Home / Business and Economy / Microsoft Feared OpenAI Bolt to Amazon
Microsoft Feared OpenAI Bolt to Amazon
8 May
Summary
- Microsoft execs feared OpenAI joining Amazon for $300 million.
- Early concerns questioned the ROI for OpenAI's research.
- Microsoft invested $1 billion after initial skepticism.
Court documents from the Musk v. Altman trial offer a glimpse into Microsoft's early considerations of a partnership with OpenAI. In 2017, following OpenAI's success with a Dota 2 bot, CEO Sam Altman proposed a significant collaboration, requesting approximately $300 million in Azure compute resources.
Initial reactions from Microsoft executives, including Azure chief Jason Zander, raised concerns about the direct revenue generated by such a deal. Further discussions involved Xbox exploring gaming collaborations, but the research costs remained a significant hurdle.
Microsoft CTO Kevin Scott expressed apprehension in January 2018, questioning the direct benefits and fearing OpenAI might move to a competitor like Amazon. He noted the PR risk if OpenAI left "in a huff and shit-talk" Microsoft.
Scott later admitted to underestimating OpenAI and Google DeepMind's AI efforts, particularly shifting towards natural language processing. Within a month of his revised assessment, Microsoft announced a $1 billion investment in OpenAI.
Nearly seven years later, as of May 2026, OpenAI is reportedly renegotiating its deal with Microsoft to enable access to AWS, a move occurring shortly after concerns about OpenAI "shit-talking" Microsoft were realized. This new arrangement would allow OpenAI to meet enterprises on platforms like Amazon Bedrock.
