Databricks. SpaceX. Stripe. That is what Amit Bhatti of TrueBridge sees when he looks at his firm's venture funds that are ten-plus years old: portfolios that started as spreads of small positions and, a decade on, have consolidated into a few names. In an interview TechCrunch published Friday, he says he hopes this generation of funds ends up holding “a lot of Anthropic and a lot of OpenAI.” His reasoning is the oldest rule in the business:
Venture has always been a power law business, so I actually view the concentration as a good thing. If we were getting to the maturity of one of our funds and there were still lots of tiny positions, something’s gone wrong.Amit Bhatti, TrueBridge, to TechCrunch
He's right about the power law and wrong about what it licenses. The old funds got concentrated by compounding: checks written early and cheap grew until they were most of the book. Concentration you grow into is an outcome. Concentration you buy is a position. What Bhatti and Beezer Clarkson of LGT Capital Partners describe in the rest of the interview is the second kind, and the two only look alike on a statement printed in year ten.
Concentration you grow into is an outcome. Concentration you buy is a position.
Bhatti supplies the mechanism himself: “If Anthropic is raising a giant quantum of capital and you put 20% of the fund into it, you’re obviously going to be ahead of pace.” That is how deployment that used to take 24 to 36 months now takes 12 to 18 for some funds, and why Clarkson is seeing managers come back to LPs every year. A fund like that doesn't wait for a winner to grow into a fifth of the portfolio. It starts there, at the round price. And because nearly everyone has piled into the same two companies, often several funds inside one firm, the LP who backs thirty managers for diversification holds the same two positions thirty times.
The power law pays on entry price, not on the name. Run Bhatti's 20% through OpenAI's new number: about $20 billion at a $1.25 trillion valuation. For a 20% position to return the fund on its own, the standard venture test, it has to 5x. At that price, 5x is a $6.25 trillion company. The name on the cap table is the one every seed fund wanted. The instrument is closer to a concentrated public-equity bet with worse liquidity.
The price is already moving. That $20 billion is $10 billion less than the round's target, with the valuation trimmed from $1.4 trillion and a first close expected below $20 billion after some investors reduced their allocations. The same day, the Nasdaq fell 2.9% and closed 7.8% below its October 6 high. None of that is a crisis. It is a reminder that what gets negotiated is the price, and the price is where the power law lives.
The best case for Bhatti is that the companies are real, and he makes it well. The last time pacing ran this hot, he says, managers were buying “zero-revenue crypto moonshots”; now companies are going from zero to $100 million in revenue. If these two labs are this decade's Stripe and SpaceX, the funds that own a lot of them will look smart and the ones that diversified away will look timid. Grant all of it. It answers a different question. Whether OpenAI and Anthropic win is a question about the companies. Whether a fund that put a fifth of itself into one of them at a trillion-dollar mark earns venture returns is a question about the check, and revenue growth doesn't merge the two. Clarkson, looking at the 2021 vintages, puts it more gently: “it’s typically harder to do a one-year vintage and get it right.”
Databricks, SpaceX and Stripe grew into their weight. This vintage is buying its weight on day one. Same names on the statement. Different bet.