Why buy the company when you can hire its best people? That’s the logic behind a series of reverse acquihires — as they’re known — by the big tech companies as the AI talent wars continue to rage.
Such moves often make sense both for the acquiring companies and the startup teams that jump ship. But it can leave the startup’s investors — and junior employers — in the lurch. And it also increases the risk of the AI landscape becoming a less vibrant big corporate monoculture.
In the UK this month, Sifted revealed that Anthropic had hired three cofounders and several senior employees from the London-based AI startup Humanloop, backed by Y Combinator, Index Ventures and Albion. In a similar deal in March, Microsoft poached the two cofounders and other senior staff from London-based AI video startup Haiper.
Europe is seen as a particularly attractive talent pool for frontier AI companies, given local pay is so much lower than on the West Coast. Cohere CEO Aidan Gomez has praised the region’s skills base, recently signing a deal with the UK government to cultivate homegrown AI talent.
As ever, this corporate fashion started in the US and is on a bigger scale. Last year, Microsoft hired Mustafa Suleyman, the former Google DeepMind co-founder, and most of the top team at Inflection AI. This June, Meta scooped up Alexandr Wang and his top researchers at Scale AI in return for a $14.8bn investment in the startup.
Buying a startup’s top talent rather than swallowing a company whole is clearly a smart form of regulatory arbitrage. Given the uncertain state of the antitrust landscape on both sides of the Atlantic, acquirers will naturally prefer cheaper and quicker workarounds. Adobe’s costly failed bid for Figma, when it was forced to pay out a $1bn break fee after regulators stymied the deal, highlighted the risks of trying to make outright acquisitions of startups.
From the startup perspective, it is understandable that founders might want to cash in early and join a bigger, and better-resourced, company. One founder I met recently, who had accepted a job at an established firm after working ferociously hard for years to build her own startup, told me that she had never slept better than the first night after joining her new company.
Trying to build a startup in any field, let alone one in as competitive and fast-changing an industry as AI, can be a nightmare. It’s certainly a lot less stressful to pocket a corporate pay cheque and stock options, enjoy the big company perks and reacquaint yourself with your family.
Whether it is such a good thing for the rest of the startup ecosystem is another matter, however. Junior employees at affected startups are likely to lose out with the value of the “remainco” most likely to plummet, assuming it survives at all.
It’s harder to generalise for VC investors. The critical determinants of whether they make money or not are how early they came on board and what the details of any exit are. Acquiring companies are often keen to keep investors whole, at the very least, to minimise the fuss.
However, Nathan Benaich, an active AI investor at Air Street Capital, says that outright acquisitions of startups remain important for the market because they create liquidity for all participants and keep the system going.
“Traditional acquihires in AI (i.e. buying the whole company, IP and team) was common during the first deep learning boom post-Imagenet and it helped bootstrap the AI efforts of all the major big tech AI players we know of today. In general, investors and teams made money,” he says.
Encouragingly, there is evidence to suggest that many budding AI entrepreneurs, and other early stage employees, remain keen to go down the startup route.
Indeed, according to Zeki Data’s 2025 State of AI Talent Report, which tracks the world’s top AI researchers, smaller companies with fewer than 50 employees have been increasing their share of the global talent pool in spite of the rockstar salaries being offered by the Big Tech companies. These startups accounted for 19% of the top AI talent, identified by Zeki, in 2024 compared with 12% in 2019.
It is vital for the health of the AI innovation economy that these startups scale and exit and return money to their founders and investors. Not everyone should go and work for Meta no matter how much money they slap on the table.
Read the orginal article: https://sifted.eu/articles/acquihire-reverse-ai-startup-europe-meta/
