20VC: Is Seed Investing Dead Without a $1BN Fund? | Does Ownership and Price Matter When Companies Can Be $1TRN Exits | Are AI Revenue Numbers Real and What to Watch Out For with Venky Ganesan, Menlo Ventures

Menlo's Venky Ganesan on owning outliers in the AI boom

Venky Ganesan argues that venture's answer to the AI boom is selectivity rather than retreat. Seed checks are options on outliers, so ownership and position size should rise only as an outlier is proven. A small slice of a trillion-dollar company beats a large slice of a small one, while gamed metrics and leverage-driven cycles are the risks to watch.

Takeaways

• Seed investing is an option bet on outliers. Venky Ganesan says each seed check buys an option on whether a company becomes an outlier. Seed is hard in core AI because neolabs raise huge sums, but a snapshot in time should not set long-term strategy. Firms should be selective.
• Ownership matters before an outlier is proven. Venky would rather own 2% of a trillion-dollar company than 20% of a $100m one. Menlo owns under 2% of Anthropic yet put 20% of one fund into it. Position size should follow new data, and he expects about 60% dilution to exit.
• Any heavily weighted AI metric gets gamed. Run-rate revenue extrapolated from a single good day is murky. Venky says net revenue retention would read 100% on a $1 order. The real test is whether founders focus on terminal value rather than on markups.
• Cycles usually crack through debt and leverage. Venky links rapid markups to Soros' reflexivity and says cycles usually break through debt defaults because people use leverage. Tranche financing began as a way to separate value-adding capital but is now used regardless of company quality.
• Missed deals are the costliest mistakes. Venky says passes cost more than bad investments. He skipped a meeting with Sean Parker that could have meant a $50,000 check. Refusing to sell a small AvanEx IPO stake that fell 90% taught him to take some chips off the table.
• Incumbent acquirers are not downside protection. Venky cites deals such as Nvidia's $14bn purchase of Hugging Face and Stripe's reported $8bn purchase of OpenRouter, and compares them with dot-com buys by Nortel and Lucent. He expects more exits but warns against assuming a strategic will rescue a startup.
• LPs want DPI, and IRR beats multiples. Venky defends fast deployment because AI companies need real capital. Menlo Eight, invested over ten months in 2000 and 2001, is the one fund that failed to return capital. Menlo Nine and Ten underperformed. LPs want DPI, and public markets set the bar for IRR.

How Venky judges people and firms

• He values founders who explain complex ideas simply.
• He asks founders what their five best friends would say about them.
• He tells LPs to judge firms by their partners, and calls Benchmark hard to beat.