David Frankel, Head of a Veteran Seed Fund, Deep Dive Interview: The Elimination Wave Behind the AI Frenzy, Valuation Discipline, and Exit Strategies

"The AI Boom Will Create Enormous Roadkill: Who Wins & Loses? | David Frankel" (20VC with Harry Stebbings interview), here are the key points summarized. 1. AI Frenzy and the Inevitable Crash (Roadkill & Crash) • An adjustment is certain: This wave of AI is the biggest technological change experienced by this generation (surpassing the internet, SaaS, and mobile), but it will inevitably be accompanied by significant overheating and bubbles. • High elimination rate (Roadkill): Looking back over the past 25 years, there have been fewer than 100 sustainable companies in the U.S. with a market value over $10 billion. The vast majority of AI startups currently popular will become "roadkill" in the next 5-10 years, with over 95% failing to meet expectations. • Seed survival philosophy: Seed investors do not need to hit all the giants. Since the median market value of top companies is around $2.6 billion, holding 5% equity in quality companies early on is sufficient to return the fund, even if the company exits for hundreds of millions to over $2 billion. 2. Evolution of the Seed Stage and Valuation Traps • Sky-high Seed rounds and "insurance policy" strategy: • In the face of super Seed rounds of $8-10 million and valuations often in the tens of millions, Founder Collective rarely leads the entire amount but adopts a side-by-side investment strategy, betting $500,000 to $1 million. • Smart founders view these veteran seed funds as "insurance policies"—leveraging their brand and patience to guard against the risk of larger funds abandoning investment before reaching $10 million ARR. • Abandoning uncapped SAFEs: Uncapped SAFEs and high valuations disrupt the mathematical logic of seed rounds (mean reversion and multiple expansion). However, in the case of extremely scarce founders, exceptions may be made to support. • Framework and missed opportunities (FOMO): Adhering to venture capital discipline (such as maintaining post-money valuation caps and requiring appropriate ratios) may lead to missing out on some $10 billion opportunities (like 11 Labs, Klaviyo, etc.), but the framework is fundamental to protecting the fund from collapse over multiple cycles. 3. Founder Character and Partner "Alchemy" • The golden combination of CEO and CTO: The preferred co-founder structure is "technical wizard (CTO) + strong sales and leadership entrepreneur (CEO)." • Founder's learning curve: The CEO's key task is to quickly transition from "doing technology/products" to "building teams and recruiting top talent (bums on seats)." Outstanding CEOs will dedicate at least 30%-50% of their energy to talent recruitment long-term. • Deeply rooted "Nepo Babies" in vertical fields: Preference for founders who have been immersed in a specific vertical industry (like family pharmacies, HVAC mechanical engineering, audio technology) from a young age, possessing deep "native insights (Edge)" into industry pain points. 4. SaaS Dilemma, Suno Phenomenon, and Exit Mechanisms (Secondary & DPI) • SaaS killers and the "last 5% barrier": • The market is overly concerned about AI consuming traditional SaaS (SaaS Apocalypse), but deeply embedded software that carries core business flows (like contracts, pharmaceutical R&D, supply chains) has strong resistance to replacement. • For companies that are extremely embedded, the market may misjudge; for lightly embedded software, replacement by AI efficiency tools is an irreversible trend. • The consumer-level explosion of Suno: The rapid emergence of AI music/audio generation tools like Suno is unprecedented. Its core essence is the consumer experience and product interface (similar to how Spotify replaced traditional records), with the underlying large model being invisibly encapsulated for users. • Secondary market (Secondary Market) and DPI priority: • Current secondary market liquidity has reached historical highs. For mature top projects, appropriately discounting 20%-25% of equity in the secondary market to lock in DPI (distributed profits) is much wiser than waiting for an IPO and lock-up period 5-6 years later. • Emphasizing that venture capital funds should not blindly pursue inflated TVPI (paper returns), "velocity of cash" and real DPI are the hard truths for the longevity of funds. 5. Macro Trends, Hardware Reconstruction, and Future Super Waves • Rise of Physical AI: Future giants will emerge from deeply embedding AI into highly commoditized hardware and physical entities (like drones, security, medical devices). • Photonic computing disrupts computing power: Predicting that in the next 10 years, photonic chips will completely replace traditional electronic chips and transmission, solving the massive energy consumption bottleneck of data centers, and may even pose a fundamental physical-level disruption to existing chip giants (like NVIDIA). • The duel of the giants: China and the U.S. will become the only two AI superpowers globally. China is rapidly advancing in foundational photonic/energy research, biomedicine restrictions, and open-source ecosystems, which should not be underestimated.