At a glance
WHAT IT’S REALLY ABOUT
Consumer AI shifts to agents, but spend remains power-user dominated
- The 7th Top 100 Consumer AI Apps report broadens methodology by adding consumer card-spend revenue data, which reveals many high-earning products that don’t rank highly by traffic.
- Consumer AI adoption is broad but monetization is narrow: roughly half of Americans use AI, yet only ~4.5% pay, and revenue is dominated by a small cohort of power users.
- Personal agents are the emerging wave, moving AI from conversational assistance toward executing tasks, but growth is constrained by privacy/trust concerns, platform permissions, and high serving costs.
- The current subscription-heavy monetization landscape is viewed as temporary and historically atypical, with ads and transaction models expected to expand as inference costs fall and user density rises.
- Among major assistants, ChatGPT remains dominant, while Claude shows surprising paid momentum versus Gemini, and specialized creative tools retain advantages where workflow depth and modality focus matter most.
IDEAS WORTH REMEMBERING
5 ideasRevenue rankings reshape what ‘winning’ looks like in consumer AI.
The 7th edition expands beyond traffic rankings by adding consumer card-spend revenue data, revealing a different leaderboard than web/mobile usage alone. Notably, 29 of the top 50 products by consumer spend didn’t appear on either traffic list, implying monetization and engagement can diverge sharply from raw reach.
Consumer AI is currently a power-user market, not a mass-market one.
While about half of Americans report using AI, only ~4.5% pay for an AI subscription, and spending is extremely concentrated. The median payer spends ~$25/month, but the top 1% spends ~$903/month, with the top 10% of payers driving more than half of payer revenue.
Agents are the biggest new trend, but privacy and platform access are the scaling bottlenecks.
Personal agents are shifting AI from ‘tool that helps’ to ‘software that does,’ but mainstream adoption is constrained by costs, unclear everyday use cases, and the trust/privacy barrier of granting agents access to email, credit cards, and personal data. Platform permissions also matter—e.g., some sites block agent-driven purchasing while others partner (e.g., Shopify).
Subscriptions dominate because costs are high—but ads and transactions may be the endgame.
Most consumer AI apps monetize via subscriptions/credits today (the report cites ~85% subscription monetization on the web list), which the speakers call historically ‘inverted’ versus prior consumer internet eras dominated by ads and transaction fees. High inference/COGS pushes companies to charge early, but they expect ads/transactions to re-emerge as costs fall and density rises.
AI-native ads can scale unusually fast, but trust-sensitive execution is critical.
OpenAI’s advertising is cited at ~$1B annual run rate, enabled by massive usage density (a referenced figure of ~1.2B weekly active users) and the possibility of superior targeting due to conversational, high-intent context. However, ads must be ‘deft’ to avoid damaging trust in a highly personal interface.
WORDS WORTH SAVING
5 quotesAbout half of Americans report using AI. Around 4.5% of US consumers are paying a subscription to an AI product. The top 1% user is spending $903 per month personally on their personal credit cards on AI.
— Olivia Moore
The top 10% of users are driving this market.
— Olivia Moore
We have transcended the need for everyone to buy a subscription to AI, and we need to see these other business models come back.
— Olivia Moore
If you don't trust the thing that you're giving it your most intimate access to your email or your life or your credit card, and it might do rogue things that you wouldn't expect or share things with other people you wouldn't expect, you know, that's gonna be the biggest impingement on these things really growing as, as massive consumer apps.
— Josh Elman
People are now building software, and they're building these really rich products, and they're actually, when they're delivering this whole experience to a user, it is much more than just a, a shim on the model or a harness that brings some context to the model.
— Josh Elman
High quality AI-generated summary created from speaker-labeled transcript.
