Microsoft executives told court they feared becoming overly dependent on OpenAI

Microsoft executives used testimony in Elon Musk’s lawsuit against OpenAI to describe a problem that large corporate partners rarely spell out in public: dependence. In federal court in Oakland this week, Microsoft CEO Satya Nadella said the company worried that its deepening alliance with OpenAI could leave Microsoft too reliant on a single outside lab for critical artificial-intelligence technology. That concern surfaced as jurors heard evidence about Microsoft’s multibillion-dollar investments, its access to OpenAI intellectual property and its effort to protect itself if the partnership shifted. The testimony offered one of the clearest public accounts yet of how Microsoft viewed the risks inside its most important AI deal. ### What exactly did Microsoft say in court? Satya Nadella testified on May 11 that Microsoft was “outsourcing essentially a lot of the core IP development” and taking a “massive dependency on OpenAI,” according to courtroom coverage of the trial. Nadella said he wanted Microsoft to secure access to the intellectual property created through the partnership while continuing to build internal capabilities. (cnbc.com) An internal email discussed in court showed Nadella comparing the situation to Microsoft’s historic relationship with IBM. GeekWire reported that Nadella wrote in April 2022 that he did not want Microsoft to become “the next IBM” while OpenAI became “the next Microsoft,” a line that framed Microsoft’s concern as strategic dependence rather than a dispute over the existence of the partnership itself. (geekwire.com) ### Why was that fear showing up during a Musk lawsuit? Elon Musk’s 2024 lawsuit names Microsoft as a defendant and argues that the company aided and abetted OpenAI’s alleged breach of charitable trust after OpenAI moved away from the nonprofit structure Musk says he backed. Because Microsoft’s money, contracts and influence are central to that claim, the trial has pulled normally private deal terms and internal thinking into open court. (geekwire.com) CNBC reported on May 13 that Microsoft executives described seeking safeguards during early partnership negotiations as they weighed the risk of becoming overly reliant on OpenAI. That testimony matters in the case because it shows Microsoft recognized the concentration risk while negotiating commercial protections for itself. (cnbc.com) ### What protections was Microsoft trying to secure? Microsoft entered the partnership with OpenAI in 2019 and has invested more than $13 billion, according to the U.K. Competition and Markets Authority’s published summary of the relationship. The CMA said Microsoft’s rights include financial-investor protections and extensive commercial arrangements, but that Microsoft does not have a right to appoint a director to OpenAI’s board. (cnbc.com) Nadella testified that Microsoft saw a clear commercial element from the start. CNBC reported that he said Microsoft gave OpenAI discounted computing resources in the early years and expected commercial and marketing benefits in return, rather than treating the arrangement as a donation. That helps explain why access rights, exclusivity questions and ownership of resulting technology became central in negotiations. (assets.publishing.service.gov.uk) Reuters, citing The Information, separately reported this week that OpenAI and Microsoft agreed to cap the total revenue OpenAI shares with Microsoft at $38 billion. That report points to how the companies have continued revising financial terms as the relationship has matured. (cnbc.com) ### What does this show about the Microsoft-OpenAI relationship? The CMA said the Microsoft-OpenAI relationship is “complex” and has evolved as Microsoft increased its investment and the scope of collaboration expanded. The regulator concluded earlier this year that the partnership, in its current form, did not amount to a relevant merger situation or give Microsoft de facto control over OpenAI. (msn.com) Court testimony this week added a different level of detail. Nadella described the investment decision as a “one-way door,” according to GeekWire, because Microsoft could not simultaneously build separate supercomputing tracks for itself and for OpenAI without major tradeoffs. That description underscored the scale of Microsoft’s resource commitment and why executives were focused on preserving optionality. (assets.publishing.service.gov.uk) ### What happens next in the case? Thursday, May 14, was set for closing arguments in the federal trial in Oakland, according to local and national courtroom coverage. Courthouse News reported that jury deliberations are set to begin on May 18 after testimony concluded on May 13. (nbcbayarea.com) (geekwire.com)