment, meaning that it processes the entire conversation from the beginning each time. In contrast, Frontier is intended to operate as a “stateful” environment that remembers the context and previous work. According to OpenAI, this “stateful” model architecture does not fall under the agreement with Microsoft. Therefore, the company is not technically breaking the rules.
Microsoft disagrees, however. Even Amazon Web Services, where Frontier is supposed to run, has issued internal instructions to employees. They are to state that AWS only powers Frontier, not that it provides access to it. This shows that Amazon, too, is aware of how sensitive the entire situation is.
Is It Worth Going to Court?
According to the Financial Times, a source familiar with Microsoft's position said that the company knows very well what is in its contract. And if it is breached, it will file a lawsuit. Nevertheless, the situation is complicated. A legal dispute would be unpleasant for both sides.
Microsoft is facing regulatory investigations in the US, the UK, and the EU over its practices involving Azure, while OpenAI is preparing to go public and offer its shares to the public. A legal dispute, however, would cast the company in a bad light.
Whether the whole matter ends up in court or at the negotiating table, one thing is clear: the golden age of friendship between Microsoft and OpenAI is apparently gone for good.
Sources: The Reuters, Futurism, Outlook Business, UC Today
