No Exit

The six most valuable companies in the world were once venture-backed startups. Alphabet, Amazon, Apple, Meta, Microsoft, and Nvidia, started out as small, private companies. They raised money from venture capitalists (VCs) to fuel their growth. They developed new technologies—search engines, online marketplaces, personal computers, social networks, operating systems, and graphics processing units. And then they did what most fast-growing private companies used to do—they went public.

Venture-backed startups used to have predictable lifecycles. A startup would raise a new round of capital every twelve to twenty-four months. After several rounds, the startup’s founders and employees would want to convert their shares to cash, and its VCs would need to deliver returns to their limited partners (LPs). The startup would exit the private market, so its shareholders could exit their investments. For some startups, the exit was an initial public offering (IPO). For others, the exit was an acquisition by a larger company.

The predictability of the startup lifecycle made the private-public divide coherent. When a startup grew enough to have a significant impact on society, it would usually become a public company or be sold to one.

Around the turn of the century, though, the startup lifecycle began to change, and the private-public divide began to blur. The number of IPOs fell precipitously. The share of startups exiting by acquisition rose. Many startups stayed private even as they grew into large companies. Some became “unicorns”—private companies valued over $1 billion.

No Exit, our new article forthcoming in the NYU Law Review, is about what happened next. Antitrust policy during the Biden administration reshaped the startup ecosystem. The antitrust laws give the government broad authority to sue to block mergers that could substantially lessen competition. Antitrust enforcers have long policed M&A deals among public companies. But they generally ignored acquisitions of startups. Startups rarely attain a market share that would raise concern under traditional merger analysis. So while acquirers had to notify antitrust enforcers about startup acquisitions over certain thresholds, the enforcers rarely sued to block these deals. When Facebook bought Instagram and Google bought DoubleClick, enforcers didn’t mount a challenge.

The era of permissive antitrust ended under President Biden. He appointed Lina Khan to chair the Federal Trade Commission (FTC) and Jonathan Kanter to lead the Antitrust Division of the Department of Justice (DOJ). Khan and Kanter stepped up enforcement, challenging more deals and settling fewer cases with consent decrees. They also made changes to the merger review process that increased the burden on merging parties, making it lengthier, more expensive, and more uncertain.

Khan and Kanter believed—arguably with good reason—that incumbent tech companies were buying startups to choke off nascent competitors. So they started to challenge them. Between 2012 and 2019, enforcers challenged only three startup acquisitions. Between 2020 and 2023, they challenged fourteen. Lawsuits or the credible threat of lawsuits killed deals between Adobe and Figma (design software), Sanofi and Maze (pharmaceuticals), Qualcomm and Autotalks (automotive), and Google and Wiz (cybersecurity). The chilling effect spread across Silicon Valley, putting would-be acquirers, founders, and VCs on notice that acquisitions by large incumbents were risky.

So how did startups respond? One might expect that they would substitute one kind of exit for another—that blocking acquisitions would lead to more IPOs. But while IPOs and acquisitions both provide liquidity, they are not perfect substitutes. The price that a startup commands in an acquisition may exceed the valuation it can achieve in an IPO because of synergies, economies of scale and scope, or the premium incumbents are willing to pay to eliminate potential competitors. The IPO market is also highly cyclical, so going public at the wrong time could be costly. And heightened antitrust scrutiny can reduce the value of an IPO by undermining one of its main advantages: access to publicly traded equity that can be used as currency for future acquisitions.

Instead of choosing a different exit, many startups are choosing no exit. In this Article, we show how startups and their would-be acquirers have developed new strategies to achieve their goals. The antitrust crackdown has increased the importance of two trends that pre-date it—the rise of employee tender offers and continuation funds. In an employee tender offer, a startup lets its employees cash out some or all of their shares while the startup remains private. Secondary sales of startup shares used to be small deals, and they were typically limited to founders or other key employees. But now rank-and-file startup employees can sell their shares, and the secondary market for startup equity has multiplied in value. Continuation funds let VCs stay invested in their portfolio companies longer than traditional venture funds allow. And as a result, startups face less liquidity pressure and can postpone exits.

Thwarted acquirers have also figured out strategies to evade merger enforcement. One of their strategies is a new structure that we call a centaur—a private company that is funded primarily by public company cash flows. The two largest centaurs are OpenAI and Anthropic, the companies behind ChatGPT and Claude. OpenAI has raised $13 billion from Microsoft, and Anthropic has raised $8 billion from Amazon and $3 billion from Google. Corporate VC investments by operating companies have been rising for the last two decades. But the centaur structure goes beyond traditional corporate VC by crowding out other investors and tying the fate of the startup and the public company with a deep commercial partnership. These deals allow Big Tech to access cutting-edge technology—and exercise influence over potential competitors—without an acquisition.

Another strategy is the reverse acquihire. In this kind of deal, a large tech company persuades the founders and key employees of a startup to quit en masse. Then it hires the former employees and makes a payment to the shell of the startup they left behind, which is ostensibly a fee to license the startup’s tech but is really a means to pay off its VCs. A reverse acquihire is an acquisition in substance but not form. And it’s becoming popular—Microsoft, Amazon, Alphabet, and Meta have all used it.

Our main claim in this Article is descriptive. We aim to show that heightened antitrust enforcement has transformed the private-public divide. We want securities regulation and antitrust scholars to appreciate how their seemingly autonomous fields of law are beginning to interact. But our arguments also carry implications for public policy.

We argue that antitrust enforcers should consider the hydraulic effects of their actions. Challenging a startup acquisition doesn’t just affect that deal. It also affects how founders and VCs perceive the exit options of other startups. Likewise, revisions to the merger review process that make acquisitions more expensive, time-consuming, and uncertain for merging parties also reduce the appeal of exiting by acquisition. And that in turn affects whether the next entrepreneurial engineer decides to found a startup and whether a VC decides to invest in it. The social benefits of enhanced antitrust enforcement can make these tradeoffs worthwhile, but we argue that there are real costs to narrowing the merger aperture, which enforcers ought to consider.

At the same time, we think that in some circumstances, enforcers need to be more aggressive. They need to prevent acquirers from making an end run around the antitrust laws. A reverse acquihire can harm competition just as much as a formal acquisition. And it’s possible that the centaur structure or other large corporate venture capital investments will be used to neuter competition. We argue that antitrust enforcers shouldn’t let companies circumvent merger enforcement structuring their transactions to avoid Hart-Scott-Rodino (HSR) filing requirements. HSR Rule 801.90 provides that the economic substance of a deal, rather than its form, triggers the requirement for merging parties to file with the government. The agencies can and should seek significant fines for these types of HSR violations.

It remains too early to tell how the antitrust crackdown—if it persists during the Trump administration—will affect the availability of startup finance. But we are not as worried as some other scholars about capital formation. VC deal value in 2023 and 2024 was down from 2021 and 2022 but is still near historical highs. VC fundraising has likewise fallen from its peak in 2022 but is still above pre-pandemic levels. The emergence of employee tender offers and continuation funds shows that startups and VCs are finding new ways to access capital and liquidity. Private ordering is resilient

We are worried, however, about transparency. In the old venture capital cycle, by the time a startup’s technology began to significantly impact society, it would have been exposed to the scrutiny of the public markets—disclosures, short sellers, and securities litigators. The startup would have an IPO and become subject to public company regulations. Or the startup would be acquired by a company already subject to public company regulations. Today, some of the most important firms developing artificial intelligence are being funded by public companies, but because the AI firms themselves are still private, they can keep the public in the dark.