The Twenty Minute VCThe AI Boom Will Create Enormous Roadkill: Who Wins & Loses? | David Frankel
CHAPTERS
- 0:00 – 2:03
Seed investing’s “pyramid” and why missing mega-winners makes funds unsellable
David explains how venture has narrowed into a pyramid where access to a tiny set of massive outcomes drives fundraising success. He frames seed as still viable, but increasingly crowded and commoditized, with too many “unreasonable bets” driven by fund-raising incentives.
- •Venture as asset-management + access-selling ecosystem (Cambridge Associates, fund-of-funds)
- •Why being outside the “top names” makes it harder to raise capital
- •Outcomes are extremely skewed; missing the biggest companies changes fund math
- •Seed is tougher because it’s faster, noisier, and more commoditized
- •Seed remains compelling when you can wait for truly exceptional founders/teams
- 2:03 – 7:42
The $50–$100M seed fund trap and the role of the “insurance-policy” micro-check
They debate Harry’s thesis that $50–$100M seed funds are stuck: too big to be collaborative micro-check investors, too small to lead large seeds. David describes how Founder Collective often participates as a credible, patient backer—sometimes viewed by founders as an insurance policy against large funds abandoning them.
- •Why mid-sized seed funds can struggle to define a clear role in rounds
- •Founder Collective often writes $500k–$1M alongside $8–$9M rounds
- •Founders may value FC for patience, brand, and network when others move on
- •Not all seeds are huge; FC still finds $3–$4M rounds where price moves more
- •Hot AI rounds are not yet proven capital-efficient
- 7:42 – 9:15
“Everyone’s AI now”: funding non-hype startups and off-piste opportunities
Harry challenges whether non-AI companies can get financed; David argues the distinction is collapsing—AI is now a baseline tool rather than a category. He describes backing domain-expert founders using modern AI tools to rework real pain points, often in overlooked markets with massive TAMs.
- •AI as default tooling, analogous to ‘using the internet’
- •Examples: domain experts using Claude/modern tooling to rebuild broken workflows
- •Off-piste sectors can still hide enormous markets
- •Financing risk matters, but ‘non-AI’ is increasingly a false label
- •Value can come from approach + edge, not hype category
- 9:15 – 10:21
Price matters (still): uncapped notes, access-selling, and seed math reality
They confront the claim that price doesn’t matter if you’re in winners. David pushes back: entry price drives required outcome size, and uncapped notes often distort seed economics as “access” is sold to oversubscribed rounds.
- •Outcome scale required is a direct function of entry valuation (pure math)
- •Uncapped notes ‘suck’ at seed; can become $100–$300M priced later
- •Elite access is sold (universities/brands), not always rationally evaluated
- •Saying no to overpriced access deals is necessary—even if you miss some winners
- •Seed investors must balance relationship value vs financial payoff
- 10:21 – 12:27
Startup founding is normalized—entrepreneurship (fortitude) is not
David worries that startup culture has made “being a founder” fashionable, producing many founders but fewer true entrepreneurs. He distinguishes founders from entrepreneurs by fortitude, learning speed, and the ability to recruit and lead through scaling.
- •Normalization of founding (YC-style) increases supply of ‘founders’
- •Entrepreneurship requires fortitude, energy, and steep learning-curve capacity
- •CEO work becomes recruiting-heavy (‘bums on seats’), echoed by Bezos anecdote
- •CEO vs CTO trajectories diverge as companies scale beyond early headcount
- •Many founders underestimate what the CEO journey actually demands
- 12:27 – 14:26
CEO–CTO alchemy: when a weak cofounder is a dealbreaker
Harry argues cofounders often change; David says he rarely invests when the founding package lacks the CEO–CTO magic. He looks for a sales-driven CEO and a “magician” CTO, emphasizing trust, alignment, and rare cofounder alchemy as key early predictors.
- •David’s preferred combo: sales-strong CEO + ‘magician’ CTO
- •CTO roles can become fungible later, but early cofounder dynamics matter
- •He’s said no more than yes when one cofounder seems subpar (with regrets)
- •He evaluates trust, alignment, and how founders validate each other’s competence
- •True cofounder alchemy is rare but disproportionately powerful
- 14:26 – 15:53
Has the ideal founder changed in the AI era? Less than you’d think
Harry notes a shift toward DeepMind/Gemini-style technical founders; David says his pattern recognition hasn’t changed much. He still prioritizes psychographic traits—youthful intensity, focus, and energy—over resume prestige, while acknowledging experience can help in certain SaaS contexts.
- •Founder ‘psychographics’ (focus/intensity) matter more than pedigree
- •Experience helps in some enterprise/SaaS settings, but energy remains decisive
- •Early-stage journeys demand enormous stamina; intensity is ‘intoxicating’
- •Not ageist: intensity can persist for decades, but is common in early 20s founders
- •AI changes tools, but not the core founder attributes he bets on
- 15:53 – 19:18
Is 1.5→5→15M ARR still venture-scale? Patience, “orphans,” and hidden traction
Harry questions whether classic growth paths are still exciting; David argues the market’s impatience creates opportunity. He emphasizes that long timelines are normal, that traction isn’t only revenue, and that abandoned ‘seed extension’ situations can become attractive capital-market moments.
- •10-year funds often take ~18 years; patience is core to venture reality
- •SeatGeek as a long-duration winner (investment since 2010)
- •Revenue isn’t the only metric: retention, DAUs, expansion are story drivers
- •Impatience creates ‘orphans’—seed extensions can be opportunistic entries
- •Execution quality matters more than rigid growth heuristics
- 19:18 – 21:57
AI boom roadkill and seed’s survival: outcomes are huge but failure is the default
David predicts another crash is inevitable, even as AI becomes the wave of a lifetime. He explains why seed still works: you don’t need to catch only trillion-dollar outcomes—owning meaningful stakes in $500M–$2.6B outcomes can still return seed funds.
- •Another dot-com-like crash: ‘definitely’; timing unknown
- •AI is transformative but Hollywood-like: most companies won’t survive
- •Extreme skew: few $100B+ outcomes vs many mid-sized winners
- •Seed isn’t dead; it’s crowded/commoditized, but still fund-returning at right ownership
- •Brand and portfolio references now often just ‘get you to the table’
- 21:57 – 28:52
Mega-platform funds: why they can fail entrepreneurs (orphan risk) and why FC stays small
They debate whether mega-platforms win as price separators. David argues large funds often buy call options, staffed by junior champions who may leave—creating orphaned companies when follow-on support disappears; he explains why Founder Collective prioritizes DPI and alignment over management-fee scaling.
- •Mega funds can orphan companies when the internal champion departs
- •95% of companies may not fit a mega-fund’s follow-on mandate
- •Entrepreneurs believe they’re the exception; statistics are brutal
- •Founder Collective’s GP as biggest LP aligns incentives toward returns/DPI
- •Raising larger funds is tempting but conflicts with their strategy and discipline
- 28:52 – 42:08
Frameworks vs momentum: pro rata as ‘original sin,’ preemptive rounds, and ownership discipline
Harry pushes that $1B valuations are ‘new Series A’; David counters that this is a momentum game and not his core business. They explore pro rata rights, the difficulty of concentrating in fast preemptive rounds, and why FC uses frameworks to act quickly and avoid style drift.
- •David distinguishes value-seeking early entry from momentum investing
- •Pro rata viewed as a founder-unfriendly call option; uneven rights are worsening
- •Preemptive rounds compress decision cycles and complicate concentration
- •Frameworks set limits (post-money thresholds) and enable fast action
- •Following every up-round with tier-1 leads may boost absolute return, not multiples
- 42:08 – 45:31
SaaS correction: are embedded SaaS companies being oversold in the ‘SaaSpocalypse’?
Harry raises cannibalization risk (Airtable/Snyk-style pressure) and faster innovation cycles. David argues the market may be overcorrecting, especially for deeply embedded, mission-critical systems where switching costs and operational dependence create resilience.
- •Innovation cycles are faster, threatening category leaders before liquidity
- •Market-cap erosion may be ‘throwing the baby out with the bath water’
- •The last 5% of embedment (mission-critical dependence) is underestimated
- •Less-embedded SaaS is easier to replace with AI tooling
- •Contrarian basket: beaten-down top SaaS stocks could rebound
- 45:31 – 49:13
Decision heuristics: the “I love it because…” test and the ‘nepo baby’ edge redefined
David shares FC’s internal rule: if you can’t complete “I love it because…,” you shouldn’t invest—anchoring decisions in conviction, not valuation. He also explains his ‘nepo baby’ thesis as vertical-native founders who grew up inside an industry and therefore carry compounding, unfair advantage.
- •Investment meeting opener: ‘I love it because…’ must be finishable
- •Founder obsession and clarity beat valuation-first thinking
- •Examples of ‘I love it because’: insight, edge, acquisition ability, founder honesty
- •‘Nepo baby’ means vertical-native operators (not trust-fund kids)
- •Industry-embedded upbringing creates durable, underappreciated edge
- 49:13 – 1:13:54
Secondaries, DPI, and dilution: managing liquidity in a more tradable private market
They discuss how unusually liquid secondaries have become, especially in top names, enabling more precise pricing and earlier DPI. David advocates occasional sell-downs to return capital while staying long, and reflects on dilution differences between fast AI journeys and long hardware paths.
- •Secondary markets are highly liquid in top private names; discounts often minimal
- •Secondaries can reveal near-term momentum (boards planning next rounds)
- •Taking ~20% off can meaningfully improve DPI while retaining upside
- •Uber sell-down reflection: likely sold a bit early but remained net long to IPO
- •Dilution: faster AI winners dilute less; long hardware journeys raise more capital
- 1:13:54 – 1:18:33
Geopolitics, China, defense, and photonic computing: the next disruptions beyond today’s models
The conversation widens to AI’s macro impact: US vs China as dual superpowers, defense urgency, and the inevitability of platform displacement (even for OpenAI/Anthropic). David spotlights photonic computing as a potential step-change that could disrupt Nvidia or be acquired, while warning US R&D cuts risk ceding advantage.
- •China and the US as the two AI superpowers; defense importance is rising
- •Even dominant AI platforms will be disrupted; China is a plausible source
- •Regulatory asymmetry: China moves faster; Europe is slowest; US is in-between
- •Photonic computing as an energy-efficient compute inflection (optical chips)
- •US competitiveness depends on sustained R&D (DARPA/universities), at risk from cuts
- 1:18:33 – 1:30:47
What would justify a bigger fund, plus rapid-fire lessons: mistakes, marriage, and what’s next
David says he’d raise a larger fund only if a new arbitrage appeared—e.g., orphaned Series A/B/C value opportunities—because seed’s risk premium has largely been competed away. In quick-fire, he covers surprises about AI adoption, controversial deals, framework mistakes (Klaviyo miss), relationship advice, and optimism about health and autonomy breakthroughs.
- •Bigger fund trigger: a clear value dislocation in later stages due to abandonment
- •Seed-stage arbitrage has narrowed as the market crowded in
- •AI surprise: less impact than expected in some consumer areas; voice/UI improving
- •Frameworks both save you and can cause misses (e.g., valuation shorthand)
- •Personal principles: presence/kindness in relationships; excitement about health advances and autonomy
