CharlieKey takeaways from your favorite podcasts
← 5 October 2026a16z

The Current State of Consumer AI

Consumer AI revenue is concentrated among a small group of power users, while personal agents, ads, and software-layer startups open the next wave, and categories built around spending time remain largely unbuilt.

Watch on YouTube ↗

Charlie's summary

  1. Consumer AI is a power-user game: about 4.5% of Americans pay, and the top 10% of them drive over half of revenue

    Olivia says about half of Americans use AI, but Ebit card data shows roughly 4.5% pay for a subscription, double a year ago. Within that group the top 10% generate more than half of revenue and the top 1% about 20%. The top 1% spend $93 a month personally, the median $25. Top spend goes to tools like n8n, Granola, Higgsfield and Manus, which are used to build, make and sell.

    ▶ 14:18
  2. The traffic list is settling, but the spend list shows a different market

    Olivia notes only 11 new products appeared across the web and mobile lists, the fewest in any edition. Yet 29 of the 50 products ranked for spend were on neither traffic list. Ranking by traffic alone misses what consumers actually pay for, so this edition added consumer card spend data for the first time.

    ▶ 1:31
  3. Personal agents are the biggest new trend, but they still haven't reached mainstream users

    Olivia says OpenClaw sparked the shift from AI as a productivity tool to AI that gets things done, then faded from the rankings as safer products emerged: Muse, Instinct, Poke and Town. Instinct reported 100,000 users growing 10% a day. Muse hit about 5 million US and Canada downloads in 22 days, against roughly 16 million for Threads. Growth is fast, yet Olivia says it is not mainstream.

    ▶ 3:09
  4. Agent costs of thousands of dollars a month reflect coding-heavy early adopters, not typical consumers

    Olivia cites assistant benchmark data showing the top use case, even on consumer agents like Muse and Instinct, is coding and technical automation. That explains the huge serving costs. Mainstream-oriented agent products see costs in the tens of dollars. Josh adds that AI feels like a blank box to newcomers, and that people learn by copying what others share, which hasn't yet happened with personal AI.

    ▶ 11:41

Give your email to unlock the rest of the key points — for every episode.

Do more with Charlie

Free — sign in with your email, no password.