Internal Skepticism Ignited
The recent release of emails, stemming from the legal dispute between Elon Musk and OpenAI, provides an unprecedented window into the formative period
of Microsoft's engagement with the AI research firm. While Microsoft is widely credited for its prescient investment in OpenAI before generative AI, like ChatGPT, captured public imagination and initiated a global AI race, internal discussions reveal a complex picture. Key executives, including CEO Satya Nadella, grappled with the extent of their conviction in the fledgling company. Questions swirled about the true confidence held by Microsoft's leadership, the hidden doubts that surfaced internally, and whether the potential loss of ground to cloud rivals, specifically Amazon, was the decisive factor in committing resources. These insights were largely obscured until Musk's lawsuit, which accuses OpenAI of deviating from its original nonprofit mission of developing artificial general intelligence (AGI) for humanity's benefit, brought these private communications to light. Musk's legal action, seeking $150 billion in damages from both OpenAI and Microsoft, has unsealed sensitive exchanges that illuminate the strategic considerations behind a partnership that has, in more recent times, seen significant renegotiation, including OpenAI's ability to strike deals with Amazon Web Services (AWS) and Google Cloud.
The Amazon Threat
A significant catalyst for Microsoft's continued and deepening commitment to OpenAI, despite initial reservations, appears to have been the strategic concern over Amazon's potential to capture the AI startup. Emails from early 2018 show Microsoft's CTO, Kevin Scott, expressing apprehension about the public relations fallout if Microsoft declined to fund OpenAI, leading to the company "storm[ing] off to Amazon in a huff" and subsequently disparaging Microsoft and its Azure cloud services. Scott recognized OpenAI's rapidly growing credibility within the AI community and its potential to become an influential voice. He acknowledged that while he would prefer OpenAI to be a strong advocate for Microsoft and Azure, he was uncertain if that alone justified OpenAI's demands. Another executive, Eric Horvitz, echoed this sentiment, articulating a 'worst-case scenario' where OpenAI would abandon Azure for AWS, badmouth Microsoft, and then launch a major innovation shared with their competitors. These concerns highlight how the competitive landscape, specifically the rivalry with Amazon for cloud dominance, played a critical role in shaping Microsoft's strategic decisions regarding its investment in AI.
Early Research and Doubts
The initial seeds of Microsoft's involvement with OpenAI were sown even before the widespread recognition of advanced AI capabilities. In 2016, Microsoft provided OpenAI with a substantial discount on Azure computing resources for a 'Dota 2' project, where OpenAI paid $10 million for computing services valued at $60 million. However, by 2017, following a milestone achievement where OpenAI's AI bot defeated a professional Dota 2 player, Sam Altman, then OpenAI's CEO, communicated with Microsoft CEO Satya Nadella, signaling the need for significantly more investment to advance to the next phase of AI research, estimating costs around $300 million at Azure's list prices. This proposal, however, was met with considerable skepticism from various Microsoft executives. Nadella himself expressed uncertainty about the specific research OpenAI was conducting and how it would benefit Microsoft. Other executives, like Harry Shum, reported not seeing any immediate breakthroughs in AGI during a visit. Jason Zander, who was the Azure chief at the time, calculated that for such substantial investments to be financially viable, Microsoft would need to generate over $500 million in incremental revenue directly attributable to the deal, which could not be achieved more efficiently through other means. This internal debate underscores the significant hesitation and lack of clear business justification that characterized early discussions about deeper collaboration.
Shifting Perceptions and Investment
Despite earlier reservations, a palpable shift in Microsoft's perspective towards OpenAI's potential became evident over time, culminating in a substantial $1 billion investment in 2019. This change in stance was influenced by evolving assessments of AI's broader implications and competitive pressures. Kevin Scott, who had previously expressed significant doubts, later admitted in an email to Nadella and Bill Gates that he had been "highly dismissive" of the AI efforts at both OpenAI and Google DeepMind when they were primarily focused on impressive game-playing demonstrations. This self-reflection suggested a growing realization of the foundational importance of these research endeavors. The timing of this shift is particularly noteworthy; just a month after Scott's email acknowledging his prior dismissiveness, Microsoft officially announced its $1 billion investment in OpenAI. This substantial financial commitment marked a significant escalation of their partnership. Following the groundbreaking release of ChatGPT in November 2022, which catapulted generative AI into the global spotlight, Microsoft further solidified its relationship by investing billions of dollars more, demonstrating a clear strategy to capitalize on OpenAI's advancements.
