A16z’s Olivia Moore on the state of consumer AI

The Subscription Fatigue Problem

We have reached a point in the AI cycle where everyone is tired of the $20-a-month toll booth. If you are building in this space, you probably feel it. The initial rush of users paying for ChatGPT Plus or Midjourney has plateaued into a reality where people are scrutinizing their credit card statements. A16z partner Olivia Moore recently pointed out that the industry is at a crossroads regarding how we actually make money from these tools.

For founders, the subscription model was a safe harbor. It is predictable and easy to code. But it is also a barrier. When every new AI tool asks for a monthly commitment, the consumer shuts down. Moore’s observation is that the next phase of consumer AI won't just be about better models, but about more creative ways to capture value without scaring off the user before they even start.

Moving Beyond the API Toll

Right now, most startups are just wrappers around someone else's compute. You pay OpenAI or Anthropic for every token, and you pass that cost plus a margin to the user. It is a thin-margin game that favors the big tech companies providing the plumbing. Moore suggests that the real opportunity lies in tapping into revenue streams that aren't tied to the raw cost of the technology.

Think about how the web evolved. We didn't end up paying a subscription for every single website we visited. We paid through attention, through transactions, and through data. In the AI world, this means moving toward performance-based billing or embedded commerce. If your AI agent helps a user buy a house or plan a wedding, the value is in the outcome, not the number of queries it took to get there.

The Multi-Model Reality

We are also seeing a shift away from the one-model-fits-all approach. Builders are realizing that they don't need a massive, expensive LLM to handle simple UI tasks. This is where the founder's perspective becomes vital. You have to be an architect of efficiency. If you can swap a high-cost model for a smaller, specialized one that costs a fraction of a cent, your business model suddenly has room to breathe.

Moore highlights that the consumer doesn't care what is under the hood. They care if the output is good and if the price is right. For builders, this means the competitive edge is no longer who has the best prompt, but who has the best unit economics. If you can provide a service for free that your competitor charges $20 for because your backend is optimized, you win.

The Retention Wall

Retention is the silent killer of consumer AI startups. It is easy to get a million downloads when a tool goes viral on social media, but it is incredibly hard to keep those users after the novelty wears off. The data suggests that many AI apps have a high churn rate because they are treats, not staples. They are fun to play with for ten minutes, but they don't solve a recurring, painful problem.

To build a staple, you have to integrate into the user's workflow or daily life in a way that feels invisible. This is where Moore’s push for diverse revenue streams comes in. If a tool is free to use but makes money when the user achieves a goal, the friction of keeping the app installed disappears. You are no longer asking for a permission to bill their card every month; you are offering a partner that pays for itself.

Where the Builders Should Look

If you are starting a project today, don't start with the pricing page. Start with the value exchange. Ask yourself if the user would pay for this if it weren't 'AI.' If the answer is no, you are just selling a gimmick. The AI should be the engine, not the product. We are seeing a massive shift toward 'agentic' workflows where the AI actually does the work instead of just talking about it.

• Focus on outcomes: Don't charge for the process, charge for the result.
• Optimize the stack: Use smaller models for smaller tasks to preserve your margins.
• Avoid subscription bloat: Look at affiliate models, transaction fees, or tiered usage that doesn't feel like a penalty.

The Skeptic's Corner

I have to stay grounded here. While the vision of diverse revenue streams sounds great, it is much harder to execute than a simple Stripe subscription. It requires deeper integration with payment processors, more complex legal frameworks, and a higher volume of users to make the math work. Not every founder has the luxury of waiting for transaction fees to pile up while their server costs are skyrocketing.

However, the alternative is a slow death by churn. The companies that survived the early days of the internet were the ones that figured out how to make the technology feel free to the end user while building a massive business behind the scenes. AI will be no different.

The Takeaway

The honeymoon phase of consumer AI subscriptions is over. If you want to build something that lasts, you have to look at the business model as part of the product innovation. The winners of the next three years won't be the ones with the smartest chatbot; they will be the ones who figured out how to make AI disappear into the background of a sustainable, non-predatory business model. Stop charging for tokens and start charging for value.

Read the original at TechCrunch AI →