Suno valued at $5 billion Suno, the AI music company backed by Founder Collective at seed, is now worth $5 billion.
David Frankel says another dot-com crash is inevitable, OpenAI and Anthropic will be disrupted — and China could be the one to do it.
The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
David Frankel says another dot-com crash is inevitable, OpenAI and Anthropic will be disrupted — and China could be the one to do it.
TL;DR
David Frankel, co-founder of Founder Collective, delivers a candid masterclass on seed investing — why the $50M–$100M seed funds will be the worst performers [1] — Harry Stebbings "$50–$100M seed funds worst performers: Frankel agrees that the worst-performing funds of this vintage will be the $50M–$100M seed funds — t…" 05:21 , why pro rata is "the original sin" [2] — Harry Stebbings "I think a billion dollar valuation's the new Series A. You enter at a billion and you hope it becomes 20." 32:45 , and why a dot-com-style crash is inevitable [3] — David Frankel "I think Microsoft have done a crappy job of AI generally." 58:58 . With 18 years of backing Uber, Coupang, Suno, and Shield AI while keeping funds deliberately small, Frankel argues that owning 5% of a $2.6B outcome still returns a seed fund — and that the best opportunities are always found off the beaten path, not in the hottest rounds.
David Frankel, co-founder of Founder Collective, joins Harry Stebbings to discuss seed investing discipline, AI market dynamics, fund sizing, secondary markets, and long-range technology predictions including autonomous vehicles and AI-driven healthcare.
The episode opens cold with David Frankel's most provocative lines — on growing bubbles, inevitable roadkill, and the liquidity of secondary markets — before Harry Stebbings introduces him as one of the great seed investors of our time. Harry makes the case for why Frankel is exceptional: unlike nearly every peer, he resisted scaling funds, kept the boutique discipline alive, and successfully navigated from the pre-AI era (Coupang, Uber, PillPack, SeatGeek) into the AI wave (Shield AI, Suno at $5B). The intro establishes the episode's central tension: small and disciplined versus big and resourced. Three sponsor spots follow — Fireworks AI (intelligent model routing for production AI), Asana (AI-native operating system for human-agent teams), and Superhuman (AI email assistant) — before the conversation begins properly.
An honest and revealing section where Frankel admits what Founder Collective has become in this market: an insurance policy. He's seen it clearly in the last 20 deals — founders take the multi-stage fund's $8–9M but keep FC in at $500K or $1M, knowing their champion at the big fund might leave, might lose mandate, or might simply move on to the next hot company. FC's brand and reputation as patient capital is itself a product. Frankel also issues a clear-eyed warning: there is very little evidence yet that the hot AI companies raising massive rounds are capital efficient — if anything, they're anything but. The discussion then moves to valuations: uncapped notes are economically bad for investors but Frankel has written one because he loved the founders. He also raises the normalization of startup founding through YC and questions whether there's a meaningful difference between 'founders' and 'entrepreneurs' — arguing the latter is a rarer, more demanding breed.
With Benchmark, A16Z, and even the most disciplined growth funds raising billions, Harry asks directly: is Founder Collective really not tempted? Frankel admits the tension is real and recurring. But the answer always comes back to the same fact: the GP is the largest LP. No external investor has more skin in the game than the partners themselves. When you're aligned that way, the calculus on fund size becomes entirely about return multiples, not management fees. Frankel also notes that the firm has been deeply disciplined about DPI — how much real cash comes back to investors — and that growing the fund size makes that discipline structurally harder. He also reflects honestly on what they've left on the table: had they followed on in Uber, Coupang, or Suno at Series A, the absolute dollar returns would be higher, though he doubts the fund multiple would be better.
David Frankel makes his most provocative structural argument: pro rata rights are bad for entrepreneurs. They're a call option against the founder, not for investors. Founder Collective has never led a follow-on round in its entire history — and when pro rata is offered only to the lead, Frankel questions whether that should be the norm. The conversation then pivots to investment frameworks: Peter Thiel once said simply following every up-round from good brands would have been the best strategy in the golden era. Frankel concedes the data probably supports that retrospectively. Harry then drops his most controversial line of the episode: a $1 billion valuation is the new Series A. He cites Cursor at $60B, Cognition at $26B — if you're entering at $1B and underwriting to $20B, that's just the new math. Frankel gently pushes back, arguing that's the momentum game, not his game.
Harry asks how concentrated Founder Collective's returns actually are — referencing Ho Nam's Altos Ventures where Roblox was the single overwhelming driver. Frankel's answer is surprising: less concentrated than you'd expect. Fund 2 has Vacata, Shield, Whoop, and PillPack all as major contributors. Fund 1 has the Trade Desks, Ubers, and Coupangs, but also Airtable, Simply, and SeatGeek. He then introduces one of his firm's most distinctive practices: every investment pitch in a team meeting must begin with 'I love it because…' If you can't finish that sentence compellingly, you don't invest. And critically, you cannot finish it with valuation — it must be founders, opportunity, or edge. His best recent answer? 'I love it because every question I ask, I get a better answer than I expected, and they're never evasive.' He also reveals his current thesis on 'nepo babies': founders who spent their formative years inside the vertical they're now disrupting, using TJ Parker's pharmacy childhood and Suno's Kensho roots as examples.
Harry raises a pointed question about founder loyalty and focus: today's founders often have angel portfolios as large as small VC funds, run side projects, and leave companies in 12–18 months. Is the breed changing? Frankel acknowledges the evidence at the margins — he's been dazzled by founder brand power that didn't translate to staying power — but argues the vast minority of founders actually abandon ship. His most interesting observation: second-time founders who had a life-changing but not enormous first outcome are often the best bets. They're hungry, they've learned lessons, they have one or two loyal team members ready to follow them, and they're in a hurry. By contrast, founders who had genuinely great outcomes and 'go again' often get bored when the second company isn't big enough fast enough.
Harry raises the job displacement question. Frankel's position: mass unemployment is not coming, but a massive productivity divide is inevitable. The haves and have-nots used to be about who had data; now it's about who can use the tools. Young people tinkering with AI in dorm rooms have a structural advantage because they're mentally plastic to the technology. Simon and Claire at 45, with decades of accounting experience, face a harder path. Their only real edge is vertical knowledge — they can sell to people who look like them and trust them, and in high-stakes service situations (litigation, insurance, auditing), the human interface still has value. Frankel sees the greatest retraining opportunity in lower-cost global environments, and is broadly optimistic that productivity gains will outweigh displacement. The TAM expansion play in services — suddenly being able to afford a lawyer for things you couldn't before — is a genuine economic benefit.
Harry asks the hypothetical: what would cause Frankel to raise a larger fund? His answer is specific: if he saw a genuine arbitrage at Series A or B — companies growing steadily but being abandoned by momentum investors, trading at unfair discounts — that would pull him upstream. He explicitly says it's not about momentum, but about perceived value dislocation. He then surfaces one of the most honest moments of the episode: Founder Collective was introduced to Klaviyo early, through the same person who sent them Suno. He loved the founders but said no because the valuation didn't fit his framework. Klaviyo went on to IPO at a multi-billion dollar valuation. The framework saved them from many bad deals — but it also cost them Klaviyo. Frankel's verdict: frameworks are imperfect tools, and the best investors know when to override them for the truly extraordinary.
The quickfire round opens with a genuinely surprising admission from Frankel: AI should have impacted consumer applications far more by now, and he's been disappointed by how little has changed in consumer AI relative to the hype. He types less and speaks more, but the category hasn't been fully played out. He then reveals what investors at Suno's $5B valuation are actually underwriting: a Spotify and Apple Music disruptor thesis — moving from a creation tool to a consumption platform, which is why Jack from Snap was brought in as CEO. The most memorable moment is Suno CTO Martin Camacho's answer when asked whether he'd swap out his own model for a better one: 'Wouldn't think twice about it.' The product is the interface and experience. How it gets built is irrelevant to the user.
Harry asks what Frankel is most excited about in the next 10 years. The answer is deeply personal and wide-ranging: autonomous vehicles — slow, slow, slow, then overnight — will mean the kids of today never need a driver's license. And AI-driven medical compute will produce treatments that make current chemotherapy look prehistoric. He names the friends he's lost to cancer and the conviction that something much better is coming. Harry, who has no driver's license, takes personal delight in the prediction. The two exchange warm closing remarks, reflecting on 11 years of friendship and mutual admiration, before the episode ends with a second full run of the three sponsor reads: Fireworks AI (intelligent model routing), Asana (AI-native workflow for human-agent teams), and Superhuman (AI email assistant).
Chapter 1 · 00:00
The episode opens cold with David Frankel's most provocative lines — on growing bubbles, inevitable roadkill, and the liquidity of secondary markets — before Harry Stebbings introduces him as one of the great seed investors of our time. Harry makes the case for why Frankel is exceptional: unlike nearly every peer, he resisted scaling funds, kept the boutique discipline alive, and successfully navigated from the pre-AI era (Coupang, Uber, PillPack, SeatGeek) into the AI wave (Shield AI, Suno at $5B). The intro establishes the episode's central tension: small and disciplined versus big and resourced. Three sponsor spots follow — Fireworks AI (intelligent model routing for production AI), Asana (AI-native operating system for human-agent teams), and Superhuman (AI email assistant) — before the conversation begins properly.
Suno, the AI music company backed by Founder Collective at seed, is now worth $5 billion.
The $50M–$100M seed funds are caught in a brutal no-man's land: too large to write the friendly $100–$250K collaborative checks, and too small to lead a competitive $8–$10M seed round. David Frankel agrees with Harry's thesis and argues these funds will be the worst performers of this vintage.
Chapter 2 · 05:21
An honest and revealing section where Frankel admits what Founder Collective has become in this market: an insurance policy. He's seen it clearly in the last 20 deals — founders take the multi-stage fund's $8–9M but keep FC in at $500K or $1M, knowing their champion at the big fund might leave, might lose mandate, or might simply move on to the next hot company. FC's brand and reputation as patient capital is itself a product. Frankel also issues a clear-eyed warning: there is very little evidence yet that the hot AI companies raising massive rounds are capital efficient — if anything, they're anything but. The discussion then moves to valuations: uncapped notes are economically bad for investors but Frankel has written one because he loved the founders. He also raises the normalization of startup founding through YC and questions whether there's a meaningful difference between 'founders' and 'entrepreneurs' — arguing the latter is a rarer, more demanding breed.
Frankel agrees that the worst-performing funds of this vintage will be the $50M–$100M seed funds — too big to be a collaborative friend, too small to lead an $8–$10M seed round.
Forget trillion-dollar outcomes. Founder Collective's internal analysis shows the median valuation among the top 500 companies created in the last 25 years is $2.6 billion. Own 5% of one of those, and you've returned your fund. Seed isn't dead — it's just math.
Of all companies created in the last 25 years, fewer than 100 have sustainably maintained a valuation over $10 billion.
Founder Collective's analysis of the top 500 companies created in the last 25 years shows a median valuation of $2.6 billion — meaning 5% ownership returns a seed fund.
Smart founders are increasingly taking $8–10M from multi-stage funds while quietly keeping Founder Collective in at $500K–$1M. They know the big fund's junior associate might leave, and FC is patient capital that won't orphan them. It costs very little to have a real insurance policy.
Mikey Schulman, CEO of Suno, told Frankel he spends 30–40% of his time on recruiting — echoing Jeff Bezos's famous claim that he spent 50% of his time on hiring.
Frankel is looking for one specific combination: a CEO who is a great salesperson and a CTO who is a genuine magician. The alchemy between them — not identical, not finishing each other's sentences, but deeply aligned — is what separates fundable companies from great companies. In 18 years, he's seen it five times.
Every technology wave has produced bigger bubbles than the last — internet, SaaS, mobile, AI. The AI wave is the biggest yet, and Frankel has no doubt there will be enormous roadkill. But the survivors — OpenAI, Anthropic, SpaceX — will be the Googles and Metas of this era.
The GP is the largest LP in every Founder Collective fund. That single fact explains why they've never raised a growth vehicle. When you eat your own cooking, optimising for management fees becomes impossible — you only care about DPI.
Chapter 3 · 25:06
With Benchmark, A16Z, and even the most disciplined growth funds raising billions, Harry asks directly: is Founder Collective really not tempted? Frankel admits the tension is real and recurring. But the answer always comes back to the same fact: the GP is the largest LP. No external investor has more skin in the game than the partners themselves. When you're aligned that way, the calculus on fund size becomes entirely about return multiples, not management fees. Frankel also notes that the firm has been deeply disciplined about DPI — how much real cash comes back to investors — and that growing the fund size makes that discipline structurally harder. He also reflects honestly on what they've left on the table: had they followed on in Uber, Coupang, or Suno at Series A, the absolute dollar returns would be higher, though he doubts the fund multiple would be better.
In recent Founder Collective funds, the GP (David Frankel and partners) are the single largest LP — more than any external investor — ensuring deep alignment with entrepreneurs.
Pro rata rights sound like a founder-friendly perk, but Frankel calls it 'the original sin.' It's a call option against the entrepreneur — not for them. Founder Collective has always had to work to earn the right to put in more money, and they've never led a follow-on round in their entire history.
Chapter 4 · 30:10
David Frankel makes his most provocative structural argument: pro rata rights are bad for entrepreneurs. They're a call option against the founder, not for investors. Founder Collective has never led a follow-on round in its entire history — and when pro rata is offered only to the lead, Frankel questions whether that should be the norm. The conversation then pivots to investment frameworks: Peter Thiel once said simply following every up-round from good brands would have been the best strategy in the golden era. Frankel concedes the data probably supports that retrospectively. Harry then drops his most controversial line of the episode: a $1 billion valuation is the new Series A. He cites Cursor at $60B, Cognition at $26B — if you're entering at $1B and underwriting to $20B, that's just the new math. Frankel gently pushes back, arguing that's the momentum game, not his game.
Twenty times a $50M post used to be the dream. Now you're entering at $1B and hoping for $20B. Cursor sold for $60B. Cognition is at $26B. The math has shifted so dramatically that refusing to enter at unicorn valuations might mean missing the entire category. A billion is the new Series A.
Chapter 5 · 32:59
Harry asks how concentrated Founder Collective's returns actually are — referencing Ho Nam's Altos Ventures where Roblox was the single overwhelming driver. Frankel's answer is surprising: less concentrated than you'd expect. Fund 2 has Vacata, Shield, Whoop, and PillPack all as major contributors. Fund 1 has the Trade Desks, Ubers, and Coupangs, but also Airtable, Simply, and SeatGeek. He then introduces one of his firm's most distinctive practices: every investment pitch in a team meeting must begin with 'I love it because…' If you can't finish that sentence compellingly, you don't invest. And critically, you cannot finish it with valuation — it must be founders, opportunity, or edge. His best recent answer? 'I love it because every question I ask, I get a better answer than I expected, and they're never evasive.' He also reveals his current thesis on 'nepo babies': founders who spent their formative years inside the vertical they're now disrupting, using TJ Parker's pharmacy childhood and Suno's Kensho roots as examples.
Shield AI was a defense drone company in 2016 — controversial, unfashionable, and misunderstood. Today it's one of the most important defense technology companies in the world. Frankel's core thesis: the best investments are never in the hot theme of the moment. Applied AI in 2016 was off-piste. The job is to be there 5–10 years ahead.
Founder Collective has held its SeatGeek position since the 2010 investment and has never sold a single share, calling it 'spiritual' at this point.
Founder Collective will not invest in any company unless they believe there is a credible path to a 10x return — a non-negotiable framework used to open every team investment meeting.
Forget trust-fund kids — Frankel's 'nepo babies' are founders who grew up inside the vertical they're now disrupting. TJ Parker worked in his dad's pharmacy at 14. The Suno founders built their entire careers on audio AI at Kensho. Evan at Rebar watched his uncle's HVAC business get rolled up by PE. That embedded knowledge is an edge no amount of funding can buy.
Top-50 AI company secondary positions are trading at or above last-round price — sometimes at a premium when insiders know a new round is coming. Frankel recommends taking 20% off the table even in your winners: returning 25% of a fund early is almost always better than waiting for a potential double.
Frankel says he has never seen secondary markets as liquid as they are now, with top-50 names trading at or above last-round price — sometimes at a premium when insiders know another round is coming.
Chapter 6 · 46:39
Harry raises a pointed question about founder loyalty and focus: today's founders often have angel portfolios as large as small VC funds, run side projects, and leave companies in 12–18 months. Is the breed changing? Frankel acknowledges the evidence at the margins — he's been dazzled by founder brand power that didn't translate to staying power — but argues the vast minority of founders actually abandon ship. His most interesting observation: second-time founders who had a life-changing but not enormous first outcome are often the best bets. They're hungry, they've learned lessons, they have one or two loyal team members ready to follow them, and they're in a hurry. By contrast, founders who had genuinely great outcomes and 'go again' often get bored when the second company isn't big enough fast enough.
Chapter 7 · 58:03
Harry raises the job displacement question. Frankel's position: mass unemployment is not coming, but a massive productivity divide is inevitable. The haves and have-nots used to be about who had data; now it's about who can use the tools. Young people tinkering with AI in dorm rooms have a structural advantage because they're mentally plastic to the technology. Simon and Claire at 45, with decades of accounting experience, face a harder path. Their only real edge is vertical knowledge — they can sell to people who look like them and trust them, and in high-stakes service situations (litigation, insurance, auditing), the human interface still has value. Frankel sees the greatest retraining opportunity in lower-cost global environments, and is broadly optimistic that productivity gains will outweigh displacement. The TAM expansion play in services — suddenly being able to afford a lawyer for things you couldn't before — is a genuine economic benefit.
Conventional wisdom says Microsoft's OpenAI investment put them at the front of the AI race. Frankel disagrees sharply. Google is actually in pole position — they come from an AI-native background, and search with context is where AI creates the most value. Microsoft's AI feels second-rate. Recovery is unclear.
Frankel argues Microsoft has done a 'crappy job' of AI generally, and that their AI products feel second-rate compared to the top 3 or 4 players, making recovery unclear.
Frankel says another dot-com-style crash is definitively coming — 'if' is not a question, only 'when' is unknown — driven by the massive capital and valuation inflation in AI.
In the 1820s, China accounted for 25% of global economic output, making it the world's largest economic machine — a historical precedent for its current AI superpower ambitions.
Every dominant technology platform has eventually been displaced — mainframes, Microsoft, Google, now OpenAI and Anthropic. Frankel says their disruption is 'unequivocal.' The most likely source? China. And the speed of the innovation cycle means they may not even have time to establish their incumbency first.
Chapter 8 · 1:08:27 Harry asks the hypothetical: what would cause Frankel to raise a larger fund? His answer is specific: if he saw a genuine arbitrage at Series A or B — companies growing steadily but being abandoned by momentum investors, trading at unfair discounts — that would pull him upstream. He explicitly says it's not about momentum, but about perceived value dislocation. He then surfaces one of the most honest moments of the episode: Founder Collective was introduced to Klaviyo early, through the same person who sent them Suno. He loved the founders but said no because the valuation didn't fit his framework. Klaviyo went on to IPO at a multi-billion dollar valuation. The framework saved them from many bad deals — but it also cost them Klaviyo. Frankel's verdict: frameworks are imperfect tools, and the best investors know when to override them for the truly extraordinary.
Frankel says it is 'unequivocal' that OpenAI and Anthropic will be disrupted, and that China has an excellent chance of being the disruptor.
Nvidia looks invincible today — just as Intel once did. Frankel believes photonic computing, chips using photons instead of electrical signals, will be the disruptor. Optical fiber already handles all connectivity in data centers. The last frontier is the chip itself. When it arrives, the energy consumption argument against AI collapses.
Frankel predicts photonic (optical) computing — using photons instead of electrons — will be the key disruptor to Nvidia's dominance in AI chips, with major energy efficiency implications for data centers.
Olo, backed by Founder Collective from the very beginning with Noah Glass, was taken private by Thoma Bravo at approximately a $2 billion valuation after a 17-year journey.
Chapter 9 · 1:16:22 The quickfire round opens with a genuinely surprising admission from Frankel: AI should have impacted consumer applications far more by now, and he's been disappointed by how little has changed in consumer AI relative to the hype. He types less and speaks more, but the category hasn't been fully played out. He then reveals what investors at Suno's $5B valuation are actually underwriting: a Spotify and Apple Music disruptor thesis — moving from a creation tool to a consumption platform, which is why Jack from Snap was brought in as CEO. The most memorable moment is Suno CTO Martin Camacho's answer when asked whether he'd swap out his own model for a better one: 'Wouldn't think twice about it.' The product is the interface and experience. How it gets built is irrelevant to the user.
Chapter 10 · 1:19:56 Harry asks what Frankel is most excited about in the next 10 years. The answer is deeply personal and wide-ranging: autonomous vehicles — slow, slow, slow, then overnight — will mean the kids of today never need a driver's license. And AI-driven medical compute will produce treatments that make current chemotherapy look prehistoric. He names the friends he's lost to cancer and the conviction that something much better is coming. Harry, who has no driver's license, takes personal delight in the prediction. The two exchange warm closing remarks, reflecting on 11 years of friendship and mutual admiration, before the episode ends with a second full run of the three sponsor reads: Fireworks AI (intelligent model routing), Asana (AI-native workflow for human-agent teams), and Superhuman (AI email assistant).
The driverless car promise that seemed '5 years away' for 20 years is finally tipping. Frankel believes the children of today's founders will never need a driver's license. And with AI accelerating medical compute, treatments like chemotherapy will look as barbaric to future generations as bloodletting looks to us.
No indexed bits in this chapter.
Sign in to keep viewing
Create a free account to keep exploring this episode's insights, snapshots, quotes and claims.
We scan show notes for social handles, websites and apps. Nothing matched on this episode.
We use essential and analytics cookies to run Vuci. To understand how the site is used:
Privacy Policy. Install Vuci on your phone
Add it to your home screen for a faster, app-like experience.
Install Vuci on your phone
Tap the Share button, then “Add to Home Screen”.
A new version is available
Reload to get the latest Vuci.