Pocket FM says artificial intelligence now supports 93% of its audio catalog and 99% of newly produced content. The Indian serialized-story platform links that shift to sharply lower production costs, much greater output and an annualized revenue run rate that has doubled to $500 million in a year.
The figures come from Pocket FM executives and describe the current pace of a private company, not audited full-year results. Even with that limitation, they show a media business attempting to turn generative AI from an experimental tool into its primary production system.
Human Ideas Feed an AI Production Line
Pocket FM still relies on people to originate ideas and shape stories. Co-founder and CEO Rohan Nayak said the company wants enduring intellectual property and believes that requires human creators. AI is used around that creative work to convert concepts into finished audio at scale.
The company has trained its own models for creative writing and text-to-speech, according to AI head Vasu Sharma. Training draws on years of production information and signals about how listeners engage with stories. Pocket FM says the resulting process makes content production roughly 80 times cheaper and can create 100 hours of material in one day, compared with about a year previously.
More than 550,000 creators now produce around 2.5 million hours of AI-assisted content annually. Two years ago, the platform’s entire catalog contained about 100,000 hours. It has since accumulated more than 770,000 audio series. Those comparisons are company-reported and do not reveal how quality, creator compensation or production costs are measured.
Revenue and Retention Rise With Volume
Pocket FM’s annualized revenue run rate rose from about $250 million a year ago to $430 million in April and $500 million now. The company calculates the number by multiplying one month’s revenue by 12 rather than using contracted recurring revenue. That makes it a snapshot of momentum, not revenue already earned for a full year.
Nayak said 12-month revenue retention increased from 44% two years ago to 76%, attributing part of the gain to a larger selection of stories for different tastes. Ninety-six titles have generated more than $1 million each, and 13 have passed $10 million, the company said.
Pocket FM reports more than 250 million listeners across over 20 countries. The United States, where the business grew about 70% over the past year, accounts for roughly 70% of its annualized revenue run rate. About $85 million of the total pace comes from advertising, while approximately $415 million comes from users paying to unlock episodes.
Growth also reflects expansion into the UK, Germany and France and the introduction of user-generated content in the US. The source therefore does not isolate how much of the revenue increase was caused by AI rather than geography, catalog growth or payment strategy.
The Model Is Moving Beyond Audio
Parent company Pocket Entertainment is extending the approach through Pocket Saga, a three-month-old microdrama app available only in the US. Nayak said its annualized revenue run rate is about $15 million. Unlike Pocket FM’s human-led story development, Pocket Saga’s content is entirely AI-produced, including videos derived from successful audio stories without conventional live-action production.
The company plans to enter at least two more entertainment formats in five years and hopes to license successful stories for books, television and movies. It is also discussing new capital with investors. Nayak did not deny a report that it could seek $100 million to $120 million at a valuation near $2 billion, but declined to confirm the amount or valuation.
Pocket Entertainment says it is profitable and produces positive adjusted cash flow, although it disclosed no profit, margin or cash-flow figures. It does not expect a public listing within 24 months. The financial claims remain incomplete, but the operating direction is clear: Pocket FM is using human concepts, proprietary models and listener data to expand content supply at a pace traditional audio production could not match.
