Court documents related to a lawsuit filed by Elon Musk, Tesla CEO, against OpenAI CEO Sam Altman have disclosed previously unknown details about cooperation between OpenAI and Microsoft, drawing attention.

The Verge reported, citing the documents on May 8 local time, that Altman replied to a congratulatory email from Microsoft CEO Satya Nadella in the summer of 2017, shortly after OpenAI’s Dota 2 bot defeated a professional gamer, and proposed a bigger partnership.

Altman said OpenAI needed about $300 million in computing support based on the Azure cloud service. Jason Zander, then head of Microsoft Azure, reacted negatively, saying that for $300 million to be reasonable, the deal would have to generate more than $500 million in additional revenue directly.

Microsoft CTO Kevin Scott also said in an email to Nadella in January 2018 that he was not sure what benefits supporting OpenAI would bring to Microsoft.

But he was worried about the possibility that OpenAI could move to Amazon. He said, "If we do not support them, we have to consider the PR risk that OpenAI will go to Amazon and, on its way out, badmouth Microsoft and Azure," adding, "They are rapidly gaining trust in the AI community."

A year later, Scott acknowledged in an email to Nadella and Microsoft co-founder Bill Gates that they had "largely ignored" what OpenAI and Google DeepMind had done on AI. He also said he became concerned that Microsoft could fall behind Google AI as OpenAI shifted its focus to natural language processing models.

A month after Scott sent the email, Microsoft invested $1 billion in OpenAI.

Cooperation between Microsoft and OpenAI is not what it used to be. OpenAI revised an exclusive contract with Microsoft to provide its tools, including AI models and Codex, to AWS as well. OpenAI told employees last month that its contract with Microsoft "has limited our ability to reach customers on Amazon Bedrock, which many companies use." The Verge said Scott’s feared scenario of "badmouthing Microsoft" had become reality.