Consumer AI Needs New Revenue Models, A16z’s Olivia Moore Says

Andreessen Horowitz partner Olivia Moore argues that consumer AI remains an early market despite ChatGPT’s clear lead among consumer AI applications. In her latest top-100 consumer AI apps report, Moore identifies durable smaller players including Suno and ElevenLabs, while highlighting a revenue structure still dominated by subscriptions and token usage.

The financial picture is uneven. Moore cited the State of Markets finding that 2.2% of US households pay for AI, and said much of the revenue generated so far is concentrated among enterprise and prosumer users rather than a mass consumer audience. OpenAI’s greater enterprise focus, in her view, is an expansion rather than an abandonment of consumer launches.

Subscriptions are not the only route

Moore questioned whether a substantially larger share of households needs to pay subscriptions for consumer AI to become attractive. She suggested that many users may prefer free access supported by advertising, with an optional subscription to remove adverts, rather than another recurring charge from their own budgets.

That approach faces a structural constraint: the marginal cost of AI services remains materially higher than that of established internet products such as Facebook or Google Search. The economics make it harder to apply the familiar ad-supported consumer internet model without attention to the cost of serving each request.

Cheaper models could widen the market

Moore said consumer products do not require frontier-level intelligence for every task. She pointed to ChatGPT’s $8-a-month Go plan as an example that she assumes uses cheaper models. Lower-cost models and open-source models could therefore become more common in consumer-focused products, particularly where the model is not itself the product.

For now, users generating much of the revenue are often using coding and technical automation products, tasks for which frontier intelligence may still be necessary. That helps explain why the strongest commercial activity sits in products with a professional rather than purely recreational character.

Prosumer demand dominates the rankings

Moore described much of what is labelled consumer AI as prosumer AI. Companies can begin with individual users and become majority-enterprise businesses within 18 months, unlike earlier consumer-first companies that could take six or seven years to add team or enterprise plans. She named Gamma, ElevenLabs and Cursor as examples of this faster transition.

The spending clusters she highlighted are product-building apps such as Lovable, Replit and Fal; product-marketing tools including Higgsfield and HeyGen; and work-management products such as Manus, Fireflies AI and Granola. Individuals may initially pay for these services, but the underlying use is closely connected to work.

Moore also found no entrants in the top 100 across social, dating, marketplaces, retail, travel, finance and health. For businesses, the practical implication is to distinguish prosumer workflows from genuinely consumer use cases, then match pricing and model costs to the specific audience rather than assuming subscription revenue will scale uniformly.