Venture capital’s new edge: betting on founders, not ideas

By mid-2026, the founder-first turn was no longer implicit; leading investors were quantifying it. On The Full Ratchet, Precursor’s Charles Hudson described a clear reweighting in his own underwriting: “when I started Inventure I was mostly a markets person and now I've become mostly a people person,” adding that “the slider is like 70% people, 30% market.” He also showed how that philosophy changed capital deployment, saying, “we've started doing more directional people bets with smaller dollars like here's 150 or 250k to go,” before committing more once an idea sharpened.

That repositioning was echoed across other top firms, showing a broader strategic shift away from classic market-size and traction screens. On The Full Ratchet, 645 Ventures’ Nandi Okike said the firm moved from being “primarily focused on financial traction and metrics as well as kind of known market size” to focusing “much more on people,” spending time to understand founders’ subjective qualities; and at Mayfield, Naveen Chadha distilled the lesson from missed seed bets bluntly: “So that's why when you ask me, people first, man, I've gone wrong on backing the 15 years back the right people because we didn't believe in their idea.”

Investors face too many possibilities to underwrite every early company through market maps alone. On Not Another CEO, one investor described tracking a universe of roughly 10,000 high potential software CEOs, saying there’s like 10,000 people that were in touch with, and that out of that pool is likely to come many of those 20 companies. He said network signals and network centricity help surface candidates, and that someone who ran all of engineering at Palunteer has probably been pitched by 900 VCs. In that setting, the scarce edge is not more data but sharper judgment about which person can navigate ambiguity.

That judgment matters because the earliest rounds are standardized while the company is not: north of 90% of companies will raise some sort of pre seed seed, and that first round is almost always a 20% dilution event and almost always up to 10 million bucks. Investors are betting on a founder they may work with for 10 to 15 year[s], often before funds clear, because the money's going to take, like, four weeks to clear, and if it's a safe note, it could be, like a week. Shardul Shah captured the speed of that trust when in 72 hours we had a term sheet signed because sometimes you just click with someone.

AI has compressed the shelf life of early startup advantages, making product and traction weaker signals than they once were. Charles Hudson said on Build Mode that in the pre-AI world the half life of a good idea felt longer because a head start was harder to catch, but we do not live in that world anymore. He contrasted that with a time when VCs would get very excited about something where the expectation, if it was successful, was that you would come in for a few million dollars on say a 20 million, a setup that assumed more durable early advantages.

That speed changes what investors trust: not the existence of an MVP or even sudden revenue, but the founder’s ability to build something that compounds beyond what others can quickly copy. Build Mode said founders now present ideas that are already taking off, with 0 to 1 million of ARR in like three months, yet investors probe for what is actually proprietary because anyone can build an MVP so quickly these days. TechCrunch Disrupt’s StrictlyVC also said that thanks largely to AI and the rate at which startups are scaling, venture capital is itself in a state of flux, underscoring why founder judgment matters more when early signals are noisier.

By mid-2026, founder-centric investing was no longer a private preference; it was being described in public as a deliberate response to how seed markets were splitting. On The Peel with Turner Novak, a seed manager said “we’ve always been sort of 70% first-time founder, 30% repeat founder,” adding that it is “totally fine” if founders are “a little raw or a little hard to underwrite” because the firm has time to work with them, while multi-stage funds get “5% better, 10% better every year” at chasing “the most obviously pedigreed people raising seed rounds,” a “vice” that “gets tighter and tighter.” They argued that “all of the great companies… are not going to come from the pool of repeat or well-known founders,” and warned LPs that if they believe “it’s 100% all the best companies are going to be repeat people…,” they are underwriting too narrow a pool.

That rhetoric matches broader evidence that venture underwriting is bifurcating toward concentrated early bets rather than converging on one standard model. The same Peel discussion cited Chameleon Ventures’ Nunu for a factor-style finding that “76% of all unicorns” did not have a fund over $100 million in their seed round, summarized as “70% 76% of unicorns, their seed round was led by a small fund,” while industry ranking data shows “the top 5% of VCs have generated roughly 90% of the industry’s profits,” “Twenty-one of the top 100 firms were founded in 2015 or later,” and “The top firm scores roughly 41 times the hundredth-ranked firm.”