Founders Fund came into Airbnb around thirtieth in line and, Brian Singerman says, made the second most money of anyone in the company.
Most firms treat a late look as a pass. His put roughly $150 million in at about a $2 billion valuation, knowing every investor who got there first would earn a better multiple on the same shares.
"It actually takes a lot of ego to put aside we passed these other people made a good decision."
Singerman spent more than a decade as a general partner at Founders Fund, where he backed Anduril, Stemcentrx and Airbnb, and became partner emeritus in December 2024. He now runs GPx, the fund he started in 2025 with Lee Linden of Quiet Capital.
I listened to the full interview so you can skip it. 63 minutes of audio, 25 minutes of reading.
Here are the 16 takeaways that matter.
👤 Guest: Brian Singerman, co-founder of GPx and for more than a decade a general partner at Founders Fund, where he backed Anduril, Stemcentrx and Airbnb
🎙️ Host: Jack Altman, a general partner at Benchmark who co-founded the HR software company Lattice and previously ran the early-stage fund Alt Capital
📰 Published: 1 September 2026 on YouTube (Uncapped with Jack Altman)
🔴 YouTube | 🟢 Spotify | 🟣 Apple Podcasts | ⏱️ 1 hr 3 min | ✅ Time saved: 38 min
Key Takeaways
He has no checklist and no rubric for founders, and says every attempt to build one failed Founders Fund tried investment memos "for like a minute" and partner meetings "for a week"
Anduril's four founders are his case study in a team of spikes rather than one complete person Only Elon Musk and Max Levchin, he said, can hold the chief-executive and chief-technology jobs at once
Founders Fund came into Airbnb around 30th and still made the second most money in it It put about $150M in at roughly a $2B valuation, knowing earlier investors had a better price
GPx sends a manager capital automatically, but only once they put about 20% of their own fund into one company
GPx charges no carry on the money it puts into managers' funds, and expects to earn on the double-downs instead
He says SPVs are dead because tier-one founders are telling the next cohort never to use one
Trust now beats brand for the best companies, because they no longer need a famous name to raise the next round
Peter Thiel's edge is assembling people who can push back on him, not his own picking
Most vintages at any firm will not hold a SpaceX, and forcing a concentration bet is how managers lose money He talked a manager he works with out of one because she could not say she was sure
He can spot the sound of someone who is genuinely the best at something because he is one himself He also says he is not a well-rounded person, and that this is a good skill in business and a bad one in a marriage
1. No Checklist, No Rubric
Altman opened on what Silicon Valley calls spiky founders — the idea that a founder carries a report card of attributes and that one of them is far off the charts. Singerman said he threw the report card away.
He keeps no list of the attributes he looks for, and says every attempt to write one down failed. "First of all, I have no list of attributes." Founders Fund brought authors in years ago to define what a Founders Fund founder was, using labels such as type A or not needy, and he said it was never accurate "And I do not have a checklist and I do not have a rubric at all."
What he looks for instead is the one thing a person is better at than everyone else, and he does not insist it be obviously relevant to the company they are building. "In fact, one of the things I love doing the most in this world still is figuring out what it is somebody is the best at and can they leverage that, right, to win at whatever it is that they're doing, right?"
Altman tied it to an argument Peter Thiel and his colleagues have made: "I feel like this kind of relates to something that Peter and all of you guys have talked about a bunch, which is like competitions for losers" — find the part of the game you are far better at, then tilt the game toward it
The method only works with self-knowledge, because someone who will not admit a weakness will not tilt the board. "I'm all about figuring out like just owning it. Owning what you're actually good at and owning what you're not good at."
Altman took the comparison from video games: "I was like a pretty big gamer as a kid and I played games like Starcraft and Warcraft and stuff and it's basically all about imbalances", where the two sides do not start from the same position and each has to exploit its own
Founders Fund ran itself on the same principle. Singerman said the firm tried investment memos "for like a minute" and partner meetings "for a week" and abandoned both — not because the practices are wrong in general, but because they did not suit that group Sean Parker rarely showed up, he said, but would surface every so often and name Facebook or Spotify. "I randomly got a chance to see like something Sean Parker wrote about Spotify" — he called it "the most amazing investment memo I've ever seen"
2. Anduril's Four Founders
Asked to connect the idea to a company, Singerman took Anduril, where he said no single founder is spiky in every direction.
The unit that works is the team, not the person. "It's the perfect example of a team of people who are very good at different things."
Palmer Luckey supplies the ideas. Singerman, who was the first investor in Oculus, described him as unbelievably brilliant and creative, a mad scientist whose ideas are not science fiction but things that can actually be built. "Do you want Palmer running the company? No." Put him on products and future products, Singerman said, rather than on running operations
Matt Grimm is the counterweight: Luckey can have one of his ideas, and Grimm can actually get it done
Trae Stephens carries the relationships — extremely good, Singerman said, at high-level connections, at high-level networking, and at building networks nobody else in the industry has
The hardest job is the chief executive's, and it is a personality job rather than a technical one. Brian Schimpf does not get riled, knows where each of the others is brilliant and where each is weak, and can put together a decision the rest will accept Singerman said there have been crazy amounts of conflicts inside the company, and that when Schimpf decides, that is the decision
Almost nobody can do all of those jobs alone. He named Elon Musk as the only person he has met who can do it all, and Max Levchin as the only other person he knows who is both a real chief technology officer and a real chief executive
Altman brought in a growth-stage chief executive he had met recently who refused to choose between an operations-minded head of sales and a selling head of sales, and said he would hire both and make them work together
3. Palantir's Shared Beliefs
Singerman said he was not there for Palantir's founding — he got involved around 2009, after Joe Lonsdale had left — so he was careful about what he would claim.
Complementary strengths are not enough on their own; the founders also have to share a core belief. He said this does not get discussed enough, that it is what sets a company's culture, and that misalignment there leads to trouble
The belief the Palantir founders shared, he said, was a hardcore pro-America outlook, which he described as really hard to find at the time
The structural parallel to Anduril is that different strengths still need someone holding them together — Alex Karp working with Thiel, in Palantir's case
He declined to go further because he was not in the room. He said he knows enough about the company but feels less comfortable describing founding-days dynamics he did not see, and offered a different example instead
4. How He Ran Pitch Meetings
He insisted on meeting every founder, not only the chief executive. "It was critical for me to speak to all the founders, all the founders, not all the head of whatever the hell, right?"
His method was to concede the founders' strongest ground at the start and then attack where he thought they were weak. He described himself as a strategy person who does not know their business — "I'm a strategy guy, but I don't know their business" — poking holes rather than testing facts
The test was who answered, not how well the chief executive answered. If a different founder took the question and clearly knew the subject, that was enough. "So it's like that was always a good litmus test for me." The failure he cannot stand is a founder who speaks about everything, including what they do not know. "I would much rather them say you know what I don't know that you should meet my co-founder."
He gets off generic answers as fast as he can. "I try and get away from generics as soon as possible." Time in Silicon Valley trains people to speak in cliché, he said, and the job is to break it: "I try and get people to like scream like this is what I'm really good at."
In hiring he uses a game rather than an interview — "I try and think of a game that we can play" — a speed round in which he names one general partner after another and asks for an off-the-cuff read, so there is no time to prepare a polished answer
Altman said his own version is silence: ask the question and let it sit uncomfortably long, because the depth is often there when someone is not used to being asked
Altman also disclosed that when the two last met, Singerman had asked him what he thought he was best at and told him there was no need to be humble, which Altman said gave him permission to answer straight
5. A+ Versus A Minus
Altman turned the question on him: what is the thing he is most confident he is good at?
He is not a diligence investor and says so plainly. "I don't do spreadsheet stuff at all. I'm not a diligence person." Everyone at Founders Fund knew it, he said, and so does his wife
What he claims as his edge is a distinction he thinks most investors cannot make. "I think people can differentiate between an A and a C founder." The one that pays is between an A+ and an A minus founder Returns follow a power law, so Founders Fund took the A+ founders and, in his phrase, "backed the truck into them"
He now thinks that framing is too coarse for the work he does. The question he asks instead is what specifically a person is better at than everybody else
The pairing is what makes it worth something. Find that, he said, then give them a company — or in his case a firm — where they can actually play to it
6. What GPx Actually Does
GPx invests in emerging managers and solo general partners rather than the big-name funds. Singerman runs it with Lee Linden, who founded Quiet Capital
The first thing it offers is advice, which he says most limited partners do not offer at all. "Like most LPs are really good LPs because they don't mess it up." He and Linden act as an adviser a manager can call about anything, and he said there are very few advisers available who have themselves returned billions of dollars in venture capital
The second is access to GPx's own base of limited partners
The third is the mechanism the fund is built around: capital that arrives automatically when a manager bets their career. "We have the ability to write programmatic and automatic capital to them when they bet their career on something so they can preempt rounds with that capital." The trigger is roughly a fifth of the manager's fund going into one company. A 1% or 2% position does not matter if it goes to zero, he said, but "But you put 20% of your fund into something, that better works." GPx gets no say in the decision. "I don't have a veto on that capital." "These are people who don't want a boss."
The capital exists to remove a delay. Without it a manager who wants the whole allocation has to raise a special-purpose vehicle, which takes a long time; with it they can preempt the round
The economics run backwards from a normal fund of funds. About a fifth or less of GPx goes into managers' funds directly, and that piece is passed straight through to GPx's own limited partners with no carry charged on it. The returns are meant to come from the double-downs
By the time a manager goes all in, Singerman expects to know the name already. He said it will not be his first or even his tenth time hearing about the company, because they will have been talking to him about it
7. Why He Quit Pitch Meetings
He stopped enjoying pitch meetings even when the founder was excellent. What he kept thinking in a general pitch meeting, he said, was how to get out of it — and that is bad when it happens in the face of greatness
The first reason he gave is standing. "I've never founded a company, right?" Meeting founders always carried impostor syndrome for him; meeting general partners does not
He compared it to a musician who wrote one hit song, loves the song, and has played it every night for thirty years until he cannot play it any more
What replaced it is not a smaller version of the same job. "Now, I still love meeting with founders and doing strategy sessions with the top founders that I can still do all day" — what he gave up is the first meeting, not the founders
Altman framed the new job as spotting people who can spot greatness. Singerman rejected the framing. "I'm looking for people who could have beaten me in my prime." He wants people who can play the game of venture capital and tilt it toward what they win at, whether or not picking is the strength they use "I'm betting on elite athletes playing a game that I know that they can win."
Selection is only one part of venture capital, he said, and he is open to backing a strategy that does not depend on it at all
8. Strategies He Will Back
Altman asked for the shape of a differentiated emerging-manager strategy, without naming firms.
The example he gave is an operator who invests in the companies he works inside. Singerman described the kind of operator known as one of the "10xers in Silicon Valley lore" — a niche skill that is irrelevant in most companies and decisive in a few In the past that person joined one good company and made it succeed. The manager Singerman described is building a portfolio of five or six non-competing companies and operating in all of them
Altman's test is that the strategy cannot be copied. A strategy that works for one fund and then attracts imitators stops working; this one does not, because the asset is the person
What he will not back is a generic answer. Singerman said "no deep tech and founders like me" is what every venture capitalist says, and the strategy has to be specific enough to hand someone a way to win
Some of it he cannot articulate, and he says so. He described meeting a manager recently and being unable to put his finger on what it was, against another case where the strategy was obvious and the person was obviously right to run it
The signal he trusts is the sense that the person could beat him. He said he had it on meeting Josh Kushner about fifteen years ago, which is why he invested in Oscar Health, joined its board and invested in Thrive He named a second solo investor whose handling of a Stripe deal Founders Fund did impressed him, and said the work would have been impressive from a whole fund, let alone one person
9. The Optimal LP Portfolio
Altman asked what he would do running an endowment or a pension fund.
He thinks an institution should be in venture capital, and should split the allocation two ways. "I put it into two buckets."
The first is the tier-one multi-stage firms and only those. Performance there is visible and the names are known; the hard part is getting a large enough check in He used his old firm as the example: whatever you can put into Founders Fund you should, and good luck getting a big enough allocation
The second is the most elite emerging managers, and the argument is the right to scale rather than returns alone. "Part two is put your money into the most elite emerging managers." An early backer of a manager who becomes the next Founders Fund gets to grow with them "Everybody loves their true believers."
This is why GPx charges nothing on the fund investments. "We want our LPs to be able to scale with these elite emerging managers, right?" He said GPx is not a gatekeeper and its limited partners can co-invest directly using its view of who is elite
The hard part is the second bucket, which is the business he is in. "There's two problems with it. A it's not enough dollars, but B there's thousands of emerging managers." Unlike the top multi-stage firms, nobody already knows who they are
10. Why SPVs Are Dead
His argument starts from where influence flows. In Silicon Valley everything runs through founders, Singerman said, so what the best founders tell the next cohort becomes the norm
The best companies are unhappy with special-purpose vehicles and are saying so. He pointed at the coverage of Anthropic and Anduril and SPVs, and said that once a company opens that door there is not much it can do about it
The consequence he expects is not immediate. The practice is not dead right now, he said, but in a couple of years people will realize "only tier four companies are doing it"
The claim is conditional on the capital cycle, and he made the condition himself. In a world with less capital a company takes what it can get and the thesis changes. "I'm very much open to adapt or die."
What he is describing is the market as it stands, which he and Altman agreed is a large oversupply of capital, particularly for the companies that are working
11. Trust Beats Brand Now
What the best founders optimize for is trust, and he said it comes from two places. "I think the best founders prioritize who they trust most" — either the person is a genuine expert the founder listens to, or they co-founded the company alongside them
That is the case for a manager who also starts companies. Singerman pointed at Founders Fund partners who founded companies themselves and said an investor who co-founds earns trust in a way advice does not
Altman said people usually hear "trust" in the personal sense — dinners, families getting together. Singerman said that is table stakes: the trust that counts is that the advice is good and the help is real
He thinks trust now beats brand for the best companies, and named the mechanism. The reason to take the famous firm was signaling that helped raise the next round. A tier-one company today does not need it: "The tier one companies just want the trust." The exception he allowed is "the topest of top tier brands", which he said can still win. A second-tier brand cannot
His example of a board built on trust is Anduril's, where he is a board observer. The board of directors, he said, is three of the four founders and nobody else
Altman named an investor who co-founds companies and said he will always get to invest whatever he wants, with the brand-name firm investing alongside him — and that a company such as Ramp or Cognition is comfortable saying so precisely because it does not need the signal
12. The Airbnb Catch-Up
Founders Fund was late to Airbnb and made the second most money in it anyway. "we were probably the 30th investor in Airbnb, but I think we made the second most amount of money"
Fixing it was expensive on purpose. "We did something 2 billion, but we put like 150 million into the company, right?"
The obstacle he names is ego rather than analysis. "It actually takes a lot of ego to put aside we passed these other people made a good decision." Coming in on top of earlier investors meant conceding they deserved a far better multiple
The lesson he draws is the inverse of throwing good money after bad. People are disciplined about refusing to add to a position they have soured on, he said, and undisciplined in the other direction, where they were wrong to pass
Ownership percentage is the wrong measure and dollars deployed is the right one, because the dollars are what set the multiple. A smaller check and a much larger one both leave the same answer: "It's like the answer is still billions."
He tied it to people who turned down early offers from Google and then would not take one in 2004 because it was already "a billion dollar company". "Like the past is the past, man."
The condition on all of it is knowing you are right. Founders Fund went big on Airbnb only once it knew, he said, and it had not known before that
Altman's version of the same point: "There's a lot of these companies that are in fact great, but it takes people a little bit of time to see it."
13. When to Concentrate
Altman said concentration is the most cited and copied strategy he hears from managers now, and that running it requires two things — having a SpaceX in the portfolio, and knowing you have one.
Singerman agreed and made both conditions honesty problems rather than analysis problems. "This is all about that whole look in the mirror thing, man."
The trap is treating your best company as your concentration bet. "The best may still not be good enough from this perspective, right?" A manager who puts a fifth of the fund behind its best name because it is the best name loses a lot of money
He said he puts no pressure on his managers to do it, because pressure would produce adverse selection
Forcing it is the named failure. A fund that turns out worse than the last one is acceptable, he said, and "forcing it loses you a lot of money"
The other missing piece is calibration. Altman: "But just knowing that like you probably are not calibrated to real greatness when you're a year or two or three into your venture career." Singerman said that is part of what managers want him for — pushing them to measure a company against Elon Musk, Brian Chesky and Max Levchin
He described talking a manager out of one. She was weighing whether to go all in on a hyped company having a strong up round and could not say she was sure, so he told her not to. His estimate of the industry: "90 plus% of managers are going to do it to play the Silicon Valley game"
Most funds at any firm will not hold a company that justifies it. He walked Founders Fund through vintage by vintage — SpaceX in one, Palantir in the next, Stripe and Airbnb in another, nothing in the one after, then Anduril — and said the funds without one are better off taking many more initial bets He still expects the funds that had one and concentrated to outperform the ones that did not
14. Inside Founders Fund
Founders Fund was never an equal partnership. "Not an equal partnership." Singerman said Peter Thiel had the final call, and used Benchmark as the contrasting model
What Thiel did not take was the largest share of the economics. A partner's carry in the current fund was driven substantially by their actual returns in the previous couple of funds, and could move up as well as down "I don't believe in entitlement." The principle, he said, is that you make a lot of other people money first and only then start making real money yourself
The advantage of variable carry is that it is not a firing. In an equal partnership the only escape valve when someone is not pulling their weight is to remove them; a variable share lets a partner stay on a smaller number and keep adding value Altman, a general partner at Benchmark, said he is biased but prefers the equal-partner model, and that its cost is a negotiation every single time — there is no version where a partner stays in the next fund at half the carry
Decisions ran on check size and often took minutes. The first step was whether a partner could persuade a colleague to meet the company; then a debate, sometimes five minutes long, settled it. A small check needed one other person, a very large one needed Thiel
A long process was itself the answer. "I could always tell it's like if somebody was doing a long process, the answer was probably not to invest."
There were no memos on the venture side while he was there, and he said the firm probably runs differently now, with its growth team carrying far more diligence. "It was really ad hoc, man."
What held the team together was respect rather than affection. Chaotic, outspoken people do things colleagues dislike, he said; there is a line nobody can cross, and short of it you respect them for their work The inverse failure is backing someone because you enjoy their company. Altman added that people who are that good at one thing are often uncomfortable to be around, and that this is what stops most investors from partnering with them
15. What Thiel Is Best At
Singerman's claim is that Thiel's edge is not his own picking. What makes Thiel the best venture capitalist in history, he said, is not necessarily his individual decision-making at the chief-investment-officer level
It is the thing GPx was built to copy. Thiel took very driven, very smart, very different people, worked out what each was good at, and let them play to it
The harder part is who he chose to hire. "I think Peter is extremely self-aware that he is not always correct and therefore you want to partner with people who you think are going to be correct in ways that you're not" Not everyone can push back on Thiel, Singerman said, and the ones who do have to be right
"This is why Founders Fund is like the greatest like in my opinion like team is very different than like a solo GP thing." He said it is the best team venture firm there is, that he is biased, and that Thiel is the reason
16. Solo GPs Over Teams
Altman asked whether GPx wants to back a team or a solo manager who later builds one
For now the model is built for solo managers, and the automatic capital is why. "I know how to do this with solo." Every manager is effectively calling GPx's money, so getting the ultra-high-conviction call right is everything
A solo manager cannot spread the blame. A manager who puts a fifth of their fund into a company has nowhere to run. In a partnership, he said, there is always an element of someone supporting a colleague's deal or half-wanting it. "This is the blame game. I hate that."
"But I think it probably works better with solo GPs just because you get the ultra high conviction stuff just nailed." He added that he is trying to work out how to make the GPx model work with partnerships
Bonus Insights
He runs a recording studio he cannot play in. "I write a bunch of my own songs, but I can't play anything." He knows enough music theory to work through digital audio workstations and MIDI controllers; he is trying to learn guitar and said it hurts The room has to cover his child's band rehearsal, family karaoke, and professional tracking and mixing
His arrangement with professional musicians is the same trade he makes as an investor. "So they get to use the studio for free." "I'll even create a label for them where they own 100% of the IP and everything." What he takes is proximity: "I get to be a fly on the wall." He drew the parallel himself once Altman pointed at it: "I'm not going to be in a band. I can't found a company."
He does not think better machines end the human version of any contest played against another person. He said he recently spoke to a chess grandmaster who pointed out that people have been "losing to computers at chess for 20 years" and still fill the stands for Magnus Carlsen On watching engines play each other: "They're alien moves, man." He compared it to AlphaGo, whose moves professionals could not explain until afterward
He drew a line between contests against people and contests against a problem. Altman raised the mathematicians who expressed sadness after OpenAI solved a set of open problems. Singerman said a solved proof does not need solving again, so mathematics may have to move to a new contest, while chess does not have that problem His summary: "I think the human experience survives regardless of how good AI gets", and "I'm bullish on humanity. I am a tech optimist." He expects real struggles over the next several years and thinks they get worked out
On whether a talent spotter has to be spiky themselves, he said he does not know. What he knows is that being among the best at something is what lets him recognize the sound of someone talking about what they are genuinely best at, the way he could tell a strong player from a conversation when he competed at games "I'm not a well-rounded person by any stretch, right?" and "I think it's a really good skill in business. It is not a good skill in like necessarily in a marriage" Neither he nor Altman could name an elite venture capitalist who is well-rounded
The job Singerman was good at — telling an A+ founder from an A minus one — is the job he now does one level up, on the managers making those calls, and his claim is that the money is made not by picking well but by knowing the rare moment to put a fifth of a fund behind a single name.
Products, Companies & Tools Mentioned
GPx (Singerman's fund with Lee Linden, which backs emerging managers and sends automatic capital when one puts about 20% of their fund into a single company)
Founders Fund (Where he was a general partner for more than a decade, and the source of nearly every example in the conversation)
Anduril (His case study in complementary founders, and a board he still observes)
Airbnb (Founders Fund arrived around 30th and, he said, made the second most money in it)
Palantir (His example of founders who shared a core belief, which he says is what sets a company's culture)
SpaceX (The standard he uses for whether a concentration bet is available to a fund at all)
Quiet Capital (Lee Linden's firm; Linden co-founded GPx with him)
Benchmark (The equal-partnership model he contrasted with Founders Fund's variable carry)
Anthropic and OpenAI (Named as the tier-one companies whose treatment of special-purpose vehicles sets the norm for everyone else, and as the lab whose math results came up at the end)
Stripe and Stemcentrx (Two of the positions he listed among Founders Fund's biggest concentration bets, alongside SpaceX, Palantir, Airbnb and Anduril)
Oculus (He was the first investor and called not going bigger a mistake on his part)
Thrive Capital and Oscar Health (Josh Kushner's fund and insurer; Singerman backed both after deciding Kushner could beat him one day)
Ramp and Cognition (His examples of tier-one companies comfortable taking a trusted individual investor and a brand-name firm in the same round)
Google (The 2004 offer people would not take because the company was already worth a billion dollars)
Spotify and Facebook (The two names Sean Parker surfaced at Founders Fund, one of them in the memo Singerman rates highest)
Stockfish and AlphaGo (The chess and Go engines behind his argument that people still turn out to watch people)
Alt Capital (Altman's fund, which he referenced when putting the trust-versus-brand question from a limited partner's side)
Books & Resources Mentioned
Sean Parker's investment memo on Spotify (Singerman said he saw it by chance and called it "the most amazing investment memo I've ever seen"; it is unpublished)
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: