Spotify's Audiobooks Weren't a Content Bet. They Were a Legal Key.

On September 20, 2022, Spotify opened a US bookstore. You could browse more than 300,000 audiobooks, but you bought them one at a time, at full price, on Spotify's website — no subscriber discount, no Premium perk, nothing folded into the monthly fee.¹ Fourteen months later, the bookstore was gone. In its place: 15 free hours of listening every month, baked straight into the Premium plan most subscribers already paid for.⁴ Same product, same catalog, opposite business model. The reversal is usually told as Spotify fixing a flop. It is better told as Spotify finding a key.

The official story is that à la carte audiobooks didn't sell, so Spotify made them a Premium benefit to drive engagement and value. That story is true and also beside the point. The pivot's biggest consequence wasn't on the listening side of the ledger. It was on the paying side — specifically, what Spotify pays the people who write the songs.

The line in a 2022 settlement that nobody outside the room cared about

To see the move, you have to look past the books and at a piece of regulatory plumbing called Phonorecords IV — the 2022 settlement between US music publishers and streaming services that sets the mechanical royalty rate songwriters earn. Tucked inside it was a carveout: a service that bundles music with a separate, distinct product can pay a lower mechanical rate, on the logic that subscribers aren't paying purely for the music anymore.⁶ This wasn't some loophole Spotify smuggled in. The carveout covered a dynamic already present in the market: Amazon and Apple both offered music alongside other services in multi-product bundles, and both companies were co-signatories to the Phonorecords IV settlement that codified the bundle treatment. The mechanism pre-existed. What was missing was a standalone music streamer willing to manufacture the 'distinct product' and walk through the door.

Audiobooks were that distinct product. Once 15 hours of audiobook access lived inside Premium, the plan was no longer a pure music subscription — it was, in the language of the settlement, a bundle. And a bundle pays less.

Read that carefully, because almost everyone gets it wrong. Spotify did not cut the per-stream rate in half. It cut the revenue base that the rate is applied to — the pool of money declared as music revenue — by almost half, by recategorizing the Premium plan as a bundle effective March 1, 2024.⁵ Same dollar coming in from the subscriber; far smaller slice now counted as the thing songwriters get paid on. The math runs downstream from a reclassification, not a rate cut. That distinction is exactly why the move survived a courtroom.

Why the cheaper version was the whole point

Here is the causal chain worked all the way down. Spotify bought the audiobook distributor Findaway in mid-2022 for €117 million, with another €13 million tied to keeping employees around.² That gave it a catalog. The catalog let it stand up a real, substantial audiobook offering — by the US Premium launch, more than 200,000 titles.⁴ A real offering is what makes the bundle defensible: you cannot dress up a token freebie as a 'distinct product' and expect a court to nod along. Then Spotify flipped the plan to a bundle, the reported music-revenue base dropped by nearly half, and the savings began to compound across roughly 99% of US Premium subscriptions, which got swept into the bundle classification.⁸

And the financials moved in the direction the thesis predicts. Spotify's gross margin climbed to 32.2% in Q4 2024 and then to a record 33.1% in Q4 2025, with the company crediting Premium gains from audiobooks for easing content-cost pressure.¹¹ 'Content cost' is, in large part, what it pays rights holders. The free books didn't just entertain subscribers. They re-engineered the company's single largest expense.

Isn't this just a content strategy that happened to save money?

The fair objection is that I'm being too cynical — that Spotify genuinely wanted to win audio beyond music, audiobooks are a real growth lane, and the royalty savings are a happy side effect of a sincere product bet. There's truth in it. Spotify did spend real money on Findaway, did build a 200,000-title catalog, and audiobooks plausibly do lift engagement. But sequence and substance betray the priority. The à la carte version came first and was quietly abandoned; the version that survived was precisely the one that triggered the carveout. You don't accidentally redesign your flagship plan into the exact legal shape of a settlement clause and then save nine figures by coincidence. The court itself underscored the substance: Judge Analisa Torres found audiobook streaming had 'more than token value,' which is the finding that lets a bundle be a bundle.⁷ Spotify needed the books to be real for the arbitrage to work. That the books are real is not evidence against the arbitrage — it is the arbitrage's load-bearing wall.

The courtroom that turned the play permanent

On January 29, 2025, the gamble paid off in the only venue that mattered. Judge Torres dismissed the MLC's lawsuit with prejudice — not on a technicality, but on the merits — ruling the regulations 'unambiguous' and Spotify's bundle classification 'the only plausible application of the law.'⁷ With prejudice means the MLC can't refile the same claim in that court. The reclassification went from contested to blessed. What had been an aggressive accounting interpretation became settled ground Spotify can now build on indefinitely.

The fight isn't over. Even before the court ruling, in June 2024, the NMPA had taken the dispute to the FTC, arguing the bundle reclassification shortchanged songwriters — proof that a court win doesn't end a war over who gets paid.¹² But the strategic verdict is already in. Spotify spent over a hundred million dollars on a book company, ran a product into the ground, reversed course inside fourteen months, and emerged with its largest expense permanently lighter. The audiobooks were never the asset. They were the lever — and Spotify pulled it not in the catalog, but in the classification.