Olivia Moore on the Economics and Early State of Consumer AI

I recently wrote about the bleak economics of consumer AI, but while some view this as a crisis, others see a significant opportunity. Olivia Moore, a partner at Andreessen Horowitz covering consumer AI, released a report on Monday detailing the top 100 consumer AI apps. While ChatGPT remains the dominant player by a large margin, smaller applications like Suno and ElevenLabs are demonstrating real staying power. Even more intriguing is what Moore’s report reveals: six consumer categories appear to be completely untouched by artificial intelligence.

Moore identifies a substantial opportunity in this sector, particularly if the industry can leverage revenue streams beyond simple subscriptions and API charges. I spoke with her this week regarding the unusual economics of consumer AI and why the category is still in its infancy. This interview has been edited for length and clarity.

It is a complex moment for a consumer AI report, especially following OpenAI’s pivot back to the enterprise this year, which has fueled pessimism about overall revenue. However, I asked Moore if she is more optimistic. She agrees that OpenAI has shifted focus, but views it not as a pivot but as an expansion, as they continue launching consumer products. She acknowledges that nearly all current AI revenue stems from subscriptions and token usage, which are heavily concentrated on the enterprise and prosumer sides.

How can consumer AI products resolve this revenue challenge? The "State of Markets" report highlighted that only 2.2% of U.S. households currently pay for AI services. Moore questioned whether this user base needs to expand before the market becomes attractive. She expressed a preference for a monetization model where consumers are compensated—possibly through ad revenue—rather than paying out-of-pocket for subscriptions.

Moore noted that it is easy in Silicon Valley, where high-income individuals and corporate cards are common, to assume users prefer direct payment. She argues that most people would prefer free access supported by ads, with the option to subscribe to remove them. She also pointed out the high marginal costs of AI services compared to classic internet services like Facebook or Google Search.

Regarding cost reduction, Moore believes improvements are underway. She noted that ChatGPT now offers a $8-a-month "Go" plan, likely running on cheaper models. She believes frontier intelligence is not necessary for every task and anticipates a shift toward lighter models and open-source solutions as more companies build on top of them.

Moore pointed out the unusual nature of technical automation in a consumer service, suggesting a blurring of lines between the two sectors. She referred to her earlier article, "The Great Expansion," to illustrate that pre-AI companies like Canva took years to adopt enterprise plans, whereas companies like Gamma, ElevenLabs, and Cursor transitioned to majority-enterprise businesses within 18 months.

Moore argues that most "consumer AI" is actually "prosumer AI." Her revenue data shows three primary categories driving spending: product-building apps such as Lovable and Replit, AI ad generators like Higgsfield and HeyGen, and general work management tools like Manus and Fireflies AI. While these are paid by consumers initially, she views them differently than traditional consumer applications.

The report also highlights vast white spaces, including social apps, dating apps, marketplaces, retail, travel, and finance, which currently have no AI entrants in the top 100 list. Moore believes this needs to change over the next six months. Consequently, she concluded that we have not yet seen true consumer AI, as it remains very early days.