The Merger That Wasn’t: the Acquihire Playbook and the Antitrust Reckoning

Between March 2024 and mid-2026, Google, Microsoft, Amazon, Meta and Nvidia deployed more than $40 billion through a dealmaking structure that has no clean name in the legal lexicon. They did not acquire companies. They did not file Hart-Scott-Rodino notifications. They hired away founding teams, licensed intellectual property at extraordinary prices and left behind hollowed-out startups, stranded investors and a regulatory establishment scrambling to build a framework that fits the facts.

The transactions go by several names: acquihires, reverse acquihires, quasi-mergers, pseudo-acquisitions. Whatever you call them, they raise a question that outside counsel advising startup founders, venture investors and strategic partners must now answer with care: when does a talent deal become a merger the law requires you to report?

The Playbook

The template was perfected by Microsoft in March 2024. Microsoft paid $620 million to license Inflection AI’smodel technology, plus a separate $30 million for Inflection’s agreement not to sue over the talent poaching, then hired nearly the entire 70-person workforce, including CEO Mustafa Suleyman, who now leads Microsoft’s AI division. Inflection’s investors received a modest return on a company last valued at $4 billion. The company was not acquired. No HSR filing was made.

The copies came fast. Amazon hired roughly two-thirds of Adept AI’s team and paid a $25 million licensing fee in June 2024. That August, Google paid $2.7 billion for a non-exclusive license to Character AI’s models and brought back founders Noam Shazeer and Daniel De Freitas. The Windsurf deal in mid-2025 illustrated just how aggressive the playbook had become: OpenAI had agreed to acquire Windsurf for approximately $3 billion but could not close due to conflicts with its largest investor, Microsoft. The day after OpenAI’s exclusivity period expired, Google announced it would hire Windsurf’s CEO and key R&D staff and license certain Windsurf technologies, a package valued at roughly $2.4 billion, without filing for merger review. By 2025, Meta had acquired a large minority stake in Scale AI while hiring its CEO to lead a new internal research lab, and Nvidia had structured a similar deal with Enfabrica.

The economic logic is straightforward. Traditional acquisitions trigger HSR pre-merger notification when the deal value clears the current threshold of approximately $120 million. A license-and-hire structure, by contrast, closes in weeks, generates no mandatory notification and avoids the months of agency review that are toxic in a market where model capabilities evolve on a quarterly basis.

The Regulatory Response

The FTC and DOJ noticed. The FTC launched an investigation into the Microsoft-Inflection deal in mid-2024, followed by a probe into the Amazon-Adept transaction. The DOJ opened a formal investigation into the Google-Character AI deal in 2025. The UK’s Competition and Markets Authority opened a parallel inquiry into Microsoft-Inflection.

As of mid-2026, no enforcement action has been taken. But the agency signaling has intensified. In January 2026, FTC Chair Andrew Ferguson announced that the agency would scrutinize Big Tech acquihires more closely. FTC Commissioner Mark Meador went further in a Silicon Valley speech, citing a December 2025 academic report and warning that acquihires could serve as a “buy and kill, but for ultra-skilled labor” strategy, with firms acquiring talent not to deploy it productively but to deny it to rivals. Senators Warren, Wyden and Blumenthal wrote to the FTC and DOJ in February 2026 urging scrutiny of transactions they characterized as “de facto mergers” designed to consolidate talent, information and resources while bypassing standard review.

The agencies are not working from settled rules. Chair Ferguson has emphasized there is no one-size-fits-all standard for when a talent deal crosses into reportable territory and has acknowledged that the FTC may need to develop additional guidance specifically for acquihire structures. Whether the path forward takes the form of formal rulemaking, enforcement guidance or case-by-case review remains unresolved.

The Antitrust Question Counsel Must Confront

The legal analysis is genuinely unsettled, and that uncertainty is the practitioner’s problem.

In favor of HSR reportability, the argument runs as follows: a perpetual license to core model technology, combined with the acquisition of the entire founding team and primary engineering staff, is economically indistinguishable from acquiring the company. The startup that remains after the deal is a shell. Its competitive capacity has been transferred, in substance if not in form, to the acquirer. Under Section 7 of the Clayton Act, the question is whether the effect of the acquisition may be substantially to lessen competition, and economic substance, not deal structure, drives that inquiry. Former DOJ antitrust head Jonathan Kanter has argued publicly that acquihires can neutralize competition as effectively as traditional mergers and that regulators must evaluate substance, not form.

In favor of the current practice, the argument runs as follows: the HSR Act triggers on acquisitions of voting securities, non-corporate interests or assets, and a license is not an acquisition of assets in the statutory sense. Employees are free agents who accepted employment offers. No stock changed hands. Applying the statute through enforcement rather than rulemaking to transaction types it was not designed to cover raises fair notice concerns, particularly under a Trump administration that has signaled a pro-innovation, deregulatory posture toward AI markets.

Neither argument is frivolous. The tension between the agencies’ stated intent to scrutinize these deals and the administration’s broader deregulatory orientation makes the outcome genuinely uncertain. That uncertainty is precisely why careful analysis matters.

Implications for Startups and Counsel

Several practical considerations follow for founders, venture investors and the counsel advising them.

The regulatory environment has materially shifted. The fact that the Microsoft-Inflection deal closed in 2024 without challenge does not mean transactions being structured today carry the same risk profile. The agencies are building dossiers, developing frameworks and signaling intent across a bipartisan spectrum.

Investors who remain behind after a license-and-hire carry different exposure than the founders who join the acquirer. The hollowed-out entity still has fiduciary obligations and, potentially, employment claims from workers who were not hired away. Outside counsel to the startup board must think carefully about duties owed to remaining stakeholders before the founding team departs.

Antitrust counsel should be in the room at the term sheet stage. The definitional question of whether a particular license-and-hire constitutes an acquisition of assets for HSR purposes is not one to answer informally or after closing. The breadth of IP rights transferred, the proportion of employees hired and the competitive significance of the startup in its market all bear on the analysis.

The venture market is absorbing the consequences. Seed-stage AI venture capital has reportedly dropped significantly year over year as Big Tech vertically integrates the features that application-layer startups offered directly into cloud ecosystems. A license-and-hire that returns 1.1x to investors on a company that raised at a $4 billion valuation is not an exit. It is a market signal.

The Outside Counsel Takeaway

The license-and-hire transaction represents the leading edge of a broader mismatch in AI competition law: statutory frameworks built for an economy of tangible assets applied to a competitive landscape where value is concentrated in people and models. The HSR Act was not designed for a world in which a 70-person team and a non-exclusive perpetual license are the primary objects of a billion-dollar deal.

That mismatch is not the client’s friend. It means the rules are unsettled, the agencies are active, enforcement guidance is forthcoming in some form and the risk of ex post challenge is real. Outside counsel to startup founders, venture funds and strategic acquirers in the AI space should treat license-and-hire structures not as regulatory arbitrage opportunities but as transactions requiring careful analysis of HSR reportability, Clayton Act implications, fiduciary obligations to remaining stakeholders and the accelerating probability that the regulatory framework will change before the deal is fully integrated.

The merger that wasn’t may still have to answer for itself.