For Pocket FM, artificial intelligence is no longer an ancillary feature in the publishing pipeline: it has become the infrastructure through which the platform produces, expands, and distributes its catalog of serialized audio stories. The Indian company stated that it has reached an annualized run rate of $500 million, double the roughly $250 million reported a year ago. The figure is not equivalent to revenue already generated in the financial year; it is calculated by annualizing current monthly revenue. Yet it illustrates the growth pace of a service that has tied its expansion to assisted content generation.
According to Rohan Nayak, co-founder and CEO of Pocket FM, 93% of the overall catalog is now produced with the support of AI, and that share rises to 99% when looking at new titles. Behind that percentage, however—at least according to the company's description—there is no system autonomously inventing series and characters: human writers remain responsible for ideas and narrative structure, while models step in to turn creative material into ready-to-publish products at scale.
The distinction is important because Pocket FM's value is built on very long stories, organized into episodes and designed to retain listeners over time. For a platform of this kind, having a synthetic voice or a text generator is not enough: it requires storylines capable of sustaining hundreds of episodes, diverse genres, and continuity between episodes. The company therefore claims a hybrid model, in which the creator defines the core of the work and proprietary tools accelerate writing, adaptation, and voice synthesis.
The economic advantage lies in scale
The primary impact of the transformation is on costs. Nayak claims that AI has made production roughly 80 times less expensive. A 100-hour project, which previously could take approximately a year of work, can now be completed in a single day. These are company claims, not independently verified data, but they help explain why AI has taken on such a central role in Pocket FM's operating model.
A reduction of this magnitude changes the economics of serialized audio. Traditional audio fiction production requires scriptwriters, voice actors, studios, direction, editing, and revisions; every new episode adds labor hours and costs. Automating a significant portion of the pipeline, on the other hand, makes it possible to test more genres, publish more frequently, and support much larger catalogs. The opposite risk naturally remains: abundance can lead to less consistent quality or make it harder for the best titles to emerge.
Pocket FM claims to work with over 550,000 creators, who generate around 2.5 million hours of AI-powered content per year. Just two years ago, the platform’s entire catalog stood at approximately 100,000 hours. Today, the library exceeds 770,000 audio series. While the numbers do not automatically indicate the success of every single work, they show how dramatically the company has expanded its production capacity in a short time frame.
The most significant commercial indicator, according to Pocket FM, is the improvement in 12-month revenue retention, which climbed from 44% to 76% over two years. Management explains that a broader catalog increases the likelihood of finding stories tailored to different tastes and habits. In other words, AI is framed not merely as a cost-cutting lever, but as an indirect personalization tool: more available titles mean more opportunities to retain listeners.
A business built on unlocked episodes
The 500 million milestone follows a rapid progression: the run rate had already climbed to $430 million in April, up from roughly $250 million twelve months earlier. Of the current total, around $85 million comes from advertising, while the dominant share, estimated at around $415 million, comes from users paying to unlock individual episodes. It is a model distinct from the flat subscription typical of many entertainment services: revenue is tied to continued listening and the story's ability to drive audiences to the next chapter.
The company states that 96 titles have surpassed one million dollars in individual revenue and that 13 have exceeded 10 million. These figures suggest that the catalog is not merely composed of long-tail productions, but also includes series capable of becoming significant editorial products. For those creating AI-generated content, this is perhaps the most delicate test: generating massive volumes is not enough; one must build intellectual properties that audiences consider worthy of following and paying for.
Founded in India in 2018, Pocket FM is now present in over 20 countries and claims more than 250 million listeners. The United States has become its most important market: growth in the country stood at around 70% over the past year, with the US market accounting for roughly 70% of the annualized run rate. The expansion has also included the United Kingdom, Germany, and France, while in the United States, the platform introduced user-generated content.
The transition from Indian roots to a heavily American revenue base highlights a central aspect of the strategy. Pocket FM is not just aiming to export an app, but to make a data- and automation-intensive editorial process replicable across different markets. The company has developed proprietary models for creative writing and text-to-speech, trained on data accumulated over years of production and on listener engagement signals. The quality of the system, therefore, also relies on the ability to determine which stories sustain attention and at what point users drop off or keep listening.
From audio storytelling to video microdramas
The project does not stop at audio. Parent company Pocket Entertainment launched Pocket Saga, a microdrama app currently available only in the United States. Three months after launch, the service has reportedly reached an annualized run rate of approximately 15 million dollars. Unlike Pocket FM, Pocket Saga offers content entirely generated by AI: there is no traditional live-action production, and several stories that were successful in audio format are being transformed into synthetic videos.
This second initiative is a more radical testing ground. While Pocket FM preserves an explicit role for humans in the ideation and narrative phase of audio, for microdramas the company claims a fully automated pipeline. The short, vertical format, already widespread across episodic fiction apps, lends itself to rapid industrial production; it remains to be seen whether reduced turnaround times and costs will be matched by enough narrative staying power to convert casual viewers into recurring customers.
For the digital media sector, the Pocket FM case highlights a concrete dynamic of generative AI: the shift does not concern only the finished work, but above all the economic structure that determines how many works can be funded, tested, and distributed. Competitive advantage, however, will not rely solely on the availability of models. As text-to-speech and text generation become more accessible, what will matter is the quality of proprietary data, relationships with creators, editorial curation, and the ability to build recognizable titles.
Pocket FM must now prove that its run rate growth and leap in productivity can consolidate without impoverishing its catalogue. The platform has already signaled its direction: using automation to multiply supply, keeping writers at the heart of audio stories while experimenting with full automation in short-form video. It is an approach that could become a benchmark for other entertainment services, but one that brings open questions regarding quality, creative compensation, and the sustainability of a market increasingly crowded with synthetic content.
