Where has all the M&A gone?
In the mythology of Silicon Valley, the initial public offering is the gold-standard exit for a start-up. But the number of I.P.O.s has dwindled over the last 25 years, replaced almost completely by acquisitions.
And now, even that exit has exited. In the last five years, the number of acquisitions has gone off a cliff.
Why the drop-off? In a new paper titled “No Exit,” a Vanderbilt law professor, Brian Broughman, and Cardozo law professors Matthew Wansley and Samuel Weinstein argue that the answer is fear of antitrust scrutiny — and the proof is in stealthy new deal-making, so-called reverse acquihires like Google’s recent hollowing-out of the A.I. company Windsurf, and Big Tech’s A.I. sugar-daddy relationships with companies like OpenAI. Wansley talked with DealBook about the deals that ate M&A.
The dot-com bust and regulatory changes ended the era of the I.P.O. So what happened to acquisitions?
In the 2000s and the 2010s, while I.P.O.s were in decline, acquisitions were on the rise — tracking the amount of money going into venture capital. The last few years, that starts to drop off, too. Our argument is that it’s out of fear that they’d get scrutiny from the Department of Justice and Federal Trade Commission.
In the years before 2020, there were, like, three start-up acquisitions that got scrutinized and challenged by the D.O.J. and F.T.C. After that, it’s 14.
How do you know Big Tech got scared? That could be a statistical blip.
The compelling evidence is deal structures that we haven’t seen before. To me, the reverse acquihire is the smoking gun, because there is no reason to do a reverse acquihire unless you’re trying to evade antitrust enforcement.
Acquihires are common enough. What makes it an antitrust dodge?
We think reverse acquihires are something new. A Big Tech company wants to acquire a start-up, but it’s worried that the D.O.J. or F.T.C. is going to challenge the deal. So instead, they get the founders and the core engineering team to leave the start-up and come work for them. But the V.C.s are not going to be happy about that, because they don’t want to see their investment walk out the door. So the Big Tech company says, OK, we’re going to pay the start-up to license their intellectual property.
That seems fine. People change jobs. IP gets licensed.
You have to understand, the license is fake. I mean, it’s real in the sense that it’s legally enforceable, but it is fake in the sense that they’re making the payment because they want to pay off the venture capitalists.
But the license fee doesn’t go to the V.C.s, does it?
The clearest evidence that this is what’s going on is that right after the license deal goes through, the start-up, or the shell of the start-up that’s left, pays a dividend to its shareholders. Dividends are common in big public corporations, but they are vanishingly rare in start-ups. The dividends are proof that what big tech is trying to do is pay off the V.C.s. It’s an acquisition in substance, but not form.
OK. Then what’s up with, say, Microsoft and OpenAI? That’s not a reverse acquihire.
That’s a more complex phenomenon. It used to be that corporate venture capital would be one of many investors in a company, including traditional V.C.s. What happened with OpenAI and Anthropic is different. The checks that Microsoft has written to OpenAI and that Google and Amazon are writing to Anthropic are in the billions. We call that a “centaur,” half public and half private. Public corporations are dominating the capital structure of a private company. So, if you’re OpenAI or Anthropic, you are much more beholden to these large public companies.
It is hard to think of an A.I. company that is truly independent from all of the incumbent big tech players.
And we’ll know you’re right about antitrust fears if the acquisitions start again under the new, more merger-friendly F.T.C. and D.O.J., right?
The Trump administration has been sending mixed signals on antitrust. There is some evidence that players in the market think that antitrust scrutiny is going to relax. Google is trying to do the Wiz deal again, right? But the persistence of the reverse acquihire structure suggests that they are still worried.
As much as I would love it if the real world would cleanly test our thesis, I don’t think the Trump administration is going to give us that opportunity.