This audio company doubled from $250M to $500M ARR after going all-in on AI.

Their shift to AI production increased annual content output from roughly 35,000 hours to 2.5 million hours. Thats ~400,000 new stories.

to get the feel for that scale, Netflix makes about 1,700 hours of original TV annually, so this is nearly 1,500x of that.

PocketFM is a storytelling app. 10-minute episodes, every episode ends on a cliffhanger. It got 200M listeners spending 135 minutes a day on it.

Two years ago they hit a wall. Demand was there. Monetization worked. They just couldn't produce stories fast enough. Writers were the bottleneck.

So instead of hiring 1,000 more writers, they did 2 things:

Partnered with ElevenLabs and cut audio production costs by 90%. 30,000+ hours of content, made for a fraction of what it cost before.

Built SOTA AI Fiction writer trained on their own data. Not the internet. 100M+ hours of episode-by-episode listener behavior. Where people drop off, where they pay, which cliffhanger converts.

Revenue went from $21M -> $500M in <3 years

Everyone thinks AI content makes stories worthless. Infinite supply, price goes to zero. And for most companies that's true.

But for them, the hard part was knowing what people will actually still be listening to and paying for. That's not on the internet. GPT didn't train on it.

AI with proprietary data turns you into a machine that gets better every time someone hits play. PocketFM pivoted to AI trained on 7 years of what actually works.

The moat was the data about which stories work.