The AI-Talent Poaching Frenzy Is Challenging Some VCs' Returns

This is, without question, the definition of a First World Problem. Some venture capital firms and their investors are not getting the returns they anticipated from AI startup portfolio companies as founders and/or key employees leave early (and quickly) for life-time wealth creating offers from big tech firms like OpenAI, Meta and others. 

Not that this means the VCs and investors are getting no return, just that the return is not what they anticipated if the top talent is poached and the investment has to be sold at less than optimal levels. So they have to downgrade from a Lamborghini to a mere Ferrari? OK, enough cynicism. Many believe it's only fair that the talent are getting paid. The real issue is the growing concentration in AI, already high to begin with, means that most of the returns - and control - is going to a few huge corporations, which might stifle innovation and better solutions. This may not seem like a big problem now, but with AI's increasing ubiquity, it could well affect many aspects of tech and human society sooner than most people realize. JL

Yuliya Chernova reports in the Wall Street Journal:

Promising startups are being defanged before they reach full potential. Some venture firms are subject to early wind-downs as a result of AI recruitment. Investors are nervous as founders of portfolio companies become targets of stupendous offers in the competition for talent in AI. High initial valuations for AI companies are exacerbating the risks for investors and misaligning incentives with founders. Even if shareholders make money, it (can be) less than they bargained for, or the exit comes earlier than expected. “You thought it was going to be a huge winner, (but) you need a high multiple to have a great fund." A quick exit at a lower amount removes the fund’s top asset. [Acquirers] are taking the upside off the table by making it extraordinarily financially compelling for founders.”

Big checks. Fast paydays. What’s not to love for investors in companies whose founders are recruitment targets in the AI talent game? Turns out many investors are wary.

Investors are getting nervous that founders of their portfolio companies could become targets of stupendous recruitment offers in the intense competition for talent in artificial intelligence. Even if shareholders end up making money, it is sometimes less than they bargained for, or the exit comes earlier than expected. 

In short, the deals might be more lucrative for the founders than the shareholders. Some of the most promising startups, some investors fear, are being defanged before they reach what investors believe is their full potential.

“It’s concerning,” said Salil Deshpande, general partner and founder of venture firm Uncorrelated Ventures. “We funded the guy’s dream and somebody else convinced him to give up on that dream,” he said about the scenarios unfolding in Silicon Valley.

“We invest in people at the early stages, we expect them to create value in the company we are investing in,” he said. “If someone will pay them enough to abandon what they are working on, that’s a risk. It’s a big risk. I’m not happy about that.”

On the bright side, some recent deals that accompanied big hires allowed shareholders to generate strong returns.

 

Earlier this month, Google hired the founders of coding startup Windsurf, as well as part of its team, and cashed out the startup’s shareholders for a total of $2.4 billion. A venture firm that first backed Windsurf in its seed round, for example, returned about 50 times invested capital, according to a person familiar with the situation. Another VC that came in at the Series B into Windsurf got a return of four times investment, according to another person. 

Even some venture firms are subject to early wind-downs as a result of AI recruitment drives. When Meta Platforms hired Daniel Gross and Nat Friedman, co-founders of venture firm NFDG, to join the company’s Superintelligence Labs, they decided that their venture firm wouldn’t raise additional funds. Gross also stepped down from his role as CEO of venture-backed startup Safe Superintelligence, which had been valued at $30 billion in a funding round earlier this year. Limited partners in NFDG were offered cash from the first fund at a markup.

When institutions back new venture firms, however, they hope to stick with the same managers, if they are successful, through many funds, not just one. That is because venture investors who create high returns for LPs are very rare, and if you find them, you don’t want them to give up investing.

Recent tech hires have generated early and quick returns for both LPs and VCs, creating high internal rates of return in some cases, a metric that measures returns over the investment’s life. IRR is higher if the exit happens faster. Yet, that is not necessarily best for investors in the long run. 

“Maybe you thought it was going to be a huge winner, and while IRR feels good, the reality is you need a high multiple to have a great fund,” said Matt Murphy, partner at Menlo Ventures. A quick exit at a lower absolute dollar amount, that is, potentially removes the fund’s top asset.

Deshpande said there is, of course, an upside to an early successful exit—it is cash now versus uncertainty. The same startup could have become an even bigger win in the future, or it could fail. 

High initial valuations for some of these AI companies are exacerbating the risks for investors and misaligning incentives with founders in acquisitions. 

In a hypothetical scenario, say a startup raises $1 billion at a $10 billion valuation and then gets acquired for $10 billion. Investors would get just their money back—not an appealing scenario—while the founders and the rest of the team would get $9 billion, making it a huge win. This math would have been true in the past, but today the rate of multibillion-dollar early valuations in AI, as well as big acquisition offers for top talent, is more frequent.

Investors say that large compensation offers to founders, such as via retention grants, can also make the deals more enticing for founders than for their shareholders.  

Murphy said the market dynamics around AI talent are making him worried. 

“As an investor, I don’t feel good about it at all,” he said. “We are still early in the AI wave, and so the outcome is asymmetric between founders, employees and investors. [The acquirers] are taking the upside off the table by making it extraordinarily financially compelling for founders,” he said.

Murphy said some of these deals are disappointing. “You are giving up on your company for a payday. That feels a bit soulless.”