| Episode | Status |
|---|---|
| Episode | Status |
|---|---|
In this episode, we’re sharing a conversation with David George, General Partner at a16z on the firm’s growth investing team. David has been involved in backing many of the defining companies of this ...
David George, General Partner at a16z's growth fund, discusses how AI is reshaping growth investing and creating unprecedented opportunities. He explains why a16z's larger funds have outperformed smaller ones, how the private markets have grown 10x to $5 trillion, and why the firm gives AI companies more flexibility on margins than traditional SaaS. The conversation covers specific investment decisions including OpenAI, Waymo, and Flow, along with frameworks for evaluating competitive threats, entry pricing, and founder quality in the AI era.
David challenges the notion that large funds can't achieve high multiples, revealing that a16z's best performing fund is a $1B fund with 7x returns from Databricks alone. He explains how private markets have grown 10x to $5 trillion and why 53% of value creation now happens post-Series C, fundamentally changing the asset allocation calculus for institutional investors.
A spirited debate on whether private or public markets offer cheaper capital today. David argues that top private companies could access cheaper capital if public, while acknowledging benefits of avoiding stock price volatility. The discussion covers why companies like Stripe stay private and the changing dynamics of when to go public.
David reveals that 50% of growth fund investments are follow-ons from venture deals, 15% are follow-ons from prior growth investments, and only 33% are net new - but all net new companies have pre-existing founder relationships. He discusses painful errors of omission like missing Deal's Series B and the philosophy of investing in 'strength of strengths' rather than lack of weaknesses.
Discussion of whether SaaS TAMs are shrinking and how AI companies must transition spend from human labor budgets to technology budgets. David shares concrete evidence from CH Robinson (40% productivity increase, 680 bps margin improvement) and explains the three levels of disruption: business model shift, UI/workflow change, and data access.
David explains how evaluation criteria have changed for AI companies. While revenue still matters, the bar for retention and engagement is higher due to faster growth. He gives AI companies 'more of a pass' on gross margins than traditional SaaS, noting that if an AI company has SaaS margins, it likely means customers aren't using AI features.
David defends investing at growth-stage prices in very early companies when backing exceptional founders like Noam Shazeer (Character AI). He addresses the challenge of paying high prices ahead of revenue scale, arguing that the best AI companies can grow fast enough to justify valuations that seem extreme on traditional metrics.
David discusses investing in OpenAI before ChatGPT and how they constantly reassess whether entry prices make sense. He uses Databricks as an example where their 2019 investment case at $6B never predicted current scale, emphasizing the need to push thinking on how large companies can become, particularly in consumer AI where ChatGPT dominates.
David shares his biggest disagreement with Marc Andreessen and Ben Horowitz over the original Waymo investment in early 2020. Despite thinking the valuation was too high, Marc and Ben pushed for it based on autonomous driving being an 'endless market.' They compromised with a smaller initial check, maintained the relationship, and later invested much larger in a subsequent round.
David explains the controversial Flow investment, defending Adam Neumann as world-class at brand building, company building, product, and hiring. The thesis centers on renters spending 30% of income on the only unbranded experience in their lives, creating opportunity for premium branded rental experience. Emphasizes betting on extremely rare founder quality.
David describes changing his mind 18-24 months ago from believing models would subsume all applications to now seeing massive opportunities for application layer companies. Uses radiology example: AI can read scans better than humans, but radiologists spend 60-70% of time on other tasks that model companies won't address, creating opportunities for vertical applications.
David highlights memorable founder meetings with Shiv from Abridge and Winston from Harvey - combining deep domain expertise with tech founder aggression. He shares excitement for two emerging categories: proactive personal health management with AI coaching, and robotics as potentially the largest AI category in both B2B and B2C over the next decade.
Do Revenue and Margins Still Matter in AI?
Ask me anything about this podcast episode...
Try asking: