Summary
Read full transcript →Agrawal opens with the thesis that every major consumer technology cycle — search (Google, 90%+ share), mobile (Apple), social (Meta) — has produced winner-take-most outcomes, and he expects consumer AI to follow. He corrected an earlier thesis that Meta and Google would dominate AI distribution through their existing user bases; instead, ChatGPT captured the moment. Altimeter's growth fund made OpenAI its largest-ever investment after correcting that mistake. Key data points shared: ChatGPT has ~900M weekly active users and over 1B MAU; its DAU/MAU ratio is 45-50% versus Gemini at 22%, indicating far higher habitual engagement. ChatGPT is one of only three apps (alongside Chrome and WhatsApp) that shows a 'smile curve' in 12-month retention — users who churn come back.
On monetization, Agrawal notes that OpenAI is capturing only ~$10-15 per user per year across ~1B users versus Meta at ~$60 and Alphabet at ~$70, leaving significant upside via price discrimination (higher tiers for power users) and potentially ads for casual users. He frames private markets in three buckets: early-stage sub-$1B (always viable), $1-10B (most dispersion, hardest to underwrite), and quasi-public giants like OpenAI, Anthropic, Revolut, and Ramp that could IPO but choose to stay private. His growth fund holds only four investments and applies a high bar: visible line-of-sight to free cash flow. He also discusses Cursor vs. Claude Code dynamics as analogous to Databricks/Snowflake competing with hyperscaler-native data products — horizontal independent tools competing with the labs' own first-party products.
“History has shown us that these consumer markets have been winner take most. Sometimes winner take all honestly. You look at the biggest technology market started the internet. We got Google out of there for search 90% plus market share. ChatGPT's daily active usage over monthly active usage is about 45-50%. Gemini is at 22%.”