Pocket FM's AI-written audio series push it to a $500 million run rate
India's Pocket FM says its annualized revenue run rate has doubled to $500 million, and AI is doing most of the production work behind it. CEO Rohan Nayak told TechCrunch that AI now powers 93% of the platform's catalog and 99% of new content — built on in-house models for creative writing and text-to-speech, trained by AI head Vasu Sharma's team on years of production data and listener engagement signals.
The economics are the story. Nayak says AI has made content production roughly 80 times cheaper, and 100 hours of audio that took about a year to produce now takes a day. More than 550,000 creators generate about 2.5 million hours of AI-powered content annually — against a total catalog of 770,000-plus audio series, up from around 100,000 hours two years ago. The result: 12-month revenue retention jumped to 76% from 44%, which Nayak credits to having more stories matched to listener tastes. Some 96 titles have each passed $1 million in revenue, 13 of them $10 million-plus. The platform claims more than 250 million listeners across over 20 countries, and the US is now its largest market at roughly 70% of the run rate, growing 70% year over year. The company is profitable and cash-flow positive on an adjusted basis, and Nayak says successful series will be licensed into books, television, and film.
The take: this is the clearest evidence yet that AI-generated long-form entertainment can hold paying audiences at scale — not as a lab demo, but as a $500 million business. The counterweight is Nayak's own framing: "We want to create great IPs that last 100 years, and that needs humans." The durable franchises stay human-led; AI fills the long tail. That's a blueprint every media company watching its margins will study. Parent Pocket Entertainment is also pushing the same engine into microdramas (Pocket Saga at a $15 million run rate three months in), with funding talks reportedly underway.
What to watch: whether retention holds as the AI share of new content climbs past 99% — and whether a US listening base this large starts demanding disclosure of what's human-made.
Would you pay for an audio series knowing it was written and narrated entirely by AI? Tell us in the comments.
