Consumer AI Is Still Early, Says a16z's Olivia Moore

Consumer AI has a money problem. Usage is huge, but very few people pay, and running the models costs far more than serving a classic web page. Not everyone reads that as bad news. Olivia Moore, a partner at the venture firm Andreessen Horowitz (a16z) who covers consumer AI, sees a market that has barely started.

On Monday, Moore published a16z's latest ranking of the top 100 consumer AI apps. In a conversation with TechCrunch this week, she explained why she thinks the category is wide open and where its revenue could come from.

What the top 100 list shows

The headline finding is not surprising. ChatGPT leads by a wide margin. Further down, smaller companies such as Suno and ElevenLabs are holding their positions over time, which Moore treats as a sign of durable demand.

The more telling part is what the list leaves out. Moore's report names a set of everyday categories with no AI entrant in the top 100:

  1. Social apps
  2. Dating apps
  3. Marketplaces
  4. Retail
  5. Travel
  6. Finance
  7. Health

She called those gaps "pretty surprising" and said they are what she wants to see filled over the next six months.

Most "consumer" AI is really prosumer AI

According to Moore, nearly everything people call consumer AI today is prosumer AI: tools paid for by individuals but used for work. The revenue rankings make this clear. Paying power users cluster in three areas:

  • Product building: Lovable, Replit and Fal.
  • Product marketing: AI ad generators such as Higgsfield and HeyGen.
  • Work management: Manus, Fireflies AI and Granola, a note-taking app that now faces a local-first rival from Google.

Individuals may pay for these tools first, but Moore argues they are not consumer products in the pre-AI sense.

This links to an idea from an essay she wrote about a year ago, "The Great Expansion." Before AI, consumer-first companies like Canva often needed six or seven years before adding team or enterprise plans. Now companies such as Gamma, ElevenLabs and Cursor launch to consumers and become mostly enterprise businesses within about 18 months.

OpenAI's enterprise turn

OpenAI's renewed focus on business customers this year has fed doubts about consumer AI revenue. Moore sees it as an expansion, not a retreat, since the company still ships many consumer products. She also understands the move. So far, almost all AI revenue has come from subscriptions and token usage, and both lean heavily toward enterprise and prosumer buyers.

Subscriptions versus ads

A figure from the State of Markets report frames the problem: only 2.2% of U.S. households pay for AI. Moore said she is not necessarily hoping that number goes up. She would rather the industry find ways to earn money from consumers that do not come straight out of their wallets.

In her view, it is easy for high earners in Silicon Valley with corporate cards to say they would rather pay. Most people, she argued, would prefer free access with ads, plus the choice to subscribe if they want the ads gone.

The cost question

Price is the other obstacle. AI services still cost much more per user than Facebook or Google Search. Moore pointed to signs of change. ChatGPT now offers a Go plan at $8 a month, which she assumes uses cheaper models. Not every task needs frontier intelligence, and a16z founders have increasingly talked about building on open source models.

The catch is that today's paying users are mostly doing coding and technical automation, where top-tier models probably are needed. As more companies build for ordinary consumers, where "the model is not the product," Moore expects cheaper models to take a bigger role.

Asked whether true consumer AI has arrived yet, her answer was short: "It's definitely very early."

Our Take

Moore's argument comes down to a business model shift: away from subscriptions and toward the ad-supported, free-first model that built the consumer web. There are early signs of that in the market. OpenAI has started showing image ads in ChatGPT, and Google has moved free Gemini users to a lighter model. Both moves fit her point that cheaper models and ads may need to work together.

The empty categories are the real test. Dating, travel, finance and health are big, crowded consumer markets. It is worth watching whether AI-native apps can break into them, or whether incumbents add AI features first. It should also be kept in mind that a16z invests in this space, so its optimism is not a neutral forecast. Whether the "six months" timeline holds is an open question.