David George says the dominant AI product isn't found, and ChatGPT has 1B users but under 50M monetized.
He is more skeptical than consensus on enterprise AI, calling white-collar labor capture hand-wavy.
Fewer than five public software, consumer and fintech companies grow 30%, while Andreessen Horowitz's portfolio grows 112% dollar-weighted.
a16z's growth chief on where AI's value will land
a16z's growth business bets that the market misprices the largest platform shifts, and AI is the biggest one. George treats this as early. He concentrates in the market leader, backs technical founders who become operators, and runs a small debating team without an investment committee.
Takeaways
AI's dominant product has not been found. George argues the chatbot interface is too limiting, and the next shift is proactive assistants with long-form memory and multimodality. ChatGPT, with a billion users and fewer than 50m monetized, is best placed to capture it.
Consumer AI monetization has room to grow. Facebook and Google went from $20 to $200 of revenue per user in the developed world. AI users spend 30 minutes a day in these products, against 50 on Instagram and 70 on TikTok, leaving room for a massive company.
Enterprise AI surplus flows mostly to customers. George is skeptical, against consensus, that AI companies will capture white-collar labour spending. He argues 90% of the surplus goes to end users, as with the steam engine and iPhone. Customer support and coding are furthest along, priced on completion rather than seats.
Technical founder-operators are his archetype. George backs deeply technical founders who become great business people, naming Ali Ghodsi, Mark Zuckerberg, Elon Musk, George Kurtz, Dylan Field and Shiv Rao. Travis Kalanick is the non-technical counterexample, and Ghodsi now knows sales operations and hiring better than most of the firm's CEOs.
Market cap accrues to the leader. The vast majority of market cap creation goes to the market leader. The model layer will look more like cloud than winner-take-all, because cloud customers are well served and the clouds remain good businesses.
Waymo was a contrarian a16z bet. Marc Andreessen and Ben Horowitz insisted autonomous driving was the mother of all markets while George thought it would take 10 years. a16z was the only outside investor in 2020, and Waymo's 400 cars in San Francisco overtook 50,000 Lyft drivers in market share.
The Figma decision
During COVID, Dylan Field told Andreessen Horowitz that now was the time.
The growth team questioned a $2bn price against the small market for designers.
The venture team argued the designer-to-engineer ratio would double and front-end engineering was merging with design.
George took the market-size risk, and the firm ran a full court press that included a board seat from its network.
How George runs the firm
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
Bob Swan told George that he and John Donahoe cut 30% of their calendars each year.
George blocks two hours on Tuesdays and Thursdays, plus 90 minutes twice a week, to think and read.
The growth team is about 10 investors, possible because the firm has early stage teams.
The firm will not run a buyout fund, because buying incumbents and squeezing customers is antithetical to helping the next generation beat them.