The a16z Show: "Do Revenue and Margins Still Matter in AI?"

Episode Date: December 18, 2025
Host: Harry Stebbings (20VC, presenting for a16z)
Guest: David George (General Partner, Andreessen Horowitz Growth Fund)

Episode Overview

In a revealing and candid conversation, Harry Stebbings interviews David George of Andreessen Horowitz to explore how investment strategy, company evaluation, risk assessment, and growth-stage decision-making have evolved in the era of rapidly advancing AI. The discussion ranges across fund performance, the changing dynamics of private vs. public markets, the new rules of revenue and margins in AI startups, the thinking behind high-valuation bets, and the principle of prioritizing founder “strength of strengths” over avoiding weaknesses. The episode is a deep dive into the mindset of a top growth investor at one of the world’s most influential venture capital firms, and how that mindset adapts in the midst of dramatic technological shifts.

Key Discussion Points & Insights

1. Large Venture Funds & Returns

• Dispelling the 'Big Fund = Low Returns' Myth

  • David George asserts large funds can outperform:

    “Our best performing fund in the history of the firm is actually a $1 billion fund... Databricks has returned 7x the fund, Coinbase 5x, and we had others like GitHub, Digital Ocean, Lyft.” (03:02)

  • Growth in private markets and tech waves’ scale have enabled venture funds to produce “3x or 5x returns” even at larger sizes.

• Private vs. Public Market Value Shift

  • Dramatic increase in private market value:

    “Private markets have grown 10x over 10 years, over $5 trillion in market cap.” (03:47)

  • Value creation is now occurring earlier and longer in private markets before companies go public.

2. The Blurring Lines Between Private and Public Markets

• Longer Stay in Private Markets

  • Some worry that companies staying private longer risk being overtaken by competitors before IPO – David disagrees:

    “I don’t think whether it’s public or private has much to do with the competitive dynamics, to be honest... The more some companies have stayed private, it’s been to our benefit because we increased our ownership over time.” (06:31–07:03)

• Liquidity & Exiting in Private Markets

  • Despite large positions, a16z rarely sells in late private rounds:

    “We could, but we historically have not. For the companies that stayed private, we’ve been really excited to stay in them, keep backing them...” (07:37)

  • David’s view: Most late-stage companies will eventually go public, and most CEOs don’t regret the move.

• Advantages of Public vs. Private

  • Public: easier access to capital, sometimes cheaper capital.
  • Private: more control, less volatility, especially for giants like Stripe and SpaceX.

3. How Asset Class Changes Affect Institutional Investors

• Quality & Size Shift in Listed Companies

  • The number of public companies has halved over 20 years, with lower average returns, shifting top-quality tech investment opportunities to private markets.

  • “It’s no longer this bespoke, small thing. It’s like the big leagues... just look at the size of the private tech companies.” (11:18)

• Advice to Institutional Allocators

  • David: Emphasize allocation to venture/private growth as more value is created there versus legacy public or PE:

    “If you just start with where the 10 most valuable companies are today... Eight of the top 10 are US West Coast technology companies, all venture-backed.” (13:29)

  • Belief that AI will make the gap wider going forward.

4. Geography: US vs. Europe & Errors of Omission

• US Dominance but European Opportunity
  • US still leads, but great European entrepreneurs are being backed — example: 11 Labs.
  • Humorous and candid admission by Harry and David of deals they missed at seed-stage.
  • David’s lesson:

    “You can’t bat 1000, dude.” (15:35)

5. High Prices for Early AI Companies & 'Strength of Strengths'

• Is Paying Up at High Valuations Smart?

  • Sometimes, yes—for exceptional founders (“strength of strengths”).

    “For extremely, extremely special people like that, we’re comfortable to step into those situations.” (17:01)

  • Examples: Early rounds in Character AI and Gnome, where asymmetric risk/upside is clear.

• Errors of Omission Philosophy

  • a16z’s “fix the mistake fund” model: invest later in companies they missed earlier.
  • David:

    “When we make an investment, we should always be investing in strength of strengths as opposed to lack of weaknesses... If you overweight fear of future theoretical competition, you can always talk yourself out of making an investment.” (21:27)

6. Disruption, Business Models & The Real TAM Debate

• AI-Driven Disruption

  • SaaS incumbents face new disruption from AI, but new challengers must innovate on business model, UI/workflow, and data access to truly win.

    “If you have all three—major change at the same time—a startup has a really good chance to beat an incumbent.” (22:50)

• TAM Overestimation vs. Underestimation

  • Many past errors come from underestimating how big verticals could get, but there’s also now a “TAM trap” bubbling up in sectors like SaaS.

7. Automation, AI, and the Shift from Labor to Technology Spend

• Technology Must 'Slap Customers in the Face' with Pull
  • Adoption happens when the value is product-driven, not top-down.
  • Real-world evidence:

    “C.H. Robinson, a truck brokerage... saw a 40% productivity increase since 2022, with margins up 680 basis points. That’s very effective AI implementation.” (25:30)

  • Margins and market pull, not just revenue, are the new signals for business quality.

8. Revenue, Margins, and Unit Economics for AI Startups

• Do Revenue Numbers Still Matter?

  • Yes, if supported by retention and engagement:

    “It does mean the same as before—if it’s high retention and high engagement... The bar has gone up significantly for us when we look at AI companies… the most important is engagement.” (28:01)

  • Organic product pull/virality is a huge sign—ChatGPT, 11Labs, etc.

• Tripling and Doubling: Is Growth Curve Dead?

  • Not dead, but should be calibrated to market speed and peer set.

• New Attitude Toward Margins

  • a16z more forgiving of lower margins in AI (input costs are dropping, usage is growing), expecting rationalization over time:

    “The history of technology inputs would suggest margins will rationalize and the margins are going to go up... On the gross margin point, today I'll say this: we give a little bit more of a pass than we used to.” (39:06)

9. Is Kingmaking Real?

• Brand & Resources Matter, but You Can’t 'Out-Fundamentals' a Weak Company
  • David:

    “If the investment thesis is our investment is going to make them a winner, it's a flimsy thesis. But investing big in a company already attracting resources can help.” (31:48)

  • The Softbank Vision Fund example: money alone doesn’t anoint the winner.

10. Sector & Investment Process Insights

• Customer Service AI: Why So Crowded?

  • The sector is enormous, and “winner-take-most” or “share-split” both possible depending on business model and speed.
  • Decagon highlighted for staggering growth and market pull.

• Entry Price for Leaders like OpenAI

  • Regular reassessment required:

    “We are constantly surprised at how big in absolute dollar terms companies can be and how good they can be... it's limiting to think you're ever at an end-state.” (43:56)

11. Inside the a16z Growth Process & Decision-Making

• Avoiding Conflicts, Early Team Coordination
  • “Fix the mistake” approach, avoid board conflicts, but recognize that “companies diverge more often than they converge.” (45:53)
• Internal Disagreement: Waymo Case Study
  • Example of disagreement: high valuation vs. enormous market size in Waymo. Solution: small initial check, kept relationship, doubled down later. (46:23)

12. Picking Founders & “Flow”

• Backing Relentless, “Strength of Strengths” Leaders

  • “Adam (Flow) has some of the strongest strengths of any entrepreneur—brand, company building, product, hiring.” (49:33)

  • The logic: in rare cases of exceptional founders, almost always “write the check.”

13. Lightning Round—Notable Quotes & Fun Moments

• What Has David Changed His Mind on?

  • Belief that models (like GPT-4) would render app layers obsolete is wrong:

    “We’ve fully changed our mind. There’s going to be application software companies built on top of models in every direction.” (51:54)

• Most Memorable Founder Meeting

  • Shiv from Abridge:

    “He was perfect archetype: knows his market, knows tech, total killer.” (54:20)

• Best Pickers at a16z

  • David: “Dixon. He has the clearest articulation of our early-stage strategy…” (56:56)

• On the Magic of the Job

  • “Speed of execution, aggression, is a huge part of success in those [vertical software] categories.” (55:22)

Notable Quotes

• On 'Fear of Theoretical Competition':

  “If you overweight the fear of future theoretical competition, you can always talk yourself out of making an investment. We try really hard not to do that.”
  —David George (21:27)

• On Market Shifts:

  “You know, the asset class is no longer bespoke, small thing. It's the grown-up leagues.”
  —David George (11:18)

• On Entry Price and Growth Potential:

  “Our investment case never would have predicted what [Databricks] became... it's limiting to think you're ever at like an end-state of productivity or new products.”
  —David George (43:56)

• On Structural Moats in AI:

  “If people are using the product a lot and getting a lot of value out of it, that's a really good leading indicator.”
  —David George (28:01)

• On Picking Relentless Founders:

  “My favorite thing about Deal: Alex is just absolutely relentless... always selling. In a market like that, exactly what you need.”
  —David George (15:41)

Timestamps for Key Segments

• Large funds & returns myth: 03:02–06:01
• Private/public market shifts: 06:01–13:29
• Advice to allocators: 13:11–14:59
• “Errors of omission” and founder strength: 17:01–22:35
• Disruption and business model shift in AI: 22:50–26:51
• AI in operations (CH Robinson example): 25:30
• Revenue, retention, and engagement: 28:01–29:26
• Gross margins, AI input costs: 39:06–40:20
• The "king making" debate: 31:28–34:57
• Decagon, SaaS, and customer support category glut: 35:42–37:27
• Entry price for OpenAI, Databricks case: 43:40–43:56
• Founders’ traits (Flow, Adam Neumann): 48:55–50:55
• Memorable founder meetings: 54:20
• Best picker at a16z: 56:56
• Internal disagreements (Waymo investment): 46:17
• Personal optimism—health, robotics as future categories: 59:40

Conclusion

This episode provides a comprehensive blueprint for how a premier growth-stage investor is navigating the most dynamic investment climate in decades. David George illustrates the importance of founder quality, the nuanced approach to evaluating margins and revenue in AI-first businesses, and how a16z leverages scale, internal coordination, and flexibility to seize both early and late-stage opportunities. With a lively and candid tone, the conversation offers actionable insight for anyone interested in venture capital, the future of AI, and the new rules for investing in an AI-eaten world.