Pocket FM, the Under-the-Radar AI App with $500M in Revenue

The chart published by @RohanNayak2 illustrates this trajectory: an annualized run rate that starts from nearly zero around August 2022 and reaches $506M in August 2026. The final section of the curve, beyond roughly $200M, is highlighted because it corresponds to the return to growth following the bet on AI.

The $500M figure is impressive, but it is not recognized annual revenue

The first clarification came from a very simple question by @davefontenot:

"ARR or annual revenue run rate?" @davefontenot

Rohan Nayak’s answer is unambiguous:

"annual revenue run rate" @RohanNayak2

This distinction matters. An annual revenue run rate takes the current revenue level, generally monthly, and projects it over twelve months. It indicates the pace of the business at a given point in time. It does not automatically mean that Pocket Entertainment has already recognized $500M in revenue over the past twelve months, nor that all of this revenue is contractually recurring.

This does not diminish the operational scale of the achievement. On the contrary, for an audio fiction app funded by pay-per-episode purchases and advertising, reaching a roughly $500M run rate implies a very real ability to acquire an audience, retain it, and then convert its attention into revenue.

The post also compares Pocket Entertainment with groups such as Netflix, Spotify, and Disney. The published table lists $500M in revenue, a 70% gross margin, and 100% annual growth for Pocket Entertainment, under the label “Content & IP + Consumption + AI.” These figures come from the company’s visual. They should therefore be read as metrics presented by Pocket, not as an independent financial analysis.

The real engine: a 0.25 percentage-point increase in CTR can trigger a growth loop

The most important part of the story is not AI-driven production itself. It is the mechanism that turns better creatives into more cost-effective acquisition.

Rohan Nayak explains that Pocket FM spent more than $100M on acquisition before identifying its core insight: increasing CTR from 2% to 2.25% would reduce CAC by approximately 30%.

"If the clickthrough rate (CTR) for an ad goes from 2% to 2.25%, our customer acquisition cost (CAC) decreases by ~ 30%." @RohanNayak2

Mathematically, a CTR increase from 2% to 2.25% represents 12.5% more clicks for the same number of impressions. The calculation is simple: 0.25 divided by 2 equals 12.5%. But CAC does not necessarily follow a linear relationship with this increase.

In Meta or TikTok ad auctions, a creative that attracts more clicks and generates better quality signals may receive more distribution, win more auctions, or turn the same media spend into more installs. This leverage effect may explain why 12.5% more clicks translates, according to Pocket FM, into a nearly 30% reduction in CAC.

The CTR improvement only creates value because it is connected to every other part of the machine.

The chain described by Nayak works as follows:

1. LLMs identify the most intriguing passages in a work of fiction.
2. These excerpts are combined into a longer promotional script.
3. Pocket FM runs a video trailer of around 90 seconds to sell an audio story.
4. The opening seconds must include a hook roughly every five seconds.
5. The creative must end with a cliffhanger strong enough to trigger a download mid-scroll.
6. If the ad fails to exceed the targets, notably a 2.5% CTR and a 55% three-second view rate, a creative director steps in.
7. Winning creatives make it possible to increase the acquisition budget without materially worsening CAC.

Pocket FM claims to produce around 1,000 ads per series, for a total of approximately 17,500 ads per month. According to the post, this creative factory made it possible to increase the acquisition budget by seven to eight times without a material increase in CAC.

That is where the playbook lies. More creatives mean more tests. More tests make it possible to identify more effective hooks. These winning hooks improve auction performance. Better-controlled CAC then enables more spending, which in turn funds further testing.

@Freyabuilds’s response sums up both the admiration for this mechanism and its main potential weakness:

"The 0.25-point CTR bump cutting CAC ~30% is the unsexy math that actually compounds. Still wondering how long that 2.25% bar holds once everyone else starts stuffing a cliffhanger every five seconds." @Freyabuilds

Rohan Nayak did not answer the question. Yet it is fundamental: the cliffhanger is a narrative convention that is easy to copy. If every competitor applies the same formulas, their ability to stop the scroll may erode. Audiences may also tire of this escalation.

Pocket FM’s advantage may therefore lie less in the cliffhanger itself than in its ability to test, produce, and iterate on thousands of them before anyone else. However, this execution advantage will have to be continually renewed.

Pocket FM does not monetize a catalog; it monetizes the desire to know what happens next

Pocket FM’s model differs markedly from that of a traditional Netflix-style subscription. The company is not primarily seeking to sell unlimited access to a catalog. It structures its product to sell at the moment when narrative desire is at its peak.

The format adopted after several experiments is the mobile audio drama: highly serialized 8- to 12-minute episodes with strong hooks and cliffhanger endings. Nayak claims that after finding this structure, average daily listening time rose from around 25 minutes to more than 150 minutes.

Monetization follows this rhythm exactly:

• users receive free listening minutes each day;
• the next episode can be unlocked daily to build a habit;
• once the free minutes are used up, users can pay per episode;
• they can also watch an ad to unlock the next installment.

This pay-per-episode monetization therefore occurs not at the beginning of the experience, but once the narrative tension has already been established. The cliffhanger serves a dual purpose: it drives downloads in the ad, then encourages a purchase or rewarded ad within the product.

Pocket FM reports that ad revenue grew from zero to an annual run rate of around $90M in twelve months. Advertising is therefore not a secondary option for non-paying users. It is an integral part of the revenue architecture, while keeping users engaged in the story rather than excluding them when they refuse to pay.

Discovery is also tailored to audio-listening behavior. Users can start a program, lock their screens, and stop browsing. Pocket FM says it uses episode playlists and recommends a new series when a program’s free minutes run out. The goal is not merely to recommend content, but to preserve the listening flow and redirect attention toward a story that can still be monetized.

AI unlocked capacity, but shifted the problem to quality

The most spectacular part of the post concerns the shift to fully AI-assisted production in mid-2024. According to Nayak, Pocket FM accepted roughly six months of stagnation at around $200M in ARR in order to transition its production pipeline, before accelerating to its current run rate.

The AI pivot is presented not as a simple cost-cutting measure, but as a bet on radically greater production capacity.

This decision caught the attention of @soy_adnan3:

"Flatlining at $200M after the AI only pivot... then adding $250M the next year is the rare case where the ugly six months actually paid off!!" @soy_adnan3

Rohan Nayak did not respond directly to this message, but he confirmed the rationale behind the decision in a reply to @kirbyman01:

"Read it multiple times. Very insightful + courageous to take that huge flatlining hit, moving over to 100% AI production. Respect 🫡🫡" @kirbyman01

"Thanks @kirbyman01. It was a tough call at the time - short-term pain, but the bet was that it'd unlock much faster growth long-term. Really glad it worked out." @RohanNayak2

The thread therefore supports the idea of an accepted transition cost. However, it provides neither detailed financial statements nor an independent methodology for isolating what is genuinely attributable to the shift to AI, geographic expansion, changes in marketing spend, or a change in product mix.

A later viral summary by @VadimStrizheus expands on the case for founders. In particular, it mentions production increasing from 25,000 to 2.5 million hours per year, LLM-as-a-judge systems, and an increase in twelve-month retention from 44% to 76%. These figures are consistent with the thesis of a change in scale, but they do not appear in the Rohan Nayak post provided here. They should therefore not be confused with claims directly documented by the CEO.

The central risk was raised by @svpino, but received no response:

"I haven't used Pocket yet, but I wonder whether you have some sort of manual filter to weed out bad AI-generated slop. Do you publish everything you generate, or do you curate it somehow?" @svpino

This is probably the thread’s most serious blind spot. If AI shifts the bottleneck from production to selection, Pocket FM must address editorial quality, audience trust, and discoverability. Producing ten times more content is worthless if users encounter generic, repetitive, or poorly calibrated content before finding a good series.

Localizing a story does not mean translating it

Rohan Nayak states that 78% of Pocket FM’s revenue remains concentrated in the United States. This concentration makes localization strategically important, even if the advertising-based acquisition engine appears exportable.

In his post, Nayak clearly distinguishes translation from adaptation: a story designed for a Spanish audience may lose its references, jokes, folklore, and tone if it is merely translated for a Norwegian audience. Pocket FM therefore says it retains the backbone of a story while adapting it to local references.

This idea qualifies another ambitious statement in the post: what enabled Pocket FM to grow from zero to $400M in ARR in the United States “works in every country.” The thread helps clarify this promise.

The acquisition engine may be exportable, but the content is only exportable after cultural adaptation.

Pocket FM specifically cites growth to $25M in the United States in 19 months, $21M in Germany in 11 months, and $10M in France in three months. @VadimStrizheus’s summary adds that a localization approach reportedly improved retention by 50% in Germany, but this additional figure does not appear in Rohan Nayak’s post. It should therefore be treated as data relayed by the summary, not as a metric confirmed in the source thread.

ElevenLabs, Pocket Saga, and the attempt to replicate the machine in video

The post and replies show that Pocket Entertainment does not intend to remain solely an audio company. In response to @shiri_shh asking about entering the short-form video drama market, Nayak replies:

"are you guys also planning to enter short form video drama?" @shiri_shh

"Yes we just launched our AI micro-drama app - Pocket Saga!" @RohanNayak2

Pocket Saga is therefore an attempt to apply Pocket FM’s model to AI-generated or AI-assisted video micro-dramas. In another reply, Nayak also presents Pocket Saga as a new anime-focused app.

The partnership with ElevenLabs is also explicitly confirmed. After congratulations from @mati, Nayak replies:

"Congratulations Team! Incredible!" @mati

"Thank you @mati. Been great partnering with you and the @ElevenLabs team along the way." @RohanNayak2

The ElevenLabs Pocket FM partnership is therefore established by this reply. However, @VadimStrizheus’s summary claims that audio costs fell by 90% thanks to ElevenLabs. This percentage does not appear in either the post or Rohan Nayak’s reply. It cannot be presented as a figure confirmed by Pocket FM in this conversation.

Similarly, @VadimStrizheus attributes figures to the Pocket FM case such as “from $0 to $500M in four years,” “ten pivots in two years,” “135 minutes per day,” models trained on more than 100 million hours, as well as 300,000 authors, 90% of whom are beginners, and $33M paid out. Some are close to Rohan’s account, which mentions ten formats tested over two to three years and more than 150 minutes of daily listening. Others do not come from the source post and must remain clearly separate.

The LTV/CAC ratio makes the acquisition strategy credible, without dispelling every doubt

Paid acquisition at a very large scale only makes sense if the revenue generated by a customer consistently exceeds their acquisition cost. @RatPoisonaut asked the question that tests this assumption:

"Impressive! What is your LTV:CAC ratio?" @RatPoisonaut

Nayak’s response provides the only direct measure of unit economics in the discussion:

"We are trending at ~3x LTV/CAC at the moment" @RohanNayak2

An LTV/CAC ratio of approximately 3x makes the narrative more credible: for every dollar spent to acquire a user, Pocket FM estimates that it generates about three dollars in lifetime value. This level can justify aggressive acquisition, especially if the reported gross margins are high.

But this information does not settle everything. The thread does not specify:

• the exact definition of LTV;
• the cohorts concerned;
• the CAC payback period;
• the share of advertising revenue included in this LTV;
• AI infrastructure and content costs;
• the sustainability of EBITDA profitability during international expansion.

The figure is useful, but it remains a self-reported, aggregated metric.

What founders can learn, and what Pocket FM still needs to prove

@VadimStrizheus’s summary outlines four lessons that align with Pocket FM’s account, provided that his additional figures are not conflated with Rohan Nayak’s statements:

1. Do not abandon a pivot too quickly. The plateau at around $200M shows that a transition can hurt the trajectory in the short term before creating new capacity.
2. Charge at the point of maximum intent. Per-episode payment at the cliffhanger is more than a pricing mechanism: it is a way to sell at the peak of narrative desire.
3. Apply AI to the bottleneck. At Pocket FM, AI is used to create and test enough content and ads to industrialize learning.
4. Turn users into the supply side. The thesis of individual studios and an expanded content offering aims to turn the platform into a production system, not merely a distributor.

The most useful lesson, however, is more specific: AI only becomes an advantage when it is connected to a complete economic loop. Pocket FM does not win because it uses LLMs or synthetic voices. It seeks to win by connecting the story that drives clicks, the video ad that drives installs, the cliffhanger that drives retention, and the per-episode payment that drives monetization.

This loop sums up its growth strategy: better creatives improve CTR, better CTR can reduce CAC, better-controlled CAC makes it possible to increase acquisition spending, and better retention increases users’ lifetime value. Per-episode monetization and rewarded advertising then turn that attention into revenue.

However, the thread leaves several questions unanswered:

• What human or automated filtering removes the low-quality AI content flagged by @svpino?
• Will the 2.25% CTR threshold remain defensible if competitors copy the same hooks and cliffhangers?
• How does Pocket FM calculate its approximately 3x LTV/CAC, and over what time frame?
• How much of the post-plateau acceleration actually comes from AI?
• Does EBITDA profitability remain robust once content, acquisition, infrastructure, and international expansion costs are factored in?

Pocket FM has demonstrated that a media company can become a performance machine. What it still needs to demonstrate is that this machine can preserve genuine editorial quality and a creative advantage once the entire industry has access to the same AI tools.