[BidClub_]
Delphi Digital · · 87 min

Solo GP: Most New VC Funds Are Built to Fail | Kevin Costa

José Maria MacedoKevin Costa

VC/PEInvestingCompany Building
YouTube ↗
TL;DR
  • Kevin Costa argues that many new VC funds are structurally incapable of exceptional returns. His target is the $5–10 million fund writing $100,000 checks: even a winner may leave too little ownership to produce more than 2–3X at fund level. He calls that a “complete waste of time,” contrasting it with his hypothetical 6.5X QQQ outcome.
  • Belief Capital looks for founder psychology that can sustain a company for decades. Costa prioritizes grit and a strong worldview, with the latter serving both recruitment and “keeping the best talent in the hard parts.” Outsider backgrounds can reveal mispriced talent, but adversity stories and prestigious résumés are unreliable shortcuts.
  • Belief concentrates early using hiring quality and shipping velocity as its main operating signals. Costa’s first-check bar is willingness to commit 5–10% of the $20 million fund; one position, subsequently identified as Matt’s biotech company Serova, represents 20% of fund one. “Hiring quality” and “shipping velocity” come before ARR, with financing terms providing a third reference point.
  • Helping founders can improve both follow-on judgment and access to allocations. Costa has moved away from believing great founders need investors simply to get out of their way: he cites Valore’s operational involvement with Elon Musk’s companies as an information advantage. Belief’s practical contribution is the “first glimpse of belief,” followed by introductions that make overlooked founders legible to stronger capital.
  • Costa treats the internet as the sourcing ecosystem and San Francisco as a place to support portfolio companies. Young technical founders are “all on the same street” online; he finds them through X, research papers and universities he considers underpriced, including Oxford and Georgia Tech. His move to San Francisco followed founders rather than a change toward local sourcing.
  • Ark and Serova illustrate Costa’s preference for founders compelled by a problem. With Ark, frontline experience provides both motivation and direct product feedback; with Serova, exposure to cancer patients drove the founders toward personalized vaccines. “They didn’t necessarily want to start the company, but they had to” captures the attraction, although risks raised include Ark’s geographic revenue concentration and literal wartime survival.
  • Augusta’s investment case begins with learning how to deploy AI inside existing businesses. Its founder initially declined venture capital and spent two to three years implementing AI through consulting, building experience now applied across large private-equity portfolios. Costa calls AI “the Ozempic for the economy”; that operating knowledge could eventually support acquisitions and changes to acquired businesses’ economics.
  • José Maria Macedo frames venture as “slow-twitch capitalism,” while Costa acknowledges an early sourcing mistake. Macedo contrasts a quick 5X markup with companies that endure and warns that activity and markups do not predict returns. Costa says passing on an ElevenLabs introduction was a mistake; his ambition for Belief is “the Berkshire for the innovation economy,” with returns as the scoreboard.
Digest · the substance, structured for research

1. Small funds can fail before they pick a company

  • Kevin Costa’s argument for his $20 million Belief Capital fund starts with ownership. He sees emerging managers copying one another into $5–10 million funds writing $100,000 checks. Even when one investment works, he argues, insufficient ownership can leave the fund at only 2–3X. His comparison is deliberately blunt: “You could put your money into the QQQ, get a 6.5X,” without doing the work.

  • The opposite end of venture also disappoints him: large firms allocating roughly 0.1% of a fund to a seed company effectively treat founders as lottery tickets, while LPs receive subpar returns. Belief is meant to serve overlooked founders with meaningful conviction. Its predecessor was Costa’s personal investing: he began at 19 and backed more than 40 companies using his own savings.

  • Costa calls pedigree-based screening “intellectually lazy,” pointing to Tobi Lütke’s German apprenticeship as an example of exceptional talent outside the usual Stanford/MIT pipeline. He acknowledges that those institutions produce great founders; his objection is how heavily investors rely on them. Founders who had to fight for access may have greater resilience, while abundant capital can make privileged ecosystems softer.

2. Grit and a worldview help companies keep going

  • The two psychological traits Costa seeks are resilience and a strong worldview. Grit matters because starting an AI company has become fashionable, while enduring the work is another matter: “Why would you do this for the next 30 plus?” The motivating event need not be major trauma. A sixth-grade teacher saying someone would amount to nothing could supply an enduring, irrational drive.

  • A worldview matters because recruiting talent is only half the task. Costa’s example is Alex Karp retaining strong people through more than five difficult years at Palantir. An enterprise ontology is not automatically an inspiring mission; “we’re defending the West” gives it one. The distinctive founder can keep people committed through the inevitable setbacks of building an enduring company.

  • José Maria Macedo challenges how easily founders can manufacture both stories. Costa’s response is that “the real ones don’t fucking care” about fitting an investor’s framework: an energy founder he backed had a happy childhood and no adversity narrative to sell. Genuine motivations may emerge only at a fourth dinner, after reciprocal disclosure creates enough psychological safety—not during a rehearsed Zoom pitch.

3. Behavioral details test the founder narrative

  • Costa sometimes brings his wife to founder dinners because he thinks she catches interpersonal and ethical signals he misses. One promising founder was persistently dismissive toward a waiter. The investment question becomes how that person might treat an engineer versus an executive. If an apparently exceptional founder later commits fraud, Costa argues, “the EV of the company that you invested in is actually zero.”

  • Another founder repeatedly redirected valuation negotiations into text instead of taking an uncomfortable call. After several days, Costa treated the avoidance as a decisive negative signal. Company-building requires confronting underperformance, putting employees on improvement plans and firing people. The founder might develop that capacity, he concedes, but the demonstrated reluctance meant passing despite otherwise strong attributes.

  • Macedo’s pushback—worth keeping—is that Sam Altman has described struggling with confrontation, while Elon Musk might conceivably dismiss a waiter unintentionally. Costa answers that self-awareness changes the underwriting: with young founders, investors are betting on a steep improvement “slope,” including their ability to identify and overcome weaknesses. An isolated weakness therefore needs to be assessed alongside evidence of development.

  • Costa also distinguishes failing to notice someone from noticing them and deliberately choosing to ignore them. He admits these readings have investor-side failure modes. His broader claim is that investing at seed, growth or pre-IPO remains founder psychology; the discipline is identifying observable behaviors rather than saying someone feels impressive. “This founder is incredibly spiky” tells him little without specifics.

4. Helping founders creates information for concentration

  • Costa says he once subscribed to the view that great founders need investors simply to stay out of their way. Valore changed his mind: he cites Antonio Gracias’s involvement across Musk’s companies, including Tesla’s production difficulties, as assistance that accelerated execution. That involvement also supplied information for repeated investments. His stated result is a higher SpaceX ownership percentage than Founders Fund “at IPO.”

  • Belief’s most credible contribution is narrower: backing someone before they are obvious, then introducing them to excellent investors and making the next round easier. Costa avoids promising transformative operating help. But useful involvement can provide both better information about whether to follow on and access to allocations in sought-after companies. Macedo discloses that he is a Belief LP.

  • Costa identifies three follow-on approaches: maximize entry ownership and stop; invest again around seed and Series A; or wait until Series C/D, when category leadership is clearer. Belief prefers the early route. Its initial bar is willingness to invest 5–10% of the fund. One company—Matt’s—already represents 20% of fund one, which Macedo describes as concentration on a cost basis.

  • The follow-on evidence Costa prioritizes is hiring quality and shipping velocity. Can the founder attract exceptional people when those people could start companies themselves, and does the team deliver quickly? He explicitly notes that he has not mentioned revenue or ARR. Those observations can then be triangulated with the Series A term sheet and, if revenue exists, its ARR multiple.

5. Generalists still need an information advantage

  • Costa questions an early-stage picker’s advantage at growth against Sequoia’s Andrew Reed, who sees 300-plus growth companies annually. Macedo asks whether that endorses copying Reed. Costa distinguishes growth from Series A, admits he does not know the firm’s returns well enough, and acknowledges that copying can work if executed well. The catch is access: missing the two hardest-to-enter winners can undermine the whole strategy.

  • Sector mandates are another constraint Costa rejects. He argues that firms repeatedly delivering 10–30X over decades have tended to be generalists; a mandate tied to existing categories could have missed Uber before its category existed. He allows that specialist funds, including fintech investors, can perform exceptionally during particular cycles. His objection concerns sustaining extraordinary returns across cycles, rather than one successful period.

  • Macedo argues that even founder underwriting benefits from understanding a sector well enough to judge decisions. Costa agrees that technical claims need checking, but says this need not all happen on the call. Three or four successive “why” questions expose substance—or its absence—while subsequent research and conversations with specialist investors can test unfamiliar biology or defense claims.

6. Belief’s headquarters is on the internet

  • Costa rejects a geographic interpretation of “outsider”: someone can live in Silicon Valley and remain outside its social machinery. “The best young technical founders tend to all live on the internet,” where they are “all on the same street.” Belief will travel anywhere to meet them. His move to San Francisco followed portfolio companies; if they had moved to South Korea, he says, he would have followed.

  • Sourcing channels change. Product Hunt was valuable for several years; X now occupies more of his attention. He also reads scientific papers, including Nature papers, and looks up their authors. The aim is an “asymmetrical flow of information” developed through sustained participation and genuine interest in technical people, including physics graduates tinkering with ideas long before they necessarily become founders.

  • Costa considers MIT and Stanford relatively overpriced pools of talent, while Imperial, Cambridge, Oxford, Georgia Tech and Vanderbilt remain underpriced. Most conversations do not become investments. The wider mechanism is that “excellence attracts excellence”: helping excellent people can lead to other excellent people. He emphasizes choosing a sourcing activity that feels enjoyable enough to sustain, rather than treating each interaction as a transaction.

7. Ark combines frontline substance with existential risk

  • Ark reached Costa through a portfolio founder and an investor, both pointing toward an unusual person deploying drones in Ukraine. “Weird” triggered his good alarm bells. The founder was not necessarily a charismatic fundraiser, but repeated “why” questions revealed unusual substance. Costa describes the investor’s role as finding that substance: “Once you see that substance, you can’t unsee it.”

  • The motivation was immediate: people on the front lines were dying, and drones and unmanned ground vehicles could help protect friends and family. Costa contrasts that with wanting a YC acceptance and a fundraising announcement for LinkedIn. Frontline proximity also supplied information about actual needs, whereas established defense primes could build products disconnected from what worked in combat.

  • Macedo, who also invests in Ark, asks why Ukrainian companies appear systematically cheaper than US counterparts despite that testing advantage. Costa’s steelman includes literal existential risk, geographically concentrated and potentially cyclical defense revenue, and uncertainty about expanding abroad. Surviving a war makes company-building harder than operating a Valley software startup; resilience does not eliminate the possibility that the company itself may not survive.

  • Macedo extends the longer-term vision beyond defense, pointing to agriculture’s labor shortages as a possible market for autonomous drones over a 30-year horizon. He calls Ark’s motivation “insular”: seeing a problem in one’s own environment and deciding to change it without a nearby cohort showing the way. He finds that striking, while acknowledging that ambition learned from a surrounding entrepreneurial culture can also succeed.

8. Fundraising ability matters without proving endurance

  • Kevin Costa presses the importance of presentation: even a substantive founder may lose to a competitor who can repeatedly attract financing. Macedo answers, “I don’t know. I would say in the short term maybe,” then argues that capital can enable recruitment, manufacturing and growth. His disagreement concerns how strongly fundraising success predicts a generational company’s eventual value.

  • Macedo invokes Bending Spoons buying companies from the funding boom to challenge that proxy. Venture is “slow-twitch capitalism”: companies take 10-plus years to form, while a quick 5X markup supplies an attractive dopamine hit. His categorical warning is that investors conflate “input, activity, number of deals, markups” with output returns; he says the two are “not correlated whatsoever.”

9. Serova targets disease beyond the language-model trade

  • Macedo contrasts AGI rhetoric with what he sees as little progress against cancers and Alzheimer’s over roughly 30 years. He rejects the implication that missing the major language-model companies means missing the opportunity set. Personalized cancer vaccines are one area he finds compelling. His belief that disease can be solved comes with an explicit qualification: “It will take a lot of work.”

  • Kevin Costa says Serova’s Matt saw hospital patients who had lost limbs, mostly because of cancer, and questioned how that could persist in 2025 amid AGI claims. He entered Oxford to study medicine, dropped out and partnered with Michael, formerly a police officer. Costa praises their execution over the preceding 12 months and their long commitment: the company “has a lot of weight, but still hasn’t been voted on yet.”

10. Augusta learned AI deployment before chasing scale

  • Costa found Augusta’s founder Rodrigo on X. At the time, Rodrigo had just left Sword Health and was working for V. In 2022–2023, Rodrigo believed AI’s impact over the next 20-plus years was underestimated. Although investors wanted to fund him, he initially declined and chose to spend two to three years consulting and implementing AI. Costa sees that learning period as foundational to what the company can now execute.

  • “AI is essentially the Ozempic for the economy,” Costa says: it makes businesses fitter, leaner and stronger. The difficult part is administration—“No one likes needles.” Augusta’s accumulated implementation experience now supports transformations across some of the largest private-equity portfolios globally. Speaking from 2026, Costa thinks people may look back in five years wishing they had participated in deploying AI across the economy.

  • That knowledge could eventually support a Bending Spoons-style approach: acquire an asset and change its economics. Costa presents this as a possible expansion, with the company still early. Another advantage is its concentration of young Portuguese technical talent. In an overlooked region, a standout company can become the obvious destination for strong people before taking that talent base into global markets.

11. Overfitting founder frameworks creates blind spots

  • Costa’s mistakes of omission often begin before a meeting: a preconceived judgment that a sector is uninteresting prevents the first call. ElevenLabs is his painful example. He had an introduction available and did not take it. “It was stupid.” The lesson is particularly uncomfortable for a founder-driven generalist: filtering by the category can prevent discovering the person who changes the assessment.

  • Psychological frameworks can also become too deterministic. A trait absent on the first call may emerge through a longer relationship. Costa admits overfitting to his preferred criteria, but refuses to declare every subsequently marked-up company a proven miss. The firm is only four or five years in; a bull market can make everything look good. “I’ll tell you in five years, or in 10.”

  • Different businesses require different strengths. An execution-heavy operation may need competitiveness more than extraordinary engineering intelligence; sports or gaming can reveal that drive. Costa says demanding an Elon Musk-like technical profile for such a business can exclude the right operator. He still favors companies with 30-year potential, while acknowledging that this preference might make Belief look foolish in the short term.

  • The calibration exchange cites Deel’s growth of around 10% week over week even as its base expanded, and founders still answering payroll questions on WhatsApp within one or two minutes at $20 billion scale. It clarifies that these lessons came from observing communication and execution, not spending extensive time embedded with the teams.

12. Belief’s ambition extends beyond the solo-GP label

  • Finite and Infinite Games shaped Costa’s preference for deep relationships with a few high-quality people. Depth can create greater network “width” because trusted contacts make meaningful introductions. He has built those relationships since entering investing at 19, rather than trying to know every GP. The same logic supports helping people over decades, even when there is no immediate investment or obvious transactional benefit.

  • That long-term orientation coexists with aggressive allocation decisions. Belief wants to lead first rounds and take as much ownership as possible; Costa openly says it is not especially collaborative at that point. Once invested, introductions and broader cooperation help the company. He sees a balance between securing what the firm needs and avoiding the scarcity mindset that makes every relationship feel zero-sum.

  • Costa says some of the best founders choose fewer partners from inception through IPO. Belief wants to support that entire journey, using information gathered early to invest at other stages. He rejects labels such as solo GP or multistage: the ambition is “the Berkshire for the innovation economy” over 50-plus years. It remains early, he stresses; returns are the ultimate scoreboard.

Full transcript
Kevin Costa

Some of the best founders started companies for bigger reasons than that. They didn’t necessarily want to start a company, but they had to. People were dying on the front lines, and someone was developing drones and deploying them there to help prevent their families and friends from dying. That’s a very different motivator from, “I want to get into YC and raise a fancy round so I can post about it on LinkedIn.”

AI is essentially the Ozempic for the economy. It makes it fitter, leaner, and stronger. The best young technical founders tend to live on the internet. Sure, you might back an incredible founder, but if they’re unethical or commit fraud, the expected value of the company you invested in is actually zero.

José Maria Macedo

All right. Hi, everyone, and welcome to another episode of the Emerging Managers podcast. I’m José Macedo, and today I’m excited to be here with Kevin Costa, founder and general partner at Belief Capital. Kevin got into tech at 17, helping build Forto, the Berlin logistics startup, and then joined La Familia, now part of General Catalyst, where he worked on Deel and Stoke Space at the seed stage. Alongside all that, he’d been angel investing with his own savings since he was 19, investing in more than 40 companies, including Bland and Superpower.

Last year, he finally took the step and turned that into Belief Capital, a $20 million solo GP fund backing young technical founders before anyone else will. He’s backed companies like Ark, Augusta Lab, and Seroba, which are already in the portfolio. What I really like about Kevin is that he has one of the best founder lenses I’ve seen in Europe. He surfaced multiple companies that I ended up backing personally, and he thinks about founders in a way most VCs don’t. So I’m really excited to jump into that with you today, Kevin.

Kevin Costa

Amazing. Pleasure to be on.

José Maria Macedo

To start with, do you want to tell us a bit about yourself? I gave a brief intro, but maybe you could put it in your own words, because I think for many people it might be the first time hearing from you.

Kevin Costa

Yeah, for sure. I grew up in the middle of nowhere, on the outskirts of Lisbon, Portugal. It’s actually not too far from where you are, probably.

José Maria Macedo

Which is where I am now.

Kevin Costa

Yeah, exactly. There wasn’t much going on in Portugal, not much of an ecosystem. I grew up addicted to video games, so I played way too much. That’s how I started meeting interesting people on the internet, which is essentially what we do for a living, I believe. We nerd-snipe interesting talent on the internet and then partner up with them. That’s sort of our job.

My mom’s Portuguese, and my dad is German. I spent some time in Germany after Portugal. I thought Portugal was boring, then found out Germany was even more boring. That’s how I stumbled into tech when I was 17. I got lucky and was exposed early on to the startup world, back when everyone said I was a loser for going into startups. I got lucky to be exposed to a very fast-growing company in Germany, and then got to see a bit of the other side early on.

Over the years, I kept seeing young technical people who were sort of my peers, and whom no one really wanted to underwrite, starting to build very interesting companies around 2019 and 2020. I saw the rise of companies like Deel and Stoke Space, and put all my savings into them, for better or worse. Over the last 6 to 7 years, I’ve put all my personal capital into backing companies. That was really the ethos, and sort of the predecessor, to what Belief is now. So that’s a bit about me.

José Maria Macedo

Why do a $20 million early-stage fund? You were working at La Famiglia, a pretty big fund. You presumably had offers from other funds. Why do your own $20 million fund, and how do you beat the big guys?

1. The Case For A Smaller Fund

Kevin Costa

I think the one thing I kept seeing was a lot of my peers and other people starting funds over the last few years. It’s become this common, cool new thing to do. I think a lot of people set themselves up structurally for failure because they want to do it simply because it’s the cool new thing.

Most people look around, see what everyone else is doing, and try to mimic that. The outcome ends up being a lot of the same: $5 million to $10 million funds, $100,000 checks, a spray-and-pray approach, and hoping for one company to stick out. Even if one does, you have such low ownership that structurally you can only get to a 2x to 3x return at the firm level, which is a complete waste of time, in my view. You could put your money into the QQQ, get a 6.5x, and not do any of the work.

I kept seeing that, with everyone basically doing the same thing. At the same time, I kept seeing young founders who were underserved by venture. A lot of the large firms treat them like lottery tickets, deploying 0.1% of their fund into a company at seed and indirectly treating it as a lottery ticket. On the other side of the spectrum, the LPs in these firms get subpar returns.

I wanted to build something completely different, specifically around the thesis of backing undiscovered talent that I still think is completely mispriced. Not necessarily the Silicon Valley mold of kids, but underdiscovered and underpriced talent.

José Maria Macedo

Yeah, maybe talk a bit about that, because I think you have a very unique view on founders.

Kevin Costa

Yeah.

José Maria Macedo

So maybe you can talk a bit about that.

2. The Outsider Founder

Kevin Costa

Thank you. Look, I think we talked about this: if you look at some of the largest companies, you see founders like Palmer Luckey, who went to Palm State College, not Stanford or MIT. You see Tobi Lütke, with a random German apprenticeship. You see Steve Jobs at Reed College. Elon was at Penn. All of these incredible founders aren’t in the Silicon Valley mold, and the list goes on and on.

Yet if you look at most venture firms, they use intellectually lazy ways of screening for talent. They’ll say, “Okay, we’re going to look at ex-Palantir people, or the Stanford and MIT kids.” There are some great founders from those places; I’m not saying nothing comes out of those ecosystems. But I think they’re overdetermined relative to what actually produces incredible outcomes.

Some of that probably stems from my childhood. I grew up in the middle of nowhere, wasn’t an insider, and wanted to find kids like me who I thought had much more talent and make them legible to capital. That’s mostly what we look for. I think some of the founders we just talked about have much more grit because they come from non-obvious backgrounds and have something to prove.

For better or worse, people from more privileged backgrounds get money thrown at them. They have it softer. It’s not their fault, but the ecosystem makes them softer by default. If you grew up in the middle of nowhere, had no venture funding, and had to really fight for it, I think you have much more resilience and grit. We tend to think those profiles are much more interesting.

José Maria Macedo

And what do you look for exactly? How do you find out that someone has grit? Presumably, you’re looking for an outsider from the beginning, right? Is that your first test, or what do you look for in them?

Kevin Costa

Not necessarily. We look for 2 psychological traits, which we don’t find too often. If we do, we’re happy to put up to 5% to 10% of the fund into a company at the super-early stage.

José Maria Macedo

What are the 2 psychological traits?

Kevin Costa

Trying not to leak too much alpha. One of the things I just touched upon is that we try to screen for resilience and grit in a world where everyone wants to build an AI company and gives up after 2 years because it’s actually really fucking hard. As Elon says, it’s like chewing glass.

I think the biggest issue now is that tech as a percentage of GDP has become very, very large. It’s no longer a niche thing; it’s just become a job, essentially. So you have a lot of tourists. That’s one thing we try to screen for.

Normally, it can appear as some type of event early on, like childhood adversity, although I want to say again that it’s very overdetermined to some extent.

José Maria Macedo

Yeah.

Kevin Costa

And also very performative, I would say. It’s nuanced how you actually find those things out. It’s not on a Zoom call, and I don’t think it has to be a big trauma or some major event that happened. For a lot of people, including some of the names I mentioned, it’s super nuanced, and we never talk about that nuance.

Maybe it was the one teacher who told you in sixth grade that you weren’t going to amount to anything, and that’s still lodged in your head. Those nuanced things that happened to you can be the irrational drive. I don’t think you have to have had a huge childhood trauma.

José Maria Macedo

Yeah.

Kevin Costa

You look at Zuck, you look at Bill Gates, who grew up in a quite wealthy environment but was bullied in school. It was his slight, nuanced chip on his shoulder that led him to build the behemoth that is Microsoft today. I think it can be very nuanced, but it’s something you screen for: in a world where everyone wants to build a company and gives up after 2 years, why would you do this for the next 30-plus? There aren’t too many people building companies for 30-plus years.

José Maria Macedo

Yeah.

Kevin Costa

So that’s one.

José Maria Macedo

So grit is one. What’s the other?

Kevin Costa

The other one dates back to the fact that there aren’t many people building companies for 30-plus years.

José Maria Macedo

Mm-hmm.

Kevin Costa

There are a lot of short-term traders. We can talk about the trader mindset in venture, but it’s about having some kind of strong worldview. I can give you an example.

José Maria Macedo

Yeah.

Kevin Costa

If you look at Alex Karp, why did the best talent stay at Palantir when the company wasn’t going anywhere for 5-plus years? It was because of Alex Karp’s strong disagreeability and strong worldview, which led him to attract some of the best talent and keep that talent through the inevitable ups and downs of a generational company.

Everyone talks about the first part, attracting the best talent. No one talks about the second part: keeping the best talent through the hard parts of building a generational company. All generational companies go through hard parts, and it’s about the stickiness of attracting and keeping the best talent. These magnetic founders, with an incredibly strong worldview of the future and how their company fits into it, do incredibly well.

I don’t think every company needs a SpaceX-type vision, because that’s just not the case. Palantir, I would argue, is quite a boring company, per se. They’re building an ontology, which is, to some extent, integration for larger enterprises to develop or implement AI. But Alex managed to create this vision of “We’re defending the West, and we’re protecting the West,” and people get behind that vision.

There are very few people who can rally talent and keep that talent through the tough times. Those are some of the high-level things I look for, but I try not to leak too much alpha.

José Maria Macedo

Is that because you’re worried about founders reverse-engineering what you’re looking for?

Kevin Costa

Yeah, to some extent. A lot of what I said has nuance, and there will be a lot of founders who don’t fit the mold. We always like to put things into buckets, and that’s where you make the most mistakes in venture. You always have to be quite open to it, so I try not to be too deterministic, although we do have some frameworks.

José Maria Macedo

What’s some of the nuance? A lot of people will say the same things, right? They’re looking for founders with independent visions, people who are gritty, and so on. Can you think of an example? You said Zoom calls aren’t good; maybe you’re meeting them in real life. What questions are you asking? Most founders, I think, will at least have a manufactured story that fits these 2 points. It’s become widespread enough that every founder—or most of the good ones you’ll meet—will have a story about how they were gritty and a story about their worldview that sounds unique and disagreeable. They’ve answered the Peter Thiel question. How do you tell the real ones?

Kevin Costa

Mm-hmm.

José Maria Macedo

Because it’s sort of become widespread enough that every founder, or at least most of the good ones you’ll meet, will have a story about how they were gritty and a story about their worldview that sounds unique and disagreeable. They’ve answered the Peter Thiel question. How do you tell the real ones?

Kevin Costa

The real ones don’t—I don’t know if we’re allowed to swear, but the real ones don’t—

José Maria Macedo

Yeah, you’re allowed to swear.

Kevin Costa

—fucking care.

Kevin Costa

A founder we backed earlier this year is building an energy company. He doesn’t fit into any of these molds. He might not have the manufactured story, and he actually doesn’t care. He’s going to build this behemoth of a company, and if some venture funds don’t have the checkbox of crazy childhood adversity, because he’s actually had a really, really nice childhood, then so be it.

The real ones don’t care about any of the molds or filters, and they don’t care about posturing. The detail is that I would never share what drives us to build belief in a huge behemoth for the next 50-plus years on a Zoom call. I might do it the fourth time we meet for dinner, maybe after a few drinks, and maybe after you share what makes you who you are and what your insecurity is, or what really drives you to build Delphi. That opens up psychological safety in the room for me to potentially open up about what drives me.

It’s nuanced. When I meet someone over dinner, I often bring my wife to meet founders. I think we men are quite good at screening for IQ, raw ambition, horsepower, and that sort of thing. But we tend to be quite bad at EQ, and maybe ethics as well.

That’s under-discussed, too. You might back an incredible founder, but if, in year 6, the company grows like a rocket and the way the founder treated the waiter—which was very nuanced and maybe only my wife would have picked up on—means the person isn’t ethical or commits fraud in year 6, then the EV of the company you invested in is actually zero. You need to screen for these things. All of this is to say there’s a lot of nuance.

José Maria Macedo

Mm.

Kevin Costa

So, yeah.

José Maria Macedo

Can you think of an example where a founder was extremely close, and you were very excited about them? They passed the grit test, they passed the independent-thinking and disagreeable-worldview test, but then something happened that made you pause or realize they weren’t the real thing?

Kevin Costa

There are definitely a few. One example is treating the waiter. I don’t want everyone to know—

José Maria Macedo

And what was it exactly? Did he snap at the waiter, or—

Kevin Costa

No, I think it was nuanced. He was dismissive toward the waiter over a long period of time in this specific case. I didn’t necessarily pick up on it very well at first, but once I started noticing it, it became quite obvious. Then you have to map that onto how the person will lead a company for the next 30-plus years, and how he’ll treat the forward-deployed engineer versus the executive. That was one example, but there have probably been a few more.

José Maria Macedo

Can you think of a few more? I’d be curious to dig into those.

Kevin Costa

Mm-hmm.

José Maria Macedo

It doesn’t have to be an ethical thing. It could be something that made you realize they weren’t as smart, ambitious, disagreeable, or whatever else you’re looking for as you thought, for someone you were otherwise very bullish on.

Kevin Costa

Yeah. There was one recently where I think we were very convinced about the founder. One of the points around disagreeableness is that, when we’re trying to negotiate the price or set a valuation for the company—

Kevin Costa

Mm-hmm.

José Maria Macedo

—it became quite apparent that the founder was trying to push all communication toward text.

Kevin Costa

And so, rather than jumping on a call and facing the confrontation head-on—which is very uncomfortable—after multiple days of trying to get someone on a call and have that very uncomfortable conversation, which is uncomfortable for both sides, he was always avoiding confrontation. That’s an incredible negative signal, because building a company means doing this every day, right? You have to tell an employee who maybe isn’t performing that he’s on a PIP or might be fired. There’s a lot of discomfort you have to face every day.

José Maria Macedo

Yeah.

Kevin Costa

And that was very nuanced, because in all the other respects, you’d say he’s incredible. But that was one area where he was lacking. That’s a lack of disagreeableness, or a willingness to push through if you have a strong worldview. Again, I’m not saying he can’t develop those qualities over time, but it’s a no.

José Maria Macedo

And was that one where you said no just based on that? Sam Altman famously says that he really struggles with confrontation, and that the problems people accuse him of being unethical over are actually because he struggles with confrontation. I can imagine Elon unwittingly being rude or dismissive to a waiter, or something like that. It’s nuanced in the sense that if there’s extreme strength, you’ll ignore some of those things.

Kevin Costa

Mm-hmm.

José Maria Macedo

And, yeah.

Kevin Costa

The devil is in the details. Sam managed to be self-aware about his unwillingness to confront things, or his dislike of confrontation. I think self-awareness is one piece of it. What we’re betting on a lot with founders, especially younger ones, is that their slope is high: they not only identify their weaknesses, but get better and almost demolish those weaknesses over time. You have to underwrite the slope as well, and I think that might be the case with a few of these founders.

Kevin Costa

Mm-hmm.

José Maria Macedo

But you mentioned Elon, right? He confronts the truth like crazy, at all times, and that’s very strong. Can he be rude to a waiter? Yes. But I think you have to distinguish between being rude because you’re not aware of someone in the room and being aware of someone in the room but actively choosing to ignore that person.

José Maria Macedo

Mm-hmm.

Kevin Costa

And, again, there are a lot of failure modes on my side, or on the investor side, that might happen. There’s also a lot of nuance in how you read into that.

José Maria Macedo

As you speak, I feel like this is very much a gut instinct read for you, which you then try to transpose into words and mental frameworks. Is that accurate, or would you say it’s an analytical process?

Kevin Costa

It’s definitely psychology. It’s trying to understand someone’s wiring. I think all we do is psychology. At the seed stage, at Series C, at pre-IPO, I think everything is founder-driven investing. The rest is kind of cope. A lot of growth investors might tell you, “Hey, this is the Excel sheet and how we project all of this,” but I’d tend to think all of it is just founder-driven investing in the best companies at any stage.

José Maria Macedo

Mm-hmm.

Kevin Costa

If you look at SpaceX, even a year ago, where was the $20 billion in compute revenue? It wasn’t there, so you couldn’t necessarily have modeled that out. I think you always have to underwrite the psychology of the founder, which is incredibly hard and complex. But there are things that are factual, which you can get to: the ability to face confrontation, the ability to be disagreeable, and a few other things you can nail down. There are signs around all of this stuff.

Kevin Costa

Yeah.

José Maria Macedo

It’s not just a gut feeling of, “I feel good about it.” Even we have to get better. I hate hearing, “This founder is incredibly spiky. You should meet them.” It’s like, okay, what does that mean? What specifically about him made you interested or excited?

José Maria Macedo

A lot of this stuff is about calibration, right? When you meet a really great founder, you start off with one idea of what a great founder is. Presumably, at 19, you had one idea. Who would you say are the people you calibrate against? Who did you meet and spend time with that really helped you calibrate the way you see founders?

José Maria Macedo

I think a large part of being a very good venture investor is being exposed to greatness early on. That basically calibrates you. If the best founder you’ve ever met is the best founder in the Lisbon ecosystem, you’ll have a local maxima problem. That’s what you think is great, so when you tell me, “I met a great founder,” that will be the bar.

I got lucky, to be very honest, whilst I was at LF[?], seeing the team at Deel grow at about 10% week on week. As the base was getting larger, they were still growing at that rate, and that slope was like, “Okay, well, that’s great.” Then there was the team at Stoke Space, where Andrew was just shipping and shipping and shipping in very short periods of time. That set my bar for what’s great.

Over time, I’ve met other teams that give you a different flavor of what great looks like, because different companies have different founder archetypes that do better. But one of the main things is that you have to meet greatness to know what it is. I think that’s the difference between a very good investor in venture and one who isn’t.

Kevin Costa

And you got to spend time with them and see them operate. What was the biggest learning from each of them, or the thing that really stood out?

José Maria Macedo

I wouldn’t say I got to spend a lot of time with them. I got to meet the heads of sales at these teams. Even recently, when we got onboarded because we’re looking to run payroll, the founders and founding team still replied within 2 minutes or 1 minute on WhatsApp. These are small things that never change. It’s easy to do when you’re a pre-seed company, but really hard to do when you’re a $20 billion market cap company.

So I wouldn’t necessarily say you need to spend a lot of time with them; it’s more about seeing what the teams do. Just from being exposed to both companies, the way they act and the way they ship, I think you learn a lot about what greatness looks like.

Kevin Costa

With those two founders, was it mainly the intensity, the speed of building, and responding in 2 minutes? Or were there other big lessons you took away? Where did the differentiated worldview, disagreeableness, or grit come from? Did you hone those frameworks over time by meeting these founders?

José Maria Macedo

Yeah, a lot of trial and error, deploying my own capital over close to 10 years, and learning what specifically I found to be the best predictor. But, yes, I’ve learned about execution speed and a very strong worldview from those teams, as well as relentless execution.

For some other types of companies, relentless execution isn’t the only thing you need. For a company like Deel, that is what you need, and it worked quite well. But the very strong worldview is especially important when you look at companies in general today. There are so many companies, but only 5 to 10 really matter every year, if that. You could even argue it’s fewer.

If you look at the companies or firms with epic returns, 10x to 30x, they tend to back companies that exist and stick around for 30-plus years. If you look at Founders Fund and Sequoia, a few companies yielded most of the returns, and those tend to be companies that endure over very long periods of time and compound even at very large scales. Apple, NVIDIA, and SpaceX are the companies you really want to be involved in over long periods of time.

Kevin Costa

And those yield most of the returns in the industry. I think it’s really about increasing your surface area to be exposed to those companies, and then being the best partner to them over very long periods of time. I think that’s the second part Nolan talks about. We talk a lot about writing early checks into companies, but how about being the best partner for them if you believe the company will stick around for the next 40-plus years?

José Maria Macedo

Yeah.

Kevin Costa

And so I think that’s the other part.

José Maria Macedo

Yeah. Maybe we can talk about that a bit, then, because you do a really good job at this. All the founders I’ve spoken to really love you. But some venture investors have this view that you can’t do anything for your founders: the right ones don’t need your help at all.

Kevin Costa

Yeah.

José Maria Macedo

And that it’s kind of just a LARP that venture investors do. It seems like you disagree with that based on your actions, but I’m curious: what do you think a venture investor can do to really add value to a founder, and what do you focus on?

3. Beyond The First Check

Kevin Costa

Yeah. I think that notion is completely wrong. I actually used to subscribe to it. It’s the Founders Fund-ism: the best founders don’t need your help; just get out of the way. The case in point for why that’s wrong to some extent, though Founders Fund still does fantastically well, would be Valore. Both Valore and Founders Fund are investors in SpaceX and Elon.

Valore has doubled, tripled, quadrupled, quintupled down into the company. They’ve embedded very deeply into all of Elon’s companies, helping with operations, stepping in during Tesla’s production hell, and unblocking a lot of roadblocks. That gave them the asymmetrical information to keep doubling down into the company. At IPO, they have a higher percentage ownership of the company than Founders Fund does.

I think that’s the perfect use case: Elon is a founder who doesn’t need help from anyone. He’s probably the one person who wouldn’t need anyone’s help to get where he is, and he still openly talks about Antonio Gracias and Valore being incredibly helpful—not by helping, but by accelerating the path to him getting there. I think that’s the perfect example of what you should do if you can, because it gives you more information about the companies, which then gives you a better ability to concentrate.

We talk about concentration a lot, but a lot of follow-on decision-making from firm managers I speak to comes down to, “Is Sequoia doing the Series A or the Series B? If so, we’re going to plow in a lot of capital.” I’m being flippant here, but that’s essentially how they make follow-on decisions. I think that’s factually very wrong. Again, you look at the data: a lot of Sequoia’s follow-on rounds end up not being good follow-on investments, so it’s probably not the best mental model for how to do those.

I do think there’s value you can provide. The best founders probably don’t need you; you won’t be consequential most of the time. But I do think the best founders appreciate you being early and being the first glimpse of belief. If you help accelerate their path, that gives you the ability to know when and if to follow on, and, second, the ability to get the allocation in the most wanted companies.

José Maria Macedo

Mm-hmm.

Kevin Costa

The most wanted companies.

José Maria Macedo

It’s a good point. I never thought of it that way. What do you do, then? Antonio Gracias is a seasoned operator. He really stepped in with a bunch of Elon’s companies in that sense.

Kevin Costa

Yeah.

José Maria Macedo

What do you think your biggest value-add can be, and what do you focus on?

Kevin Costa

Yeah. Again, I tend not to like to overpromise, especially in a live format. But for us, if you look at some of the companies we’ve backed, we were the first glimpse of belief in them when they were in the middle of nowhere, not in the Silicon Valley mold. The best thing we have done, and still do, for them is make them legible to really good investors and really good capital.

I think the best thing you can do is be the first glimpse of belief in someone, and then bring them to some of the best folks along the way—in other words, help make the next round easier. I think that’s what most investors can do well. The really good ones do that really well, and for the rest, the value-add stuff is very incremental. We try to do a lot, and it is a big part of how we’re going to build belief, but I don’t want to overpromise.

José Maria Macedo

Mm-hmm. Yeah, you talked about this in your previous answer: how being value-add helps you concentrate because it gives you more data to evaluate a founder. How do you think about concentration?

Kevin Costa

Yeah.

José Maria Macedo

Are you mostly doubling down on a company once you’ve made the first investment and seen them execute? Or will you put a large percentage of your fund into a company before seeing the founder operate? And after seeing them operate, how concentrated do you go?

Kevin Costa

Yeah. Follow-on investing, or concentration, is one of my favorite topics because I think there’s a lot of intellectual laziness in the industry. I talk to some of the people I arguably think are among the best pickers and investors out there. There are three schools of thought on follow-on investing. One is: don’t do it. Just get as much entry ownership as possible from the beginning. Basically, don’t deal with it.

The second is that you do it early, at the seed and Series A, when you’re close to the team and pay a lower entry price. The failure mode here is a false positive. Again, a lot of Sequoia’s Series A investments end up not being good follow-on investments. The third school of thought is to plow into companies at Series C or D, when you know they’re category-defining, and avoid the false positive.

I can talk about how we see it at Belief. We’re very concentrated. To answer your question, we’re mostly only willing to write a check if we’re willing to put up to 5% to 10% of the fund into a company at the first check, at the belief stage. That tells you about the bar. One of the companies where we have 20% of Fund I is quite early.

José Maria Macedo

Which one is that?

Kevin Costa

That is quite early.

José Maria Macedo

If you don’t mind sharing.

Kevin Costa

Matt’s company.

José Maria Macedo

Oh, cool.

Kevin Costa

Yeah.

José Maria Macedo

Awesome. Congrats. Twenty percent on a cost basis.

Kevin Costa

Time will tell.

José Maria Macedo

What gets you to the conviction to put 20% in? Because I know you went—

Kevin Costa

Yeah.

José Maria Macedo

Mm-hmm. Yeah.

Kevin Costa

So, to finish the thought: there are three schools of thought. One is don’t do it, one is do it early, and the third is plow money in when you know it’s a category-defining company. What I think is quite interesting for Belief is doing it early. We spend a lot of time with a small set of teams assessing hiring quality, which comes back to what we look for in founders: can they attract the best talent?

Especially in a power-law economy, the best talent can do whatever they want. They can raise money to do their own thing, join a company, or whatever. Hiring quality is crucial. The second thing is shipping velocity. Those two tend to be asymmetrical data points we can get by spending a lot of time with the teams. Note that I haven’t said anything about revenue or ARR. Those are the 2 metrics we look for.

Then you can triangulate those two with a third factor: the term sheet from the Series A fund, and maybe the ARR multiple on that if the company has revenue. Between these 3 things, you can make a really good decision about whether to follow on.

José Maria Macedo

Mm-hmm.

Kevin Costa

That’s how we do it at Belief. My challenge to a lot of growth investors is that I think it’s delusional to believe that even a top 0.1% early-stage picker here can have an advantage at Series C against Andrew Reed at Sequoia, who looks at 300-plus growth companies and the best of them every year.

And so I tend to think it’s a fool’s errand to think that you can do better here.

José Maria Macedo

Doesn’t that tell you to just copy-trade Andrew Reed? Isn’t that a vote for the strategy we were hitting on before, of just copying Sequoia into these rounds?

Kevin Costa

At the growth stage, it’s different than at Series A.

José Maria Macedo

Okay. So you think that at the growth stage, copy-trading Sequoia or whatever is a viable strategy, actually?

Kevin Costa

I don’t know their returns too well. There are individuals within the firm who are notoriously very good.

José Maria Macedo

Yeah.

Kevin Costa

Yeah.

José Maria Macedo

I think it is actually well—

Kevin Costa

Notoriously very good.

José Maria Macedo

I haven’t checked either.

Kevin Costa

Yeah.

José Maria Macedo

Someone told me there was a fund whose thesis was to copy-trade Sequoia and a handful of other Tier 1s into rounds: if they were in, you’re in. They actually did pretty well, which was surprising to me, but interesting. An interesting data point.

Kevin Costa

Yeah. No, it does well if you do it well. But again, this is the thing: if you don’t get into the 2 companies that are the hardest to get into, because even Sequoia had to fight really hard for that allocation at Series A and B, your returns go out of whack. If you don’t have the Apple or Nvidia allocation, but you have all the other companies—

José Maria Macedo

Yeah, that’s fair. And you’ve invested in companies across defense, biotech, and health. By the way, I’m an LP in your fund. I feel like I should have said that in the intro, but full disclosure. You’ve invested in companies across all these areas, and I know your biggest, most concentrated position is a biotech company.

Kevin Costa

Mm-hmm.

José Maria Macedo

Obviously, you’re not a specialist investor in any of these areas. To what extent do you try to understand the sector and the business? Are you looking at the competitors and trying to understand the technical differentiation and all of that? To what extent do you need to do that as a founder-driven investor?

4. An Internet Native Firm

Kevin Costa

I think you can get there. We’re completely founder-driven. Sector-specific funds tend to be quite stupid. If you look at the last 50-plus years and the most generational firms in terms of returns, the firms that spin out 10–30x fund returns back to back tend not to be sector-specific. You couldn’t have caught Uber in a sector-specific fund; the sector didn’t exist. You couldn’t have caught Airbnb in a sector-specific fund; the sector didn’t exist. The list goes on and on.

I think it has worked for some firms, where throughout cycles you can make great returns. In fintech, for example, there was a cycle where some firms had incredible returns. But in terms of the duration of returns and the resilience of those extraordinary returns over time, generalist firms tended to prevail.

We normally just try to look for these 2 psychological traits, and that’s really it. It can be a team building cancer vaccines or a team building defense. With defense, for example, once you ask 4 “whys” and drill down into someone, making the first call very uncomfortable, I think you can get a high degree of fidelity on how technical someone is and how well they’ve understood their space.

I’ve asked 3 or 4 “why” questions of a lot of Silicon Valley teams building in defense, and at the third “why,” they crumble. I don’t think you necessarily need to be a biotech expert or a defense expert. You need to be really good at understanding psychology, understanding the individual, and seeing if there’s truth.

José Maria Macedo

But the “whys” will eventually get down to hard, first-principles answers about the technology, right? If you drill down enough into the “whys,” your position is that they should be able to explain it simply enough that you can understand the rationale without being an expert in the sector.

I agree that you don’t need to be a sector expert. But even if you’re a founder-driven investor, it’s useful to understand the sector somewhat so you can judge some of the founder’s decisions, right?

Kevin Costa

Yeah, but I don’t think that has to happen on the call.

José Maria Macedo

Okay.

Kevin Costa

You might learn something about biotech on a call that I don’t understand, through a consensus check. I can speak to some of the best biotech investors, who are close friends, and try to get their point of view. I can do research on my own. I can sense-check the proof of work to some extent. But what you want to get on the call is just a sense of how much substance is really there. Mm-hmm.

José Maria Macedo

Interesting. And your conviction to concentrate is based on your founder read again: seeing them execute over time, deciding who you rate higher, and traction to some extent. That’s basically how you decide?

Kevin Costa

There are a few trends. Some of the best founders, over the life cycles of their companies, are choosing fewer partners from inception all the way to IPO. If you’re building a venture product, your product should be about how you can support these founders all the way through. When you find some of the incredible founders, which doesn’t happen too often, you should double, triple, or quadruple down if you have the confidence and the ability to. I think that’s almost part of the product.

Yes, it’s mostly based on these factors, and also the longevity of having spent time with someone and seeing whether they do what they say, and at what level they do it. You can assess from that and be that lifecycle partner.

José Maria Macedo

Yeah.

Kevin Costa

Yeah. Good, yeah.

José Maria Macedo

You keep mentioning, with some disdain, the Silicon Valley founder or Stanford archetype that’s a bit soft. But you did move to Silicon Valley this year. I’m curious why you did that if you think European founders have it harder and are a better source for you. And how has it been since moving there?

5. Founders Beyond Silicon Valley

Kevin Costa

Again, I don’t like to generalize, so I want to be careful there. We can back, and have backed, founders in Silicon Valley. There are a lot of incredible founders that come out of the ecosystem. I think generalizing tends to be wrong. It’s not European founders; it’s founders from all over. We don’t care. We back teams in Africa and anywhere else.

José Maria Macedo

But it’s outsiders, right? Your archetype is, to some extent, outsiders.

Kevin Costa

Yeah, but you can be in the Valley and be an outsider. I don’t know, I’m here. I spend most of my time on the internet here. I don’t go to many of the events, and I’m not super extroverted. I love this one-to-one that we’re having. You can still be an outsider, right?

José Maria Macedo

Yeah.

Kevin Costa

It doesn’t matter where you are geographically. We moved here mostly because most of our companies, while they don’t start here, eventually move here. If we’re supposed to be a concentrated firm that deploys a lot of capital into a few companies, we have to be close to them to get the data points we just talked about. The joke I always used to make is that if all of our founders had moved to South Korea, I would’ve had to move to South Korea. I think that’s true.

José Maria Macedo

Okay, makes sense. Interesting, because from the companies I know, they…

Maybe we don’t want to speak about it on air, but I’m curious which ones have moved there. Has it affected things? I know the biggest positions in your current fund are founders you didn’t find in San Francisco, and potentially wouldn’t have found there. Maybe you found them on the internet, so potentially you would’ve found them either way—

Kevin Costa

Yeah.

José Maria Macedo

—and just done a call with them.

Kevin Costa

Yeah.

José Maria Macedo

But I think you were in London before, right? And—

Kevin Costa

London and New York, yeah.

José Maria Macedo

—I think—for, yeah, London and New York. So I’m curious: do you think it’ll affect your sourcing? Maybe say a little bit about sourcing as well. How are you mainly sourcing these founders?

Kevin Costa

Yeah. You touched on the point. Maybe because of the way I grew up, but I genuinely believe this: whilst everyone here in San Francisco says, “San Francisco is the best ecosystem; you need to be here,” people in London will say, “London is the best ecosystem,” and people in New York will say, “New York is the best ecosystem,” I fundamentally believe the best young technical founders tend to all live on the internet.

If you build your firm so that its headquarters are on the internet, and you live online enough to know where those people are, they’re all on the same street, as I like to say, and you’re able to speak to all of them. Again, that means talking the talk in the same way and living on the internet. That’s what I built, and what we’re building Belief to be: our firm’s headquarters are on the internet, and we’re willing to fly anywhere in the world to partner with the best founders. It might be Australia, it might be Africa, it might be here in the U.S.

I genuinely believe that, and that’s what we’ve done. We don’t care whether companies are here in the U.S. or outside of it; they all tend to be internet-native. That’s where we spend most of our time. To talk about sourcing, I like to talk a lot about asymmetrical flows of information. I think that’s a core pillar people should think about: where do I have an asymmetrical flow of information, and how do I really make use of it?

For us, since a very young age, we’ve been chronically online. I’m still chronically online all the time, and we still think there’s a lot of asymmetrical information that comes from doing that. We meet a lot of people. Many of them don’t end up being founders, but we enjoy talking to all these individuals, all these physics graduates. They’re tinkering—

José Maria Macedo

What’s an example? Where are they? Where are the places you meet them online? Is it Discord servers? Is it X? Are there specific groups you’re in?

Kevin Costa

Yeah, it tends to change over the years. Product Hunt was incredible for a few years. Mostly, it’s X now. I also spend some time reading Nature papers. By reading papers and looking at the individuals who wrote them, you can meet a lot of interesting people.

We also like to spend time in places where we think human capital is mispriced. I always say you have to go where human capital flows are not abundant and where human capital flows are mispriced. In my view, MIT and Stanford might be slightly overpriced, while Imperial, Cambridge, and Oxford are still very much underpriced. Georgia Tech and Vanderbilt are also very much underpriced.

We like to spend time in those ecosystems, where there’s incredible young technical talent, and speak to a lot of people. Most of them don’t end up being investments, but excellence attracts excellence. If you speak to enough great people, and do well by them and help them, things normally come back to you. But you have to play a game that doesn’t feel like work to you. For us, it’s just fun.

José Maria Macedo

That’s cool. All right, maybe let’s talk about some companies, because I think that’ll be interesting too. Let’s start with Ark, which we’re both investors in, for full disclosure. It’s a Ukrainian drone and defense company, right? I think you wrote the first check at the pre-seed and then doubled down as well. They’re doing really well. How did you find them, and what was the process like for getting to conviction?

Kevin Costa

Yeah. Ark is a defense prime, as you mentioned. We met them because one of the founders we’d backed told us, “There’s this interesting guy in Ukraine deploying drones on the front lines,” or, “There’s this weird guy in Ukraine.” Whenever I hear “weird,” my alarm bells normally ring. It’s a very good signal.

José Maria Macedo

The good alarm bells.

Kevin Costa

Yeah, the good alarm bells. I heard that and thought, “Okay, very interesting.” An investor of ours also mentioned, “There’s this guy in Ukraine you should meet.” We took a call, and we were quite starstruck by someone who didn’t fit the Silicon Valley mold. He wasn’t necessarily the most charismatic at pitching, but once you drilled into why, he had so much more substance than anyone else.

I tend to find that our job is to be arbiters of substance; we have to find it. Once you see that substance, you can’t unsee it.

José Maria Macedo

What was the substance there, if you remember? What was the thing—the “why” you asked about—that made you think, “Oh, okay”?

Kevin Costa

For one, the best founders don’t necessarily start companies just because it’s the new cool thing to do. Some of the best founders started companies for larger reasons than that. They didn’t necessarily want to start the company, but they had to. Ark was very similar: people on the front lines were essentially dying, and someone was developing drones, UGVs, and other things, and deploying them on the front lines to help prevent their families and friends from dying.

That’s a very different motivator than, “I want to get into YC and raise a fancy round so I can post about it on LinkedIn.”

José Maria Macedo

Yeah. On the motivation, I totally agree. He’s a very impressive guy. Was there something about the company, or the way he thought about what he was building, that really impressed you? You’ve spoken to a lot of defense startups, I guess, in your time.

Kevin Costa

Yeah. One thing was the asymmetrical information that comes from being on the front lines. No one else is really on the front lines. You have all these large defense primes building products that aren’t really working on the front lines, so there’s a big cognitive dissonance between what they’re building and what the front lines actually need. He had incredible asymmetrical information about what people actually need.

José Maria Macedo

I have a follow-up on that. Why is this so non-obvious right now? It seems like companies in Ukraine are systematically underpriced versus their U.S. counterparts, despite having, as you say, this asymmetric information from being on the front lines and actually having their stuff tested in production. Why are people discounting that?

If you had to steelman it, is it just that the majority of their revenue is in Ukraine, and people are worried that revenue isn’t sticky? If you had to steelman the case for not investing in these companies, or for them being underpriced versus their U.S. counterparts, what’s the reason?

Kevin Costa

There are multiple reasons. One is existential risk from being in those countries. The company might not exist if the war hits them quite hard.

José Maria Macedo

Yeah.

Kevin Costa

Right?

José Maria Macedo

That’s fair.

Kevin Costa

That’s just an existential risk. By the way, it makes building these companies much harder than building a software company in the Valley. That’s why I think these founders have much more resilience and grit.

José Maria Macedo

So that’s one. The second one is access. To some extent, revenues: can you scale revenues outside your main base? And I think with Ark, it is the vision. It’s not that drones are not interesting. My point of view was that defense is interesting, but if you look at a 30-year horizon, this is already happening all over the world, where drones in China deliver your food. It’s not some guy on a bike.

Guess where there are huge labor shortages? Agriculture. Autonomous drones will, at some point, do the crops in places where there’s just no labor. I think he had a very strong vision on that, and not just defense. But if I were to steelman it, a defense company’s revenue is cyclical to some extent and very much tied to one geography. Can they scale outside of that country?

Kevin Costa

Yeah.

José Maria Macedo

And the fifth one is—

Kevin Costa

Literally existential risk.

José Maria Macedo

Yeah. Literally, yeah, exactly. Can they not die?

Kevin Costa

Yeah. Like, actually, though.

José Maria Macedo

Yeah, yeah. Like—

Kevin Costa

Not like—

José Maria Macedo

Yeah.

Kevin Costa

…the company die. Yeah, it’s crazy.

José Maria Macedo

Yeah, yeah.

Kevin Costa

Nice. And I interrupted you while you were going into Ark. I don’t know if you remember what you wanted to say, if there was more you wanted to say about them and why. What got you bullish when you were digging into the whys?

José Maria Macedo

Yeah. I think there are 2 types of ambition. There’s the insular one, where you see something that’s your truth in your environment, and you want to change your environment for the better. I would say that’s Ark’s. And then there’s the very classic sort of ambition, which is nurtured ambition, right? You grow up somewhere, and people around you are founders of X company. Or if you’re in Hollywood, everyone wants to be an actor.

You almost memetically want to do that just because it’s the thing around you. I think it’s very rare to find the first type: you look at your truth and you want to change the future. It’s very insular, and it’s hard. You don’t have any comparison, and you don’t have a cohort of people next to you who have done it before. That’s why maybe companies like, to some extent, even Revolut, were quite insular. Nick was in London. There weren’t too many big companies with a big vision that had done it before.

And he wasn’t surrounded by a lot of people like here. Again, both can be massively successful. But I find that insular type of ambition very striking, and that’s something Ark definitely had and still has.

Kevin Costa

And on Ark, what do you think? How do you think about a founder who—you’re drilling down to the substance, as you said.

José Maria Macedo

Mm-hmm.

Kevin Costa

Right? Do they have the psychological characteristics I’m looking for? But there’s also a way in which whatever the opposite of substance is—I guess the form matters too, right? The way you pitch your vision, whether you can communicate it to investors who aren’t willing to go as deep as you are, has some path dependence in your success, especially in venture.

Ian from Cantos said this: “Actually, a really important thing if you’re doing early stage is to think: will they be able to attract, keep attracting funding? Because even if a company can generate revenue, if you can’t attract funding, you might lose to the company that can.” You’re very much trying to get to the truth, but form does have some path dependence on the company’s outcome. It’s hard to deny that. How do you think about that?

6. Building For The Long Term

José Maria Macedo

I don’t know. I’d say in the short term, maybe. Again, I think that’s very much the times we live in, where we conflate funding attracted with impact. At the end of the day, the only thing that matters is whether the company is around in 30-plus years. If there’s a case to be made that raising a lot of money and being very legible to capital doesn’t lead to the best EV in 20 to 30 years, just look at what’s happening with all the companies that raised a lot of money in 2020.

The Airtables and Miros of this world are being scooped up by Bending Spoons. It’s quite ironic, given Bending Spoons wasn’t the company raising a lot of money in the Silicon Valley mold, but now it’s scooping up a lot of the companies that were in the Silicon Valley mold and raised a lot of money. If we were in 2020, you’d be saying to me, “Man, you need to raise more money if you want to be like Airtable. They all have a higher chance of making it.” But in the end, it doesn’t matter.

Either you build a company that generates free cash flow, which you can then reinvest into the company’s growth, or you can go the super-hard venture route. But again, I tend to think that these short-term goals—being incredibly legible to capital and raising the next round—are very short-term. They’re interesting if you’re looking for a 5X markup or things that don’t matter.

That’s why I think younger investors tend not to do well in venture—or investors in general tend not to do well. Everyone is geared toward fast-twitch capitalism, which is fast dopamine hits. That’s more attuned to being a founder, right? You build a product, sell it to someone, get ARR, get the dopamine hit, and rinse and repeat. I think there are a lot of younger investors, but also investors of all types, in venture, and it’s not slow-twitch capitalism.

It can be if you want to generate the quick 5X markup and look very good. But that’s not the point. The point is slow-twitch capitalism, where it takes 10-plus years for a company to really form and become enduring. It’s not super sexy. In the meantime, you might have some dopamine hits, but not always. This is something we have to keep reminding ourselves: there aren’t that many feedback loops.

The main mistake younger investors make is conflating inputs—activity, number of deals, markups—with output returns. The two aren’t correlated whatsoever. So yes, attracting capital in the short term might give you the ability to hire interesting talent or scale up manufacturing. You need it to some extent, but I think we very much overestimate how much that actually leads to the company having generational EV.

Kevin Costa

Interesting. So next up, do you want to talk about Augusta? You mentioned it as the biotech company, so I’m not sure if you can’t talk about it or if it’s in stealth. Or do you want to talk about Recast? Which one do you think?

José Maria Macedo

Again, we’ve been talking a lot about companies that stick around for 30-plus years. One thing I’ve been thinking quite a lot about is how we keep talking about AGI being here. I think there’s complete cognitive dissonance: you look at a lot of cancers and Alzheimer’s, and we’ve made little progress over the last 30-plus years. It’s still a lot of the same medicine, still a lot of the same efficiency.

Yet we talk about AGI being here, and if you’re not in Anthropic, you’re not going to make it. But if you look at the real world, there hasn’t been much progress. One of the areas I’m most excited about is solving disease and eradicating disease. One of the companies we backed is building personalized cancer vaccines. I fundamentally believe this is something we will fix and solve. It will take a lot of work, but that’s where I think there’s a lot of progress.

Again, groupthink would make you think that if you’re not in one of these large language model companies, you’re not going to make it. Yet there’s so much opportunity. I actually think there’s an abundance of opportunity, but no one wants you to think that, because that’s the strategic move, right? Get everyone to not be agentic with their life so that more power and value accrue to you. It’s sort of the beer strategy, which is fine.

Kevin Costa

Mm-hmm.

José Maria Macedo

And what did you see with Matt from Serova?

Kevin Costa

Yeah. Same thing as with Arc. The goal wasn’t necessarily to build a company. Matt got to see a lot of individuals in the hospital who didn’t have their limbs, mostly due to cancer, and thought to himself, “How is it that we live in 2025 and we talk about AGI, yet cancer still hasn’t been eliminated?”

So he got himself into Oxford, studied medicine, dropped out early, met Michael, his co-founder, and they started Serova.

José Maria Macedo

Who also has an insane story, by the way.

Kevin Costa

Who also—

José Maria Macedo

Co-founder Michael, yeah.

Kevin Costa

Yes, yes. Matt himself has an incredible story, but Michael as well. And again, both are not in the mold. Michael was a cop. But both know that this is the thing they want to do over the next 30-plus years.

They’ve executed incredibly well over the last 12 months, and I think it’s one of the companies you’ll hear a lot about. Using the voting machine versus the weighing machine concept we talked about, I think it has a lot of weight, but still hasn’t been voted on yet. Those are the companies we tend to find interesting.

José Maria Macedo

Nice. I’m curious about Augusta. I think it’s interesting to talk about because we’re both Portuguese, and I think it’s one of the most exciting companies out of Portugal.

Kevin Costa

Mm-hmm.

José Maria Macedo

How did you meet them? And also, it’s an AI consulting company. You wouldn’t think this would be venture-scale. I think that’s the first reaction people have when you pitch them the idea, although it’s becoming more consensus. How did you find them, and how did you think about the investment?

Kevin Costa

Yeah. I found them on X. Rodrigo, the founder of the company, and I bonded over being chronically online. He was a young kid in Portugal who looked up to Daniel Gross, I think. Talk about being chronically online.

That’s how we got to meet. He had just left Sword Health. He was working for V, and it was 2022 or 2023. He said AI was the next big thing, and he wanted to build a company around it. He said everyone was still underestimating how much of an impact this would have over the next 20-plus years.

I sit here today in 2026, and I tend to think it’s still very practical. But if you look at the world around us, I think in 5 years everyone will look back and be like, “Ah, I should have jumped on implementing AI into the economy at large.” Basically, AI is the Ozempic for the economy, right? It makes it fitter—

José Maria Macedo

Mm-hmm.

Kevin Costa

—and leaner and stronger. The question is: How do you administer it into the economy? Which is actually quite hard. No one likes needles. That’s essentially the challenge of what they’re doing.

At the time, he had a very singular motivation. He could have raised capital; a lot of venture capitalists wanted to give him money, and he said no. He said, “I’m going to build this for the next 2 to 3 years as a consulting company where we’re implementing AI.” But the learnings they gathered over those 2 to 3 years have now led to them implementing and executing transformations in some of the largest private equity portfolios globally.

They’re doing it so well that at some point they’ll be able to do much grander things with it, which could be like a Bending Spoons-style approach: acquire an asset and re-flip the economics. But I think that learning path was very important.

They’re a very insular team out of Portugal. The two of us are Portuguese, and there haven’t been too many teams out of Portugal. I’ve been trying for the last 10-plus years. They’ve monopolized some of the best young technical talent in Portugal, and that’s almost one of the advantages of the company. If you’re in these non-obvious regions, you can almost monopolize some of the best talent because you stick out as the one company in the country.

I think Terra has done this in Africa, and I think that is a very special strategy. Then you can scale that to the world. I’m very excited about what they’re doing. Again, it’s still very early days.

José Maria Macedo

Nice. I’m curious, what are your biggest mistakes in investing? Both mistakes of omission—deals that you didn’t do that you think you should have—and mistakes of commission, deals that you did and think were a mistake in hindsight.

7. Learning From Investment Mistakes

Kevin Costa

As a whole, we always focus on the mistakes of omission. They tend to be around preconceived notions that lead us not to take the first call with a founder because we think the space isn’t interesting. Again, it’s the sector-specific mind, which I should always get out of. We all still make mistakes, though, so there have definitely been cases where I should’ve just taken that first call.

José Maria Macedo

Is there a particularly painful one that you’re remembering?

Kevin Costa

I think I had an introduction to the ElevenLabs team at the time.

José Maria Macedo

Oh, wow. That is a painful one.

Kevin Costa

Yeah.

José Maria Macedo

And what about once you’re on the call?

Kevin Costa

I think it got flagged to me, and I was like—

José Maria Macedo

You just didn’t think voice was interesting?

Kevin Costa

No, no, no. I didn’t take the call. It was stupid.

José Maria Macedo

And once you’re on the call, is there one you remember where you misjudged? I’m curious also if there’s a class of mistake that you tend to repeat, like a leak that happens repeatedly.

Kevin Costa

Mm-hmm. Specifics, definitely—

José Maria Macedo

Let me think about it.

Kevin Costa

The whole point around nuance in the beginning is that we tend to overdetermine these frameworks and psychological traits. To some extent, that’s been one of the mistakes we’ve made, and we’re much more attuned to it now. Hopefully we won’t, but as you know, we might still make mistakes as we go.

We look for these traits way too much, when they might not be there. They might not be there on the first call; they might only be there after you’ve met someone for a long time. Overfitting to the criteria or framework we set ourselves has definitely been a mistake sometimes.

José Maria Macedo

Is there a specific one where you think you passed on someone, or a founder who has ended up being great, because you overfit?

Kevin Costa

That’s the thing with venture. Potentially—

José Maria Macedo

Yeah.

Kevin Costa

The thing is, we’re 4 or 5 years in. It’s early. Some of these companies—

José Maria Macedo

Mm.

Kevin Costa

—might look interesting now, especially in a bull market. Everyone is getting marked up, and so it might look like a mistake. Time will tell. I’ll tell you in 5 years, or in 10.

José Maria Macedo

Okay.

Kevin Costa

Yeah.

José Maria Macedo

Okay. And in the mistakes of commission, is there a mistake you think you make? Maybe there’s overfitting there too. You didn’t mention intelligence or clock speed in the things you’re looking for. Maybe it’s obvious that those need to be there. But is there a way you can over-index on grit or disagreement? I’m curious. Maybe there isn’t anything that comes to mind.

Kevin Costa

Yeah, for sure. There are a lot of business models and companies where, especially in the short term, we’re talking about execution-driven businesses. Think about Deel in the beginning: Do you need a lot of horsepower to execute on an execution-heavy business that can scale to hundreds of millions or billions of ARR?

Even with the implementation of AI, now that we’re talking about it, do you need an incredibly high IQ? Probably not. Screening for that in a business that’s just ops-driven might be the wrong move. You should actually look for competitiveness, for traits like just wanting to beat someone early on. That might come through in sports, gaming, or whatever it is that person has done at a young age.

That might be the higher-fidelity answer to whether you should invest or not, versus it has to be Elon Musk, if you see what I mean, from an engineering perspective.

José Maria Macedo

Mm-hmm.

Kevin Costa

We definitely have some of those, yeah.

José Maria Macedo

So you’re saying you tend to look more for founders in industries that don’t require that as much? Because I guess in many of the companies you’ve backed, it actually does require that, right?

Kevin Costa

Yeah.

José Maria Macedo

Like, pretty high—

Kevin Costa

Yeah.

José Maria Macedo

—clock speed.

Kevin Costa

Yeah. We tend to—again, we’ve made a lot of mistakes, and there are companies we should probably be in to some extent. But we tend to think the 30-year companies are more interesting. Time will tell. In the short term, we might look very stupid.

José Maria Macedo

Mm-hmm.

Kevin Costa

If—

José Maria Macedo

Okay.

Kevin Costa

Interesting.

José Maria Macedo

And you do have some companies that are sort of in the Silicon Valley mold, right? I’d say Superpower is very much a darling at this point, no?

Kevin Costa

Yeah, but it wasn’t initially, right?

José Maria Macedo

It wasn’t initially.

Kevin Costa

I mean, Max is an Australian founder. He has a very interesting strategy for acquiring talent. In the beginning, he used a strategy similar to Augusta and to what a lot of our teams do, which is monopolizing some of the best talent from Australia and bringing them over. In a world where everyone here in the Valley leaks a lot of their talent to the next company, bringing over some of the best talent is quite smart from a visa perspective, because you’re actually tied to the company. And second, you just have incredible loyalty because you brought someone from another country to the US to work on this super high-growth company, plus you’ve given them equity. What we’ve seen there is much higher loyalty and trust, which matters a lot. It’s not only about attracting the talent, but keeping the talent.

José Maria Macedo

Mm-hmm.

Kevin Costa

But I’d say it wasn’t that obvious. Definitely, they’re in the mold now. But if you look at their talent, it’s not necessarily only—

José Maria Macedo

Mm-hmm.

Kevin Costa

—talent from here.

José Maria Macedo

Maybe moving on now to some more personal stuff. People have incredibly nice things to say about you. To what extent is that something you try to optimize for, just being super helpful to everyone? And is there a way in which doing that costs you? Do you have to be helpful to founders you’ve passed on? I’m curious if you’ve ever felt that being helpful costs you deals or something else.

8. Playing The Infinite Game

Kevin Costa

Yeah, it’s funny that you mention that. I believe wholeheartedly in it. We’re in the Valley, where everyone talks about paying it forward and playing the infinite game. I was lucky to read a book called Finite and Infinite Games, which you might know, and I think it has shaped a lot of the way we think about conducting life in general, personally, but also while building Belief.

You must play infinite games with a few high-quality people over very long periods of time, and you’ll do fantastically well. Selfishly, on the personal side, there are a few people who really matter, and you go very deep with them. Through that depth with a few people, you actually have more width, ironically. If you want to get to someone or somewhere, because you have depth within your circle, you actually have greater reach.

The counterpoint would be to know everyone, talk to everyone, and be networked with everyone at a very high level. I tend to think you want to optimize for the complete opposite. But what we’re trying to do is sometimes lead first rounds into companies, so we’re not super collaborative. We don’t care who else is in, and we want to take as much of that first round as possible. In that sense, we’re not too collaborative; we tend to be quite aggressive.

But over long periods of time, companies take 30-plus years, and I think it’s shortsighted to think that you can’t introduce two people. People are very shortsighted in life, and specifically in venture. What you want to do is compound over 30-plus years, because that’s how generational companies compound over long periods of time. I think you need to do that from a very young age, and it will come back to you.

We’re early on there, but I think you also need to balance that with what you want to get. There is a balance. One thing I dislike is people being too zero-sum in their thinking, too focused on scarcity. That’s why the future we envision is one of abundance. I don’t know if that answers your question.

José Maria Macedo

Sorry—no, yeah, it does. I have a follow-up, actually. You’re really good at being collaborative once you’ve invested in a company. It seems like you introduce them to a lot of people, including some tier-one investors. I think it’s one of the best things an emerging manager can do. How have you built those relationships? Have you done that consciously, and are you nurturing them? How do you think about relationship-building with some of the firms, like later-stage firms?

Kevin Costa

I’ve been lucky to be exposed to this world since the age of 19.

José Maria Macedo

Okay.

Kevin Costa

So I’ve built a few deep nodes since early on. Those few deep nodes and friendships have led to a lot of width, ironically. Again, it’s not like we’re talking—

José Maria Macedo

Mm-hmm.

Kevin Costa

—to all the GPs of all the firms. As you know, I like to stick to a few people I can go quite deep with.

José Maria Macedo

Mm-hmm.

Kevin Costa

But through that, if we build our relationship in a very good way, I should be able to come to you and say, “Hey, Ze, I really want to get to X person. I’m pretty sure that you’ll pull your weight there.” I think that makes a lot of sense. That’s always how I’ve approached it, and I think over long periods of time, that really compounds.

José Maria Macedo

Nice.

Kevin Costa

So, yeah.

José Maria Macedo

Okay, last question, and then we’ll end it. It’s been really interesting. You talk a lot about infinite games and thinking on 30-year time horizons. In your ideal world, what does Belief look like 5, 10, or 30 years from now? Do you eventually want to be a multi-stage firm, able to lead growth rounds and things like that as well? Do you see yourself doing public markets at some point? I’m curious what the vision is.

Kevin Costa

It’s early days. One of the things is, even if we look at Elon and SpaceX making life multi-planetary, that was a large vision, but it wasn’t the vision from the get-go. The vision from the beginning was, “We’re going to shoot up a rocket into space, take a picture, and attract more funding to NASA.” Big things start small.

Having said that, I do believe there’s a new product that founders want, from inception all the way to IPO. You ask, “Do you want to be a multi-stage firm?” We don’t care about any of these labels. We reject any label. We just want to be the best investment firm there has ever been, and I think there’s an opportunity to do that.

If you look at the last 50 years and what Berkshire did, it’s incredible. Now, as tech is becoming a larger percentage of GDP, I think there’s an opportunity to become the Berkshire for the innovation economy. I think that means doing early stage really, really well, but also leveraging some of the asymmetric information you get to deploy at other stages, with the ultimate goal being returns.

That’s essentially the scoreboard, while helping some of the best founders and supporting them in a different way that no one else has. We’re doing this for the next 50-plus years. We have a very long arc of ambition.

But it’s still early days in the journey. I tend to reject labels like “multi-stage firms” or “funds.” When people say, “You’re building a cool fund” or “a solo GP fund,” it’s like, no, that’s wrong. We’re just trying to build the best investment firm ever.

José Maria Macedo

Very cool, man. Thank you so much for joining. I think people are going to really enjoy this. And yeah, I really enjoyed the chat, and I’m excited to keep investing in Belief and seeing you crush it over the years.

Kevin Costa

Exciting. Thanks for the time, and yeah, see you soon.