Microsoft feared OpenAI would ditch Azure, badmouth it to Amazon

When OpenAI was testing AI-powered gaming bots against human pros, Microsoft CEO Satya Nadella and OpenAI CEO Sam Altman were quietly laying the groundwork for what would become one of the most consequential partnerships in tech. Court documents from the ongoing Musk v. Altman trial now reveal the inner anxieties of Microsoft’s leadership as they debated investing in the startup, including a specific fear that OpenAI could “storm off to Amazon” and “shit-talk” Microsoft’s Azure cloud platform.

In the summer of 2017, just days after OpenAI demonstrated a bot defeating a professional Dota 2 player, Altman responded to Nadella’s congratulatory email by proposing a far larger collaboration. He wanted Microsoft to fund OpenAI’s next phase of AI research. But behind the scenes, Microsoft’s top executives worried that without a substantial commitment, the startup might pivot to a rival cloud provider and publicly disparage Microsoft’s technology. The internal communications, now exposed in court, offer a rare window into the high-stakes calculus that shaped the partnership.

These documents underscore how Microsoft’s leadership viewed OpenAI not just as a promising research lab, but as a potential competitive threat if they failed to lock in its loyalty. The fear that OpenAI could abandon Azure for Amazon Web Services and criticize Microsoft publicly drove much of the urgency behind the eventual multibillion-dollar deal. Today, that partnership has reshaped the AI landscape, but the early worries about defection and reputation damage were very real.