Selective Acquihiring

It was around 5 pm on a lazy weekday afternoon. I still remember peering out into the mostly empty parking lot, the bright Californian sun blunted with tinted glass in my tiny conference room, on the second floor of an office building that headquartered Sun Microsystems at one point.

Of course, this was the Googleplex. It must have been in the early springtime, but I was staying in the office later than most of my colleagues to interview for a role at Meebo, around 2011. Meebo was a universal chat client, built and run entirely in the browser, which was cool enough at the time that I wanted to join the team as a senior front-end engineer1. I didn’t get the job, and I eventually ended up joining another startup by the name of Square instead, but that turned out to be a prescient move, as Meebo ended up getting acquired by Google only a year later.

Now, I doubt most people remember or care about Meebo a decade and a half later. Still, I distinctly remember the acquisition not only because I had a chance to join the company and then boomerang back to Google, but it was one of the first acquisitions that Google made where they just brought on the entire team. That is, it was a full acquihire, which was one of the first ones that Google ran due diligence at the team level as opposed to picking and choosing individuals from the acquired company. For Meebo, this meant absorbing 200+ employees, which wasn’t a big deal; after all, they had purchased Motorola Mobility beforehand and taken on 19,000 employees with that addition2.

I mention this slice of tech history because of this recent Stratechery article on the trends we’re seeing in today’s tech M&A landscape. Ben Thompson has been critical of the FTC cracking down on tech acquisitions and acquihires—in the name of antitrust oversight—because it kills the ecosystem that has allowed startups to compete with Big Tech for talent:

Big Tech, starting with Microsoft, realized that the easiest way to avoid regulatory annoyance around acquihires was to separate the acquisition from the hiring; what has happened as these deals have evolved is that tech companies increasingly realize that if they are simply hiring and not acquiring then they don’t have to hire everyone. That, however, breaks the implicit social contract that made startup employment significantly less risky for rank-and-file employees.

Ben predicts that this new scrutiny and weariness around antitrust will push M&A towards this new norm of selective personnel acquisition; it has already played out across several AI startups. Founders, executives, and key employees decamp for a tech giant—earning massive exits from their equity and control of the company—while other employees are left with the hollowed-out reminder of the firm, trying to keep the business afloat, knowing full well that their best talent is now working for a competitor who has shown a willingness to deploy massive amounts of capital to win, capital that the startup has no hope of matching. It’s no surprise, then, that most of these companies, left a shell of their former selves, don’t last very long3.

Then again, Silicon Valley is run on cycles, and the outsized reaction right now is very much a product of peak AI hype. There’s only one category of startup that’s getting this type of convoluted buyout—up-and-coming AI shops—and even then, only a handful of key researchers and executives are making the news, due to the compensation packages offered. Other reasons for M&A transactions remain valid: bolstering existing product lines, filling gaps in the business with proven, close-knit teams, or hiring talent en masse based on geography/skillset/specialization. They’re less flashy than a founder getting bought out for some $x billion by tech companies worth trillions, but of course these acquisitions still happen with regularity. For instance, the mentorship startup that I’ve volunteered for the past few years, Plato, was acquihired last year by Coda—which was then itself acquihired by Grammarly.

So chances are, in a few years, this period of exuberance will fade as the demand for AI talent either subsides with the maturation of foundational models, and/or with the gradual increase in the supply of researchers and engineers who can train said models. I graduated from college when web development was the premier technology of my day4; I saw the rise and maturation of mobile engineering5 and DS/ML follow similar ebbs and flows. I’d like to believe that startups are a resilient bunch, and that even if some of these dynamics around talent concentration prove to be sticky, there’ll be natural second-order effects and responses that would restore balance to the tech ecosystem.