Microsoft CTO Kevin Scott took the stand Wednesday in federal court in Oakland, defending a 2018 email that questioned OpenAI’s commercial viability as standard diligence, not a sign that the tech giant was steering the AI startup away from its nonprofit mission. The testimony, part of a high-stakes legal battle over the Microsoft–OpenAI partnership, pulled back the curtain on early tensions between OpenAI’s idealistic charter and the commercial realities that came with billions in Microsoft funding.

Scott’s email, sent to Microsoft CEO Satya Nadella and other executives, warned that OpenAI was “not a great platform for building a business” and raised doubts about whether the startup could compete with Google’s AI prowess. In court, Scott characterized the message as a routine competitive risk assessment, not an attempt to undermine OpenAI’s structure or redirect its research. “It was my job to look around corners,” Scott said, according to reports from the courtroom.

A $13 Billion Bet with Guardrails

The email surfaced as part of a broader inquiry into whether Microsoft exerted undue control over OpenAI, potentially violating the startup’s original commitment to develop artificial general intelligence (AGI) for the benefit of humanity, not shareholders. Microsoft’s investment, which began with $1 billion in 2019 and later swelled to some $13 billion, came with exclusive cloud computing deals and, crucially, a cap on Microsoft’s share of OpenAI’s profits until its investment is repaid. But the partnership also gave Microsoft early access to GPT models and a seat at the table when OpenAI’s nonprofit board made governance decisions.

Critics, including Elon Musk, argue that this arrangement turned OpenAI into a de facto Microsoft subsidiary, prioritizing revenue over safety. Musk, a co-founder of OpenAI who left in 2018, has filed lawsuits seeking to unwind the partnership, claiming it betrays the startup’s founding agreement. Scott’s email is a key exhibit for those who say Microsoft was always focused on commercial returns, even as OpenAI publicly preached its nonprofit ethos.

The Email in Question

The 2018 email, sent months before Microsoft’s first investment, reflected a candid assessment of OpenAI’s prospects. Scott, who joined Microsoft in 2016 after leading LinkedIn’s engineering team, was a driving force behind the company’s AI pivot. He had previously warned that Microsoft was “multiple years behind” Google in machine learning infrastructure. In the message, he expressed skepticism that OpenAI’s research lab model could yield a viable business, but he also saw an opportunity: a partnership could give Microsoft a much-needed AI platform while allowing OpenAI to tap into Azure’s vast computing power.

Under cross-examination, plaintiffs’ lawyers pressed Scott on whether the email signaled a plan to “neutralize” OpenAI as a competitive threat. Scott pushed back, insisting that Microsoft never sought to constrain OpenAI’s independence. “We wanted them to succeed as a company because our fates were intertwined,” he testified.

Mission Creep or Pragmatic Evolution?

The courtroom clash distilled a fundamental tension: can a nonprofit dedicated to safe AGI accept billions from a for-profit behemoth without changing its DNA? OpenAI’s structure was unusual from the start. The nonprofit controlled a capped-profit subsidiary, which in turn attracted investors like Microsoft. But as the subsidiary’s valuation skyrocketed—reaching $90 billion in private tenders—the nonprofit’s board struggled to balance its oversight role with the subsidiary’s need for operational freedom.

That tension erupted last November when the nonprofit board briefly ousted CEO Sam Altman, only to reverse course after employee and investor backlash. The boardroom drama exposed rifts over how fast to commercialize AI, with Scott and Nadella reportedly supporting Altman’s push to ship products while some board members urged caution. Microsoft’s own governance, meanwhile, came under scrutiny: the company never took a board seat at OpenAI, but it holds significant influence through contractual rights and the threat of pulling Azure access.

The trial in Oakland is just one front. Musk’s lawsuit, filed in San Francisco, seeks to dissolve the partnership and reclaim his early donations. OpenAI has countersued, alleging Musk harassed the company and breached a founding agreement. Meanwhile, regulators in the U.S. and EU are examining the partnership for potential antitrust violations, probing whether Microsoft’s deep integration with OpenAI stifles competition in cloud and AI services.

For Windows users and enterprise IT leaders, the outcome could shape the availability and pricing of AI tools. Microsoft has woven OpenAI’s models into the fabric of its products—from GitHub Copilot to Azure OpenAI Service to the Copilot assistant in Windows 11. A forced unwinding of the deal could fragment that ecosystem, giving rivals like Google and Amazon room to gain ground. But a ruling that validates the current structure might greenlight even deeper coziness between Big Tech and AI startups, raising concerns about market concentration.

Scott’s Role as Architect and Advocate

As Microsoft’s CTO, Scott has been the architect of its AI strategy, overseeing the company’s massive investment in custom silicon, data centers, and developer tools. His testimony revealed a leader keenly aware of the competitive landscape. He acknowledged that Microsoft was initially hesitant about investing in a nonprofit but was won over by Altman’s vision and the potential to leapfrog Google.

Scott’s 2018 email, while blunt, also contained a strategic bet: that OpenAI’s research could become “the foundation of a very large business” if paired with Microsoft’s infrastructure. That bet paid off handsomely. Microsoft’s market cap has soared past $3 trillion, fueled by investor enthusiasm for AI. Yet the email’s appearance in court underscores the legal risks that come with such a high-profile alliance.

The Verdict on Trust

For the judge and jury, the core question is whether Microsoft and OpenAI colluded to subvert the nonprofit’s mission. Scott’s testimony painted a picture of a supportive partner, not a puppet master. But the email and other exhibits—including internal discussions about profit caps and governance changes—may tell a more complicated story.

If Microsoft loses, it could face limits on how it works with future AI partners and be forced to give up some of its contractual advantages. OpenAI, too, might have to restructure, potentially opening the door for more investors or even a public offering. But a win for the defendants would affirm that mission-oriented AI labs can take corporate cash without surrendering their souls.

Looking Ahead

As the trial continues, Windows watchers should keep an eye on two things: the immediate impact on Copilot features and the longer-term precedent for platform‑AI integration. Microsoft has already begun decoupling some OpenAI-specific services, offering alternative models like Meta’s Llama and Mistral in Azure, possibly as a hedge against legal headwinds. For now, the partnership remains intact, but Scott’s email and his testimony have reminded the industry that behind every transformative technology deal lies a messy, human debate over mission versus money.