20VC: Why VC Today is Worse Than 2021 | Why Vertical SaaS is a Bad Investment Today | Why We Are Deluding Ourselves on Growth Expectations | Revolut Raises $3BN at a $75BN Valuation | Benchmark Adds Their Newest General Partner

What It Covers

Jason Lemkin and Rory O'Driscoll debate venture capital's current challenges: TAM exhaustion in vertical SaaS, AI market saturation risks, Benchmark's newest partner addition, Revolut's $75B valuation, and why 2025 investing feels harder than ever.

Key Questions Answered

• •TAM Exhaustion Risk: Vertical SaaS companies hitting revenue ceilings faster than expected because everyone is in-market for AI tools now, similar to 2020 COVID buying patterns. This temporary surge creates false growth signals that won't sustain, making billion-dollar exits harder to achieve despite strong early traction.
• •Market Timing Compression: AI adoption is compressing what should be five to seven years of steady customer decisions into one to two years of frenzied buying. After initial purchases, companies face high business process change costs and won't buy again annually, creating inevitable growth deceleration that current valuations don't account for.
• •Capital Intensity Escalation: Poolside building its own two gigawatt data center signals that competing in AI now requires owning infrastructure, not just software. This transforms venture from needing $500M to breakeven into requiring $5B, fundamentally changing risk profiles and making temporal diversification nearly impossible with eighteen-month fund cycles.
• •Late-Stage Concentration: The easiest money in 2025 comes from doubling down on already-proven winners like Revolut at $75B rather than early-stage bets. Two-thirds of smart capital now flows to post-public-eligible companies with venture structures, effectively doing public market investing with private market fees and illiquidity.
• •Early-Stage Disruption: Code generation tools like Replit reaching $250M ARR in ten months makes judging early-stage founders harder because product quality no longer signals founder capability. When nineteen-year-olds can build production-ready software in days, traditional evaluation methods fail, requiring new frameworks for assessing founding teams beyond product demos.