What Does it Take to Be Good at Series A and B Today?

Rory O’Driscoll, Jason Lemkin, and Fabrice Grinda debate Series A/B pricing risk, AI bubble dynamics, and why fund-returner thresholds have permanently risen.

• At Series A today, investors pay Series A prices for seed-level risk — pricing compress leaves no margin to bury picking errors.
• Grinda’s 2017 fund sits at 4.31x / 32.56% IRR; Lemkin argues IRR is the hardest VC metric to control and largely outside a GP’s hands.
• Grinda generates most exits via secondaries on the way up, using Shares Post and Forge to hit 30% IRR across 500-deal diversified funds.
• Benchmark’s $75M lead in Manus (Chinese AI) framed as idiosyncratic portfolio risk that could become firm-level congressional/reputational risk — O’Driscoll says he would skip it.
• Grinda exited China after Xi personally killed Ant Group’s IPO; exited Russia after 2014 Crimea invasion; now backs Ukrainian defense startups on cost-per-kill metrics.
• Lemkin: half of mediocre SMB sales reps and customer success teams will be gone within two years as AI replaces them.
• 10x engineers now effectively 100x with AI tooling; Windsurf reports 50% productivity gains across engineering — the gap between elite and average widens, not narrows.
• Sequoia’s evergreen fund structure is the correct long-term insight, just launched at the worst possible timing window.

2025-05-01 · Watch on YouTube