Do Revenue and Margins Still Matter in AI?
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
Harry Stebbings sits down with David George, General Partner at Andreessen Horowitz's growth team, for an unfiltered conversation about how venture investing is evolving in the AI era. (64:00) David shares insights from backing category-defining companies like Databricks, Stripe, OpenAI, and SpaceX, explaining why traditional growth metrics might not apply to AI startups and how the extension of private markets has changed everything. The discussion covers the hot topics shaping venture today: when it makes sense to pay up for early-stage AI companies, why fear of theoretical competition kills great investments, and how to spot exceptional founders before the market catches on.
• Main themes: The conversation explores how AI is reshaping venture investing, the evolution of growth metrics, the challenge of evaluating high-priced early-stage AI companies, and the art of identifying exceptional founders with "strength of strengths."Speakers
David George
David George is a General Partner at Andreessen Horowitz where he leads the firm's growth investing practice. His team has backed some of the most defining technology companies of this era, including Databricks, Figma, Stripe, SpaceX, Anduril, and OpenAI. He is now actively investing in the next generation of AI startups including Cursor, Harvey, and Abridge.
Harry Stebbings
Harry Stebbings is the host of "20VC," one of the world's leading venture capital podcasts. He has built a reputation for conducting in-depth interviews with top-tier investors and entrepreneurs, providing insights into the venture capital industry and startup ecosystem.
Key Takeaways
Invest in Strength of Strengths, Not Lack of Weaknesses
David emphasizes Ben Horowitz's philosophy of focusing on what founders do exceptionally well rather than their weaknesses. (21:28) Many investment mistakes happen when VCs overweight the fear of theoretical future competition or worry too much about market size limitations. Using the example of missing deals like 11 Labs and Deal, David explains that exceptional founders with spiking strengths can overcome most challenges. This approach means accepting that great founders may have flaws, but their extraordinary capabilities in key areas (like product building, hiring, or market understanding) will drive success.
Revenue Quality Matters More Than Speed in AI Companies
While AI companies can scale to $100M ARR faster than ever before, David warns that rapid growth doesn't automatically mean quality revenue. (28:01) The bar for evaluating AI companies has gone up significantly - investors must look closely at retention and engagement metrics rather than just growth rates. Companies like Gamma succeed because they combine organic customer acquisition with high engagement and retention. If a company pitches as an AI business but shows SaaS-level gross margins, it likely means customers aren't actually using the AI features.
Market Pull Trumps Everything
David consistently looks for companies with extreme market pull - where customers are "starving" for the product rather than needing to be convinced. (30:17) Examples include customer service AI companies like Decagon, where market demand is so strong that every product demo converts to deals. This organic demand is more valuable than any amount of sales and marketing spend. Companies with genuine market pull can build sustainable moats because they're solving urgent, painful problems that customers desperately need solved.
The Private Market Extension Creates Massive Opportunities
The extension of private markets has fundamentally changed venture investing, with companies staying private longer and growing larger before going public. (04:38) David notes that 47% of value creation in top IPOs happens between Series A and B, while 53% happens from Series C onwards. The private technology market is now over $5 trillion, meaning institutional investors need to adjust their asset allocation strategies. This shift allows growth funds to capture more value that previously would have been available only in public markets.
Human Labor Budget Transition is Critical for AI Success
For AI companies to reach their full potential, there must be a transition from human labor budgets to technology budgets, and this transition must be product-driven, not top-down mandated. (24:53) David cites CH Robinson as a compelling example - they achieved a 40% productivity increase and 680 basis point margin improvement by using AI to replace traditional call center operations. This type of demonstrable ROI proves that AI can successfully transition spending from human labor to technology, but it requires genuine product-market fit rather than executive mandates.
Statistics & Facts
- Andreessen Horowitz's best performing fund in the firm's history is actually a $1 billion fund, with Databricks returning 7x the fund and Coinbase returning 5x the fund so far. (03:43)
- The private markets have grown 10x over ten years and now represent over $5 trillion of market cap, fundamentally changing where value creation occurs. (04:18)
- CH Robinson disclosed a 40% productivity increase in shipments per person per day since 2022, driven by AI implementation, resulting in operating margin improvements of 680 basis points. (25:58)
