Venture Trailblazers, October 1, 2026. Adeo Ressi in conversation with Charles Hudson, founder of Precursor Ventures.
On October 1, Adeo Ressi sat down with Charles Hudson for Venture Trailblazers, the VC Lab series where investors who built their firms their own way talk through how they actually did it. 1,046 people registered for the session, and we still finished the hour with more than 100 questions left in the Q&A.
Charles has raised five funds and about a quarter of a billion dollars at Precursor Ventures. He has never let a single fund get above $100 million, he runs the firm with no investment committee, and he writes the first check into 80 to 100 companies per fund, usually before there is a product to look at. Over the hour he and Adeo covered how he named the firm, how he sizes funds, how he decides, what he looks for in a founder when there is no data, what the Cursor round did to pre-seed pricing, and what 1999 taught him about this AI cycle.
Watch the full conversation here: Charles Hudson on Venture Trailblazers, the complete session.
This is the full recap. Below is what he said about each of those topics, in the order he said it, with a timestamped index at the end so you can jump straight to any part of the recording.
Nine things worth taking from this session
- Size your fund backward from the exit you believe in. Precursor targets about 2% ownership, so the fund has to be small enough that a realistic exit returns it.
- Your fund size decides which wins you get to celebrate. That is the real reason Charles never went over $100 million.
- At pre-seed there is no data for a committee to debate. One decision maker, decided in the meeting, is a strategy rather than a shortcut.
- A large portfolio is a sourcing engine. Forty percent of the current fund came from referrals by existing portfolio companies.
- Discipline is a promised multiple, not a price ceiling. Seed entry prices have never sustainably fallen, so a fixed cap ages you out of the market.
- Expensive does not mean good. Cursor’s first round was $10 million, led by funds paying ordinary early stage prices.
- The best predictor of success is the first six months after the check. Precursor rates every company at that mark and lets it drive follow-ons.
- Look for founders who got off the golden path. The performable signals have all been contaminated.
- Starting a company is easy now. Scaling one is not. Compute and distribution still cost money, and most of the hard parts never got easier.
His first job in venture was at the CIA’s venture fund
Charles did not set out to be an investor. He got there through an internship at a tech company that no longer exists, where someone encouraged him to apply for a job in venture capital.
"I didn’t even know what it was, and I ended up working for" In-Q-Tel, the venture arm of the CIA. Adeo called it an oxymoron, a CIA venture capital fund, and Charles agreed and said it was great. "It’s still around too. Like they’re doing great work."
The detail he lingered on is the one worth noting. "What I didn’t realize is my first job in venture was" the thing he was actually meant to do. "I didn’t even realize that that was, like, what I was really meant to go do." Twelve years of running his own firm started with a job he took without knowing what the industry was.
He named the firm by running a March Madness bracket
Before Precursor existed, Charles sat down and ruled things out. No trees, even though the tree firms have done fine. "Sequoia, Oak, there’s the various flavors of Redwood. God bless those folks." No naming it after himself, because "you are kind of creating a dynamic that maybe makes it harder to include people in the future." And nothing obscure: "I also don’t want one of these where it’s, oh, this is the Sanskrit word for fame and fortune that’s like, nobody knows what it means."
So he generated 100 names, split them into brackets of ten, and bought anonymous online surveys asking strangers which name sounded right for a firm with his strategy.
"It was like my own little March Madness." He ran winners against winners until he had a short list. Precursor came third. He picked it anyway, because the gap between first and fifth was not material and he liked it most. He also wanted a name where the namespace and the domain space were still reasonably open.
The comments turned out to be worth more than the votes. "Some people were like, this name is terrible. Please do not ever use this name."
He never did get precursor.vc. The person who owned it was not using it and quoted a price he would not pay.
Five funds, a quarter of a billion, and the exits so far
Charles pitched his first fund in 2014, when institutional seed had only just been accepted as an asset class. He showed up selling the stage before it. "I know you just figured out seed. Have I got a story for you. I’ve got this brand new thing you’ve never thought of called pre-seed."
Five funds later, Precursor has deployed about a quarter of a billion dollars across hundreds of companies. The thesis he pitched was that he could be a good partner to both sides of the market, founders and later stage investors, and that doing that well would compound.
The exits he named on the call: The Athletic went to the New York Times, Passport Shipping sold to Global-e, DroneDeploy was acquired by Procore for around $850 million, and Elroy Air is on a path to the public markets. Fund I backed 83 companies, and 17 of them were still alive at the time of the session.
He was also direct about the timeline. Pre-seed takes a long time to reach liquidity, 12 to 13 years by his reckoning, and the best outcomes in the portfolio have not landed yet.
The funds stay under $100 million so a win still feels like a win
The math came from what LPs believed in 2014. A billion dollars was the conceptual ceiling on a modellable outcome, and Precursor targets roughly 2% ownership at exit. If the biggest exit you are allowed to assume is a billion, the fund has to stay small or the returns do not work.
He could raise differently now. "If I put out a fund model today, I think I could probably go to LPs and say, well, we’re going to assume that we’re going to have one $5 billion and one $10 billion company."
He has not, and the reason is not modeling.
"I never wanted to be in a situation where we have some massive outcome, where the founder makes $75 million and we make $25, and it’s life-changing for that person for multiple generations of their family, and I’m kind of like, it doesn’t matter to my fund. I never want to be in a world where your success and the things that make you happy mean nothing to my organization."
When Passport Shipping sold, Precursor threw a party. "The individuals who worked at the company did well. Obviously the founder did well. But all the small funds who were in that company also did well, so everybody was happy."
Full breakdown: why Precursor keeps every fund under $100M.
One decision maker, no investment committee
Charles listed three practices people told him were mistakes when he started. He still runs all three.
First, he is the only decision maker, there is no investment committee, and he can say yes in the same meeting. "At the stage we invest, there’s no data. It’s mostly perception. And I’d rather live and die by the decisions that I make than have a partner talk me out of something that they don’t like as much as I do." Twelve years on, that has become an edge he did not fully price in. "In this environment, speed is an absolute weapon."
Second, meeting a lot of people, which everyone told him to narrow. "I don’t always know where a great founder is going to come from."
Third, portfolio size. Precursor does 80 to 100 deals a fund, and the usual objection is that nobody can support that many companies. His answer is that the portfolio is the sourcing engine. Forty percent of the investments in the current fund came in as referrals from existing portfolio companies. "A large portfolio creates a large network."
He also described how deals actually get found now. Venture has flipped from mostly inbound to outbound, and a founder can be contacted, take a coffee and hold a term sheet before they ever run a process. "That person didn’t run a process. They just replied to some high-quality inbound." If that is the market, a two week IC cycle is not caution.
Full breakdown: how Precursor decides without an investment committee.
"Nobody can unsee Cursor"
Adeo pushed him on the current market, saying it feels like 2021 again with 14-person teams raising a billion dollars at a $10 billion valuation. Charles did not dodge it.
"The hardest thing in venture right now is nobody can unsee Cursor."
You can argue the next best bid was not $60 billion and it does not matter. "Somebody paid $60 billion, so it’s worth $60 billion. Like, that’s the end of the story. We could debate it, but it happened." And it is no longer a single data point. Hugging Face had already shown the threshold could be cleared. "We’ve seen a few things. We know you can clear the $10 billion threshold."
The consequence is that a billion dollar outcome stopped feeling like a result. "Suddenly the billion dollar company, everyone’s like, oh, that’s cute, but that doesn’t really help me. Which is a weird thing to say, but we’ve gotten to that place in the business."
He is careful about what that justifies. "We’re making a lot of high-priced, high conviction bets on the future. If people’s ability to forecast the future is correct, this will all be fine. If people’s ability to forecast the future is imperfect, which is usually the case, some people will maybe be fine, and some people will be ruined."
And he pushed back hard on the idea that expensive equals good. "People forget Cursor’s first round was not a mega round. It was $10 million. Dorm Room Fund, Box Group, people whose job it is to invest in early stage companies at reasonable valuations were in that company."
Full breakdown: pre-seed pricing discipline in the Cursor era.
"There’s no vocabulary police in venture"
Charles has given up arguing about round names. "Founders choose what to call their rounds based on what they think is most advantageous. And there’s no vocabulary police in venture, so if somebody wants to call their $10 million round a pre-seed, they’re allowed to do that. I can’t stop them. Nobody can stop them. The market can’t stop them." Internally, Precursor says they invest pre-Series A.
That flexibility extends to price, and it is a real change in how he thinks about discipline. "Early in my fund management career, I thought discipline meant sticking to exactly what you said you would do. So if you said we’re going to do things below $10 million, rigorously maintaining a focus on that as the ceiling. Now I think discipline is more relative."
His reasoning is empirical. "In my entire venture career, I don’t think I’ve seen a sustained downward movement in seed entry prices. I’ve seen periods of retraction and pullback, 2008, 2021, but the curve goes up and to the right." Fix yourself to a dollar figure and you age out of the market.
So what he promises LPs is a multiple, not a price. "We invest at prices where we can get the kinds of multiples I’ve promised you based on what we believe about the exit environment today."
Approaching institutional LPs
Two Q&A questions got at the same problem. Charles started with the obvious one: allocators have flooded inboxes just like VCs do, so a warm introduction beats a better cold email.
Then the part most first-time managers get wrong. "If you want to meet with somebody at a gigantic pension fund and you’re raising a $10 million seed fund, you probably are signaling that you don’t understand that person’s world." Large allocators have programmatic minimums that make small commitments hard even when they want to do it, and sending the pitch anyway tells them something about your judgment.
On whether industry expertise justifies a fund, his test is scarcity. "I have a bunch of friends who started AI-focused VC firms ten years ago, before AI investing was a mainstream multi-stage focal theme. So they saw stuff, because that theme was not popular and not well understood. If you today said, hey, I’m starting an AI-focused VC fund, people would say, you and everybody else."
Full breakdown: approaching institutional LPs as an emerging manager.
The six-month test
Asked what actually predicts success before there is traction, Charles gave an answer he finds frustrating, because you cannot know it in advance.
"The thing that seems to matter the most is the way the company performs in the first six months after they get money."
Precursor rates every company at the six month mark, and that rating drives follow-on decisions more than anything else. Sometimes it pulls a follow-on forward, before the next priced round.
The logic is clean. Strip away the specifics and most pre-seed pitches say the same thing: if I had money, I could go faster, and money is my principal constraint. "So if you add money and the company doesn’t almost immediately accelerate, money was not the principal constraint. There was some other thing."
Reserves follow from that. Precursor keeps a small reserve concentrated on the top 20, and Charles is blunt about the part of the job he likes least: "You have to be comfortable having awkward conversations with people where you’re like, I like you, but I’m not giving you more money."
Full breakdown: the six-month test for follow-on investments.
He looks for founders who got off the golden path
On reading grit, Charles thinks the old signals have been contaminated. Founders know they are being scored on responsiveness, so they are responsive during the window.
What he watches instead is whether the person has ever stepped off the track. "Most of the successful founders in our portfolio have done something risky or gotten off the golden path. People who drop out of PhD programs. People who’ve quit big jobs to do risky things. People who’ve moved to other countries."
The pattern that worries him is the opposite one. "I’ve met a lot of people who went to an elite education, worked at two or three elite companies, and then want to start a startup. This is the first time you’ve ever done something that isn’t formulaic. If I haven’t seen the person put themselves into a position where they weren’t an expert, I always worry, do you know what you’re actually signing up for?"
Domain expertise cuts both ways. Half of Precursor’s best companies were started by domain experts and half were not. "Sometimes domain expertise is an advantage because it gives you a head start. But sometimes a lack of domain experience is an advantage because it gives you the freedom to envision things working without all the constraints and knowledge of how they work today."
His bar for a yes: "In order to say yes, I have to have a vision of how the company becomes worth at least a billion dollars." Checks run $250,000 to $500,000, with some at $750,000.
Full breakdown: what Charles Hudson looks for in founders.
Does pre-seed survive AI? Starting is easy now, scaling is not
Adeo asked the existential question for the stage. If two people and no money can build a product, what is a pre-seed round for?
Charles conceded the first half. You can get something going with one or two people and almost no capital. Then he drew the line.
"It is very hard to scale things with no money, either because you need compute, or because you need marketing."
His observation from the portfolio is that a lot of markets now get crowded fast and thin out just as fast, and that the rest of running a company did not get easier. Hiring, distribution, pricing, retention, all still hard. Two companies that look identical on day one diverge quickly, and the divergence is about the operators rather than the code.
What 1999 teaches him about 2026
Charles graduated in 2000, into the tail end of the dot-com boom. The class of 2001 had offers rescinded.
He is excited about AI, and his read on the last cycle is not the one people expect.
"People were right about the internet. If anything, we underestimated the impact it was going to have on our lives. It just so happens that that first generation of companies were not the ones to capture all of the benefits."
Some did, and he named Google, Amazon, PayPal and Amazon Web Services. Others did not, and he thinks they still mattered: "I’d argue there probably wouldn’t be an Instacart without a Webvan, and there probably wouldn’t have been a Postmates without an Urbanfetch and a Kozmo."
His forecast splits the same way. "In five to seven years, white-collar professionals in your industry, pick your favorite industry, will be using AI. Will they be using it from the providers that exist today? That I’m more skeptical of."
For a fund manager that distinction is the whole game. Being right about the technology and wrong about which companies capture it is the most common way a thesis fails.
Where he lands on AI risk
Charles did not pick a camp, and he was clear-eyed about the costs. There will be job losses and he called that unavoidable, naming customer service as an example, while agreeing those were often terrible jobs. New jobs are being created too.
On who owns the problem: "Nobody wants to hold that bag. The people building it would love the government to be responsible. The government would like you to hold that problem."
And then the line that got the most reaction in the chat: "We should be careful because this is the most powerful thing we’ve probably ever" built. "I don’t think this is a move fast and break things technology. I think you could really, truly break things."
Charles is coming back
We ran two minutes over and still had roughly 100 unanswered questions in the Q&A. Adeo asked on the call if he would do a second session.
"I would love to come back. If you have me, I’d love to come back."
Round two is being scheduled. Apply to VC Lab if you want to be in the room for it.
Read the deep dives
- Why Precursor keeps every fund under $100M
- How Precursor decides without an investment committee
- The six-month test for follow-on investments
- What Charles Hudson looks for in founders
- Pre-seed pricing discipline in the Cursor era
- Approaching institutional LPs as an emerging manager
Session index
Timestamps refer to the full recording on YouTube.
- 03:37 How he got into venture, and the CIA’s venture fund
- 07:03 Naming Precursor with a March Madness bracket
- 09:38 The thesis, and why pre-seed barely existed in 2014
- 12:24 Five funds, a quarter of a billion, and the exits
- 13:22 Fund I: 83 companies, 17 still alive
- 15:44 Why the funds stay under $100 million
- 19:04 Is this 2021 again
- 22:36 Nobody can unsee Cursor
- 24:11 There is no vocabulary police in venture
- 27:25 What discipline actually means now
- 29:21 Outbound, and how deals get found today
- 32:17 One decision maker, no investment committee
- 33:39 Portfolio size as a sourcing engine
- 35:04 Approaching institutional LPs
- 36:20 Check size and what he needs to see
- 37:46 Vibe coded demos
- 40:06 Domain expertise cuts both ways
- 42:02 The six-month test
- 43:50 Does pre-seed survive AI
- 46:47 How he reads grit
- 48:22 What return makes a fund a success
- 52:33 Advice for starting a fund today
- 55:00 What 1999 actually taught him
- 57:20 Where he lands on AI risk
Watch the full session
Everything above comes from the hour Charles spent with Adeo on October 1. You can watch the complete conversation here, and the rest of the Venture Trailblazers series is on the VC Lab YouTube channel.
Launch your own fund with VC Lab
VC Lab is a free program that takes you from a thesis to a deployed Fund I, with the legal, LP and operating work handled alongside you. The managers who come through it are the people asking Charles questions like these in the first place.
Apply to VC Lab, or start at govclab.com.