Microsoft and OpenAI’s Early Partnership Negotiations Revealed
Court documents have surfaced detailing the initial discussions and negotiations between Microsoft and OpenAI, revealing the strategic considerations and anxieties that shaped their early partnership. These documents offer a fascinating glimpse into the high-stakes environment of AI development in 2017 and 2018, highlighting Microsoft’s cautious approach and OpenAI’s ambitious funding needs.
The genesis of the partnership can be traced back to the summer of 2017, shortly after OpenAI showcased an AI bot capable of defeating professional players in the game Dota 2. Following congratulations from Microsoft CEO Satya Nadella, OpenAI CEO Sam Altman proposed a significant expansion of their collaboration. OpenAI required substantial computing resources, estimated at around $300 million at list prices for Azure, to advance its research beyond existing Azure credits.
This initial request prompted concern among some Microsoft executives. Jason Zander, then Azure chief, expressed in an August 2017 email that such an investment would necessitate demonstrating “significant incremental revenue” directly attributable to the deal, exceeding $500 million, which could not be achieved more efficiently through other means.
Several months later, Altman presented an alternative proposal. According to Brett Tanzer, now VP of Azure solutions and ecosystem, this involved a partnership with Xbox focused on gaming, coupled with an open offer to share OpenAI’s technology and intellectual property in exchange for increased sponsorship of their Dota research. While the Xbox team showed interest in exploring collaboration, they were unable to shoulder the research costs independently.
Strategic Concerns and the Amazon Factor
Microsoft CTO Kevin Scott weighed in on the developing discussions in January 2018, sharing his thoughts with Nadella. Scott expressed uncertainty about the direct benefits Microsoft would gain from the deal, particularly regarding the Dota efforts. However, he voiced significant concern about the potential for OpenAI to align with Microsoft’s primary cloud competitor, Amazon Web Services (AWS).
In a January 2018 email, Scott articulated his apprehension: “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. They are building credibility in the AI community very fast, recruiting well, and are going to be an influential voice.” While acknowledging OpenAI’s growing influence, Scott questioned whether these factors alone justified the requested investment.
A year later, Scott reflected on his initial skepticism. In an email to Nadella and Bill Gates, he admitted to having been “highly dismissive” of AI advancements at OpenAI and Google DeepMind during a period focused on “game-playing stunts.” His perspective shifted as OpenAI pivoted towards natural language processing models, leading to concerns that Microsoft might fall behind Google in the AI race. Shortly after this reassessment, Microsoft announced its $1 billion investment in OpenAI.
The relationship has evolved significantly since these initial talks. Most recently, OpenAI renegotiated its agreement with Microsoft, allowing it to offer its AI models, including Codex, on AWS. This adjustment comes shortly after OpenAI informed its employees that the existing deal with Microsoft had “limited our ability to meet enterprises where they are — for many that’s [Amazon] Bedrock,” echoing the very concerns about shifting loyalties that Scott had foreseen years prior.
Business Style Takeaway: The internal Microsoft communications reveal a pragmatic, yet competitive, approach to strategic partnerships in emerging technologies. The fear of a key AI innovator aligning with a rival like Amazon underscores the importance of securing foundational relationships in the rapidly evolving tech landscape for enterprise success.
Learn more at : www.theverge.com