In the summer of 2017, OpenAI wowed the world by having a bot beat a Dota 2 professional player. Microsoft CEO Satya Nadella sent congratulations, and OpenAI CEO Sam Altman replied with a proposal: how about a $300 million partnership to fund the next phase of AI research, mostly in Azure credits? Some Microsoft execs were spooked. Jason Zander, then Azure chief, wrote to Nadella in August 2017 that the numbers only made sense if the deal generated $500 million in incremental revenue - which seemed unlikely.
Altman later suggested an Xbox-themed collaboration, but the gaming division couldn't shoulder the research costs alone. Enter Microsoft CTO Kevin Scott, who in January 2018 emailed Nadella with a blunt assessment: he wasn't sure what Microsoft would get out of the deal, but he was definitely worried about the PR disaster of OpenAI bolting to Amazon. "I guess the other thing to think about here is the PR downside of us not funding them, and having them storm off to Amazon in a huff and shit-talk us and Azure on the way out," Scott wrote. He acknowledged OpenAI was "building credibility in the AI community very fast" and would be "an influential voice."
A year later, Scott admitted to Nadella and Bill Gates that he had been "highly dismissive" of OpenAI's game-playing stunts but became impressed when the startup shifted to natural language processing. A month after that email, Microsoft announced a $1 billion investment. Fast forward seven years, and the partnership has evolved into something that looks a lot like a situationship - OpenAI recently renegotiated its deal to bring models to AWS, just days after the kind of "shit-talk" Scott feared. OpenAI told employees its Microsoft deal "had also limited our ability to meet enterprises where they are - for many that's [Amazon] Bedrock."