The relationship between OpenAI and Microsoft, a long-standing investor and cloud partner, has become increasingly complex as the ChatGPT-maker transformed into a formidable competitor. However, newly revealed emails from 2018 indicate that Microsoft executives harbored significant reservations about providing additional funding to OpenAI when it was still a small nonprofit research lab.
These internal communications, presented in federal court during the recent Musk v. Altman trial, expose how Microsoft wavered on what is now celebrated as one of tech history’s most successful corporate partnerships. At the time, several Microsoft executives noted that their visits to OpenAI didn’t signal any imminent breakthroughs in artificial general intelligence (AGI).
Much of OpenAI’s 2017 work concentrated on developing AI systems to play video games, showing promising early results. Yet, to advance this project, OpenAI required five times more computing power than it had initially secured from Microsoft. This critical need set the stage for a pivotal internal debate within the Redmond giant.
Early Seeds of a Tech Giant (and Skepticism)
The initial connection between the two companies formed in 2016 when Microsoft agreed to provide $60 million worth of cloud computing services to OpenAI at a substantial discount, following a direct appeal from Elon Musk to Satya Nadella. OpenAI quickly consumed these services at twice the anticipated rate, signaling its aggressive computational needs.
The email trail began on August 11, 2017, with Nadella congratulating OpenAI CEO Sam Altman on an AI victory in a video game competition. Just ten days later, Altman responded with a bold request, seeking $300 million worth of Microsoft Azure cloud computing services and additional engineering support.
Altman expressed confidence in his vision, writing, “I think it will be the most impressive thing yet in the history of AI.” Nadella promptly sought input from four of his lieutenants, whose responses highlighted internal skepticism across Microsoft’s divisions.
According to Jason Zander, Microsoft’s executive vice president, the AI team saw “no value in engaging” with OpenAI, while its research team believed their own work was “more advanced.” The public relations teams were also wary, disliking the idea of supporting a group that promoted “machines beating humans.” Zander ultimately suggested that Azure could benefit from association with Musk and Altman, but cautioned against taking a “complete bath,” or large financial hit.
The Amazon Threat: A Crucial Turning Point
A subsequent financial analysis revealed that providing the requested services could cost Microsoft approximately $150 million over several years. Zander articulated the need for clear business value, stating, “Unless he can help us draw a more direct networking effect with OpenAI->Microsoft business value, we will wind up having to pass.” This strong stance underscored Microsoft’s desire for a tangible return on investment, not just a philanthropic endeavor.
The conversation paused for several months but reignited on January 10, 2018. Brett Tanzer, then a director on the Azure cloud unit, informed Nadella that Altman had proposed licensing OpenAI’s gaming AI to Microsoft’s Xbox division for $35-50 million in Azure credits. However, Xbox couldn’t commit to such a sum, and Microsoft planned to discontinue discounts for OpenAI by March.
Nadella then forwarded the email to 15 Microsoft executives, seeking their perspectives and sharing his own candid thoughts. He admitted, “Overall I can’t tell what research they are doing and how if shared with us it could help us get ahead,” acknowledging that Elon Musk was championing OpenAI as being on the “verge of some big AGI breakthroughs.”
Microsoft CTO Kevin Scott was “highly skeptical of an imminent breakthrough in AGI,” viewing OpenAI as merely treating Azure as “a bucket of undifferentiated GPUs.” He emphasized the lack of technical differentiation in their request but highlighted a significant PR risk.
Scott articulated the executives’ primary fear: “having them storm off to Amazon in a huff and shit-talk us and Azure on the way out.” Zander echoed this concern, stating his “worst case scenario is having them ditch Azure for AWS… and then land with some big new innovation that is shared with our competitors.” This worry proved prescient, as OpenAI later committed to spending $138 billion on Amazon cloud services, with Amazon investing $15 billion to $50 billion in OpenAI.
Microsoft’s chief scientific officer, Eric Horvitz, also expressed reservations about significant financial support without clear “broader ecosystem or relationship benefits.” Ultimately, Microsoft informed Altman that no internal team would commit to sponsoring OpenAI at that time.
From Hesitation to Billions: A Partnership Rekindled
Despite these early reservations, the relationship between Microsoft and OpenAI eventually rekindled and flourished. From 2019 through 2023, Microsoft became OpenAI’s most prolific fiscal sponsor, committing an astounding $13 billion in cash and cloud computing credits.
This dramatic shift from skepticism to monumental investment is central to Elon Musk’s ongoing lawsuit against OpenAI. Musk alleges that Microsoft aided and abetted the unauthorized use of his donations to the AI lab, contending that Microsoft helped transform OpenAI into a for-profit entity that corrupted its original nonprofit principles.
While Microsoft’s role in the trial has mostly been a background element, except for a few text messages and these newly revealed emails, that is set to change. Satya Nadella himself is scheduled to take the witness stand, promising further insights into this complex and defining partnership.
Source: Wired – AI
