Satya Nadella was apparently worried that OpenAI might defect to Amazon Web Services and publicly trash Azure - that phrasing surfaces in court documents from the ongoing Elon Musk v. Sam Altman trial. The documents offer a rare look at the private anxieties behind one of tech's most consequential corporate partnerships.
The communications cover a period when OpenAI was experimenting with AI-powered gaming bots, long before ChatGPT existed. Nadella and Altman were negotiating what would become a multi-billion dollar relationship, and Microsoft's leadership was concerned enough about losing the deal to document their fears about Amazon specifically.
The worry had a rational basis. Amazon Web Services runs more cloud infrastructure than any other provider, and training large AI models requires enormous amounts of it. An OpenAI that ended up primarily on AWS - and publicly criticized Azure in the process - would have been a meaningful blow to Microsoft's cloud business and its credibility in the AI industry at exactly the moment that industry was becoming important.
How the Partnership Actually Formed
Microsoft has since committed more than $13 billion to OpenAI, and its Copilot products across Office, Windows, and developer tools are all built on that foundation. OpenAI's infrastructure runs substantially on Azure. The outcome Microsoft was fighting for is the one that materialized.
But these documents show it wasn't a foregone conclusion. OpenAI had real leverage - Amazon was a credible alternative - and Microsoft was negotiating from a position of competitive anxiety, not strength.
The trial centers on Elon Musk's claims that OpenAI abandoned its nonprofit mission, a separate dispute from the partnership mechanics. What the discovery process keeps producing, as a byproduct, is documentary evidence of how the current AI industry was actually assembled: carefully, with significant fear of getting it wrong.