ReadingResearchRadarInvestment framework
Sign in / Sign up
← Back to reading
News3 min read·2026-10-19

Benchmark Partners: In Low-Frequency Deal-Making, the Absence of Formal Structure Is Itself an Asset

Benchmark's five general partners—Peter Fenton, Ev Randall, Jack Altman, Chethan Puttagunta, and Eric Vishria—explained at TechCrunch Disrupt 2026 why each partner makes only two or three new investments per year, why their Monday meetings have no agenda, and why the absence of a formal voting process constitutes an investing advantage. The core mechanism: extremely low deal frequency builds ten-to-fifteen-year deep relationships with founders, enabling the team to back contrarian bets on the basis of mutual trust (as with Cerebras in 2016), and if the fund succeeds, one or two investments will generate 90% of its returns.

ImportanceMaterialEvidenceE3 inspectableWrite-upStandard

2026-10-19 05:00 GMT+8

When Jack Altman joined Benchmark in February, someone asked him how the onboarding went. He said: "Perfect—there was no onboarding." Chethan Puttagunta or Eric Vishria had told him early on: "You're going to feel tight on your first one. Don't. You're going to make many investments while you're here. This whole game includes crazy power laws. The odds of it being your first are low." This is not a story about a newcomer adjusting. The sentence Altman said immediately afterward is what all five partners repeatedly argued from the TechCrunch Disrupt stage in October 2026: "Because the firm is structured around a low volume of deals, you don't actually want to have all of this rigor. If we invested in 100 or 200 companies a year, you'd want a CRM and a formal voting process. Because it's so rare, the lack of structure is a huge asset."

Eric Vishria made this mechanism concrete with the Cerebras story. In the spring of 2016, Benchmark invested in the first round of AI chip company Cerebras—their first hardware investment in ten years. Vishria recalled that the founders seemed good, they had a very simple thesis, and he himself "didn't understand most of the rest." On the Sunday night before the founders came back to pitch the partnership, Peter Fenton told him the deal "went against everything he'd learned in venture." After the Monday pitch, Fenton had flipped. "A big part of it is understanding each other's instincts," Vishria said. Fenton added: "Andrew and his co-founders are heretics. What they were saying was unorthodox, and you see it today, with people dumping on the stock because it's polarizing. That's exactly where we have to be as investors. If it's consensus, why is there a startup?"

Sign in / Sign up
ReadingResearchRadarInvestment framework

The reason this "Sunday objection, Monday flip" process is possible rests on each partner making only two to three new bets per year (Puttagunta confirmed the pace has not accelerated), the core fund holding at $500 million for twenty years, and a ten-to-fifteen-year shoulder-to-shoulder commitment with founders. Peter Fenton invoked a J.P. Morgan paper on concentrated portfolios to describe the ideal state: "the agony and ecstasy of a concentrated portfolio. If the fund is successful, one or two investments will generate 90% of the returns." He thinks the growth fund should hold four to six positions, maybe seven to eight.

But the model is being stretched by its own success. This summer Benchmark announced $2 billion across two funds, only the second time in its 31-year history that it has veered from raising a far smaller vehicle. Fenton admitted the core fund "has just inflated to be consistent with the existing economy, which we hadn't done and that was probably a tactical error on our part." Three of the four Benchmark partners who sat on this same stage three years ago are no longer with the firm. Ev Randall said "every firm is ultimately a collection of the people within the firm and how they work together"—both a strength and a vulnerability.

Randall's read on the current market gives this structural constraint its urgency. He said Character.AI's bankruptcy is being read as an "AI chill signal," when in fact that company's founder and team had been acqui-hired into Google a year and a half earlier and it had already been sold for $2.3 billion—"its Chapter 11 being read as an AI chill effect shows that the info environment for public equities in tech has been drawn so much into this, like, hyperactive X echo chamber." He went further: CoreWeave, Nebius and similar stocks trade at levels "due to the ARR figures of two private companies that will probably be public in the next 12 months," and the industry's capex and valuations all "run through the estimated and projected health of OpenAI and Anthropic." His conclusion: "There's nothing inherently unhealthy about the AI ecosystem. What's unhealthy is the volatility from data points that are often untrue or exaggerated for clicks."

Sources

  • 原始出处techcrunch.com↗
Continue researching →