AI Licensing in M&A: How Big Tech Buys Talent and Tech Without Buying the Company
Published by AINave Editorial • Reviewed by Ramit
AI dealmaking is moving away from traditional equity acquisitions. Instead of buying startups outright, big tech companies are increasingly licensing technology and hiring key talent in separate but linked transactions. This shift changes how deals are valued, how risk is managed, and how regulators may respond.
AI dealmaking shifts from equity to licensing and talent
According to Peter Werner, chair of Cooley LLP’s global business department, fast-moving AI deals are forcing buyers to make decisions without months of due diligence. “You’re back to unmoored financial calculations driving acquisition size,” Werner said in an interview. Buyers now rely on the perceived quality of the talent and the technology rather than traditional financial metrics. Earnouts are used to tie payments to talent retention and technology integration goals, reducing risk for buyers.
A surge in AI licensing agreements, often coupled with talent transfers, is reshaping the landscape. Rather than buying startup equity, tech giants pay fees for intellectual property licenses. Nvidia licensed Gorq’s inference-chip technology, Google paid about $2.4 billion for a non-exclusive license to Windsurf’s technology, and Google DeepMind entered a licensing agreement with Contextual AI. Each of those transactions also included hires of key leadership and technical talent.
What this means for AI builders and founders
If you are building an AI startup or negotiating a deal, this trend changes the signals you should watch. The minimum viable acquisition of assets is now a real option: a buyer may license your technology and hire your team without taking on your cap table, liabilities, or regulatory baggage. That can mean faster liquidity for founders and earlier access to resources for builders, but it also means your valuation may depend more on your team’s reputation and your tech’s integration potential than on revenue multiples.
For buyers, the compressed timeline means you cannot rely on deep technical diligence. Werner noted that buyers often lack the “sufficient horses to do deep technology diligence,” so deals become big bets. Earnouts help, but they require clear milestones for talent retention and tech integration.
The regulatory asterisk
Licensing structures can skirt mandatory merger reviews, but that is not a guarantee. Werner warned against over-relying on this: “If an acquirer is doing a truly transformational, significant transaction that antitrust regulators believe raises legitimate concerns, I don’t think it is dispositive to say it’s a license and not an acquisition.” Regulators are actively overseeing AI models, and the Trump administration’s lighter antitrust stance may not protect every deal. Builders should assume that any large licensing-plus-talent deal could still face scrutiny.
High prices and fast pace are the new normal
Werner said he stopped being surprised by the numbers long ago. “It’s more like resignation that this is just how things work now.” Cooley’s own deal volume reflects the pace: the firm ranked 10th among legal advisors by M&A deal volume in the first half of the year, with work for Nvidia, Anthropic, and OpenAI. It also guided Menlo Ventures on a $3 billion raise and represented General Catalyst in Anthropic’s $30 billion Series G. For builders, this means the window to negotiate favorable terms is short, and the bar for what counts as a “big bet” keeps rising.
