Most VCs Are Afraid to Be Wrong | Rohan Pujara

Rohan Pujara sees San Francisco venture capital as a consensus-driven bubble, where the same companies and founders circulate through tier-one funds and emerging managers, often via group chats. He argues that LP-GP incentives, pressure for short-term results, larger funds and management fees can reward avoiding visible mistakes over making the best investments. For early-stage companies, a crowd of investors can mean competition from day one, capital burned against rivals and less chance to build durable margins or a moat, even as incumbents such as Meta, OpenAI and Anthropic enter the market. Consensus may be more useful at Series A and later, which he characterizes as growth investing; a large market alone does not make a category consensus—the breadth of investor agreement does.

Pujara is not avoiding AI, but wants distinctive companies on its fringes or ideas that may become consensus in a few months. He cites Cursor as a non-consensus early-stage bet that was not initially doing coding, and says Anthropic’s first round was difficult to raise. He sees consensus forming after investors watch a product work and pattern-match; competition followed the launch of Instinct’s phone agent. His standard for an investor is to seek truth, think clearly and resist FOMO. At seed, he is drawn to polarizing founders and tries to understand why they see the world differently, backing them when he believes they are capable and right. Valhalla looks for companies he would be excited to own in ten years, regardless of market conditions—not investments premised on predicting a valuation jump from $2 billion to $10 billion.

He values founders’ courage, trustworthiness, history of taking risks, street smarts, quick reactions and intensity more than raw IQ, college or grades. Asking why someone started a company helps him distinguish durable motivation from opportunistic interest in a hot category. He prefers independent, candid thinkers who stand out through genuine first-principles reasoning, not provocation for attention, and often builds relationships before a founder starts a company or raises a round. He may help create a round outside a formal process, rarely competes on a term sheet and generally passes once another firm has submitted one. Because founder-investor relationships can last 10 or 15 years, he does not view the job as selling shares after a quick valuation increase.

Valhalla’s first fund plan narrowed from roughly 30–40 companies to about 10, but ultimately held 20. Pujara would rather find two to four leading opportunities a year and invest heavily, seeking differentiated companies, compelling founders and enduring moats. With an $80–100 million fund, he can lead a seed round with a $10 million check and wants to put 20–30% of the fund into an exceptional potential fund returner, rather than systematically reserving pro rata for every company. He estimates his judgment might yield one successful bet in every 12–15 investments, while acknowledging that concentration can leave a first-time fund without a standout. He contrasts this with Michael Dempsey’s data-based approach of about 30 checks per fund, which he says he has not examined in detail.

Pujara warns that short-term momentum can distort judgment: a company’s sharp rise after an investment does not prove the investment was good. Some private companies valued at $1 billion or even more than $10 billion, he believes, may ultimately be worth zero or sell for only enough to satisfy preferred investors. He sets a high bar for follow-ons, has used SPVs selectively and avoids treating follow-on money as guaranteed “reserves.” His example of a clear exception is investing 30% of a fund in a company he believes is “the next Elon Musk.” He looks to TCI, founded by Chris Hone, for its long-term approach to public-market monopolies, typically holding nine to 12 positions and reassessing whether each company remains a monopoly or is declining.

K2 Space illustrates his willingness to bet against a prevailing model: while investors favored smaller satellites, K2 built larger ones in anticipation of Starship increasing launch capacity. Pujara says that roughly three and a half years after his investment, K2 still had no competitors, had raised at a $6.8 billion valuation and had signed more than $1 billion in commercial and government contracts. He assessed the team by spending time with its members and checking whether they could attract strong people from SpaceX; lukewarm founder references are bearish, he says. He also backed JAWS after seeing Jeff’s iterative approach to providing electricity in sub-Saharan Africa work and checking the experience of investors in similar regional companies that had failed. JAWS raised $100 million and was growing roughly tenfold year over year, though local politics, violence and language barriers can make the opportunity look risky to investors unfamiliar with the region.

His own mistakes included backing promising sectors and business models when he rated the founders only 7–8 out of 10; those investments tended to return capital or small multiples rather than become outliers. He now looks for surprising ideas and excellent founders, then does enough diligence to judge the opportunity without overthinking himself into a pass. He says his process has shortened from nine months to three or even one, while acknowledging missed opportunities and that he has not tested whether an exceptional founder at a $10 million post-money valuation is a better bet than one at $100 million.