Small Fund, Big Waves: Founder Collective's David Frankel on Discipline in the AI Cycle
Key takeaways from 20VC's interview with David Frankel, co-founder of Founder Collective (August 2026):
Sticking to boutique seed. As peers scale into ever-larger funds, Frankel argues Founder Collective's discipline is the edge — backers of Uber, Coupang, SeatGeek, Olo and PillPack pre-AI, and Suno (now ~$5bn) and Shield AI in the current wave.
The $50–100M seed trap. He calls mid-sized seed funds the worst-performing vintage: too big to write friendly $100–250K checks, too small to lead an $8–10M round. Founder Collective still does $3–4M rounds and small $500K–$1M "insurance policy" positions alongside larger investors.
Price matters less — to a point. Frankel pushes back on "price doesn't matter if you're in the winner," calling uncapped notes at seed bad math. Of the top 500 companies created in 25 years, the median outcome is ~$2.6bn — owning 5% of one returns a fund.
Roadkill ahead. He expects enormous fallout from the AI wave — "95% are not going to be there" — while naming OpenAI, Anthropic and SpaceX as the likely generational winners, the "Googles of our era."
What he backs. Frankel looks for CEO–CTO "alchemy" (a magician CTO, a salesperson CEO), prizes youth, focus and intensity, and says the team now leans toward deep-engineering founders from DeepMind and Gemini.
The discipline tension. Even Benchmark has raised a growth fund; Frankel admits the pull to raise more is real, but says Founder Collective's GP is its largest LP — aligning the firm to chase returns and DPI over management fees.
Patience as strategy. He warns "triple-triple-double-double" is harder to sustain, notes 10-year funds now take ~18 years, and cites still owning every share of SeatGeek from a 2010 investment as proof great outcomes just take a long time.
Read more: 20VC