Most people tracking the consumer AI race are looking at the wrong numbers. Web traffic and app downloads tell one story. Money tells another. According to a16z’s seventh edition of its Top 100 Gen AI Consumer Apps report, published by Olivia Moore on October 5, the top 1% of U.S. consumer AI payers account for 19.5% of all observed consumer AI spend. The bottom 50% combined? Just 16.6%. That gap is the real story here.
Those top payers are spending an average of $903 per month on consumer card transactions for AI, a figure that is up 80% over the last 18 months. That kind of growth at the high end matters because it signals that a small group of deeply committed users is quietly funding a disproportionate share of the consumer AI economy. For founders building premium tiers and investors modeling monetization curves, this changes the math considerably.
Claude’s performance in that premium cohort is the most surprising data point in the report. Anthropic’s product has 7.3% of its consumer payers on the Max plan, which runs $100 per month. Compare that to 1.3% for Google’s Gemini and just 1.1% for ChatGPT at equivalent pricing tiers. Claude converted its paying users to its highest tier at a rate roughly six times better than ChatGPT. That’s a meaningful signal about the intent and profile of Anthropic’s customer base, even if the overall subscriber count still lags significantly.
And ChatGPT is still dominant on volume. The report puts OpenAI at three times more U.S. paid subscribers than either Claude or Gemini, with a 2.5x lead over Gemini and a 14x lead over Claude in monthly active users. So the picture is nuanced: ChatGPT wins on scale, Claude wins on premium conversion rate, and Gemini is competitive on subscribers but trailing on both ends of the spending spectrum.
The methodology shift this year is also worth paying attention to. For the first time, a16z added a top 50 ranking by observed U.S. consumer spending, drawn from a YipitData card panel. Of the top 50 vendors by actual spend, 29 do not appear anywhere on the web or mobile traffic lists. These are desktop apps and agents embedded inside messaging platforms, surfaces that tools like Similarweb and Sensor Tower simply cannot see. That’s a structural blind spot affecting how most of the industry benchmarks this category right now.
Zooming out, the broader consumer picture is one of wide but shallow adoption. Around 50% of U.S. consumers use AI in some form, 25% daily, but only 4.5% hold paid subscriptions to ChatGPT, Gemini, or Claude. Moore’s summary captures it cleanly: consumer AI usage is wide, but not very deep, for almost everyone. The report lands the same week Anthropic pushed out expanded startup pricing for Claude, which fits the pattern. The premium tier is where real revenue concentrates, and every major player knows it.
