OpenAI’s Revenue worries are real and Apoorv breaks it down in his first lecture of Economics of AI Supercycle.

OpenAI has approximately 1 billion users. 90% percent of them pay nothing.

That is the central OpenAI problem as Agrawal frames it: a massive user base generating remarkably little revenue per user. ChatGPT monetizes at roughly $10 per user per year. Google monetizes its 4 billion users at approximately $100 per year. Meta monetizes 3.5 billion users at roughly $70 per year. The gap between OpenAI and Google is not a matter of scale - OpenAI already has a billion users. It is a matter of monetization model.

ChatGPT has also just crossed what Agrawal calls the niche-app threshold: usage comparable to Spotify or Twitter, but not yet at the social-app scale of Instagram or TikTok . This matters because

the ceiling on ChatGPT's user growth may be structural. Knowledge work - the primary use case - is not something all 4 billion people who are online do.

Asking technology questions actively is a behavior that characterizes a specific subset of the online population. Getting to mandatory-utility scale would require ChatGPT to expand well beyond its current use case.

The path from $10 to $100 per user, in Agrawal's view, does not run through subscriptions alone. Subscriptions at current price points might double or triple ARPU; they do not produce 10x improvement. What produces 10x is advertising. Agrawal predicts AI advertising in conversational interfaces will be the major 2026 monetization development, drawing an analogy to mobile advertising at Facebook's 2012 IPO when the bear argument was that there was no room for ads on a phone screen. The intent-based nature of AI queries, he argues, would allow for more precisely targeted ads with better attribution than any prior platform - because the user is logged in and expressing explicit intent.

Whether that prediction proves correct is the open question. Demis Hassabis at Google has already announced that Gemini will not use advertising as a revenue model. If OpenAI follows a similar positioning, the advertising thesis does not materialize from the two largest consumer AI players.

Agrawal also identifies OpenAI as one of the two companies that dominate approximately 90% of application layer revenue in the AI ecosystem today . That concentration is both an asset and a ceiling - it reflects genuine dominance in consumer AI, but it also means OpenAI is competing primarily with Google for the same user base and the same monetization unlock.