Public transcript

Fred Wilson & Nikhil Raman on what it takes to be a succesful investor today.

By: USV
Length: 50 min
Language: English
Text from: the video's captions
Transcribed: 3 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

Fred Wilson and Nikhil Raman of USV sit down for a wide-ranging conversation about what it takes to be a successful venture investor today. They cover earning founder trust as a career investor, conviction in decision-making, chasing missed deals, the shift to a less collaborative and more competitive venture market, and the rise of mega-funds. They also discuss programmable biology, AI-driven margins, network effects, shared AI agents, and the value of going public.

Key points

Fred Wilson says great founders who are hard to get along with are often where the most money is made, citing Don Valentine's 2x2 matrix shown to Roloff Botha at Sequoia.
Career investors without founder experience must work harder to earn trust by doing the work—Wilson flew to Buffalo and did every job at a portfolio company for a week after founder Jordan Levy challenged him.
Wilson learned to treat the firm's money as his own money, deciding first whether he likes an investment and then convincing partners, rather than seeking their approval first.
USV passed on FOMO in September 2025, then spent roughly six months chasing the deal after Nikhil realized the mistake; the company tripled or quadrupled in size in the interim.
Venture has become less collaborative: investors must treat every deal like a jump ball and scrap for allocations, as USV did with News Research (makers of Hermes).
Wilson argues smaller firms like USV are at a competitive disadvantage versus mega-platforms and should go earlier, build domain expertise, and take on complications like geography—USV has three companies in Colombia.
Biology is becoming an engineering discipline ('programmable biology'), and Wilson believes making investments is how he learns best, as with Coinbase in 2012.
Wilson challenges anyone to find a great USV exit that didn't have network effects, and notes AI and crypto have not undone the network effects of large web 2 companies like Twitter.
USV companies are now showing EBITDA margins of 80-90% on ~$30 million revenue due to agentic coding and automated functions, which may justify order-of-magnitude valuation increases.
Wilson runs his Hermes agent on a carefully locked-down Mac Mini after being hacked four or five times, and believes companies like Amazon and Coinbase should provide agents for the data they already hold.

Questions it answers

00:00Great founders, hard to get along with

What kind of founders make the most money for investors?

Citing Don Valentine's 2x2 matrix, Fred Wilson says great founders who are hard to get along with are where the most money is made. These founders have great conviction and courage, stay true to their instincts, and often don't explain themselves.

Don Valentine told Roloff Botha: 'We make all our money here' on great, hard-to-work-with founders.
Founder mode means staying true to instincts despite board or management pressure.

Notes

Earning Founder Trust

Fred Wilson's self-critique of USV: the firm doesn't spend enough time on founders; the magic is getting both the right market and the right founders.
Don Valentine's 2x2 matrix (great founders, hard to get along with) — "We make all our money here" — still holds today per Wilson.
Career investors without founder experience must do the work: use the products, know the team, and earn trust. Wilson's story of flying to Buffalo to do every job for a week at Ron Schriber and Jordan Levy's company illustrates this.
Wilson's gift is simplification: helping founders craft a simple fundraising story and reframing the actual question being decided.

Conviction and Chasing Deals

Treat the firm's money as your own money: decide you like it first, then convince partners.
USV passed on FOMO in September 2025, then Nikhil realized the mistake while using the app. They chased the deal for about six months; the company tripled or quadrupled before they got in.
Gracious persistence — offers rejected without upset, attending the holiday party — won the allocation.

A More Competitive Venture Market

The business is far less collaborative than 40 years ago; treat every situation like a jump ball or rebound.
Mega-funds institutionalized by Andreessen Horowitz have lasting market power; USV is at a competitive disadvantage and should go earlier, build domain expertise, and accept complications (regulatory, geographic — three companies in Colombia: Audi, Somos, Pandino).

New Domains and Learning

Nikhil expanded from crypto into energy, developer-facing AI tools, and programmable biology — biology becoming an engineering discipline.
Wilson learns most from making investments (e.g., Coinbase in 2012); a little naivety helps — he might have talked himself out of Radiant's 1-megawatt nuclear reactors.

AI, Margins, and Markets

Portfolio companies generating ~$25M profit on ~$30M revenue (80-90% EBITDA margins) thanks to agentic coding and automation; may justify higher valuations and less-dilutive financing like secondaries.
A friend's company is doing an all-secondary IPO with no primary capital.
Going public when ready is powerful: anyone can become a shareholder.

Network Effects and Agents

Most great USV exits had network effects; AI and crypto have not undone web 2 networks like Twitter.
Shared agents (e.g., a couple's agent) are an opportunity for iMessage, WhatsApp, or Telegram.
Wilson runs Hermes on a locked-down Mac Mini after being hacked four or five times; companies holding your data should provide the agents.

Text from the video's own captions. Summary and notes written by AI from the transcript, so check anything important against the recording.