Scarcity of investment produces scarcity of return, and the two are easily – dangerously – conflated.

There is a particular kind of industry blindness that masquerades as data analysis. It goes like this: a company structures its offering in ways that constrain consumer choice, observes the constrained behaviour that results, and presents that behaviour as evidence of consumer preference. The absence of demand it has measured is demand it created – by suppressing it.

Publishing – credit where due – does this with impressive consistency.

The Storytel Illustration

Storytel has just published its annual report, and impressive reading it makes. 1.8 million titles, in 55 languages available to more than 2.6 million subscribers.

So Storytel is not the villain of this story. Quite the opposite. Under Jonas Tellander, it was arguably the industry’s most adventurous geographic expansionist – entering markets where audiobook culture was embryonic or non-existent, acquiring nascent local players, and betting that supply would generate its own demand. Some bets paid off quickly. Others were slower burns. But the underlying logic was sound, and historically validated: Storytel didn’t enter a Swedish audiobook market. It built one.

Regime change brought a different calculus.

New management, facing intensified competition and shareholder pressure, pivoted toward consolidation rather than expansion.

The result by 2026 is a revenue profile in which roughly 95% of income derives from ten core markets, with the remaining 5% spread across more than fifteen others. The late-improvers weren’t abandoned – only Russia was cut loose, and that for reasons that had little to do with commercial logic – but the expansionist chapter was effectively closed.

And herein lies the analytical trap: if you look at that 5% figure and conclude that markets outside Storytel’s core footprint lack meaningful audiobook appetite, you are not reading market data. You are reading the arithmetic of strategic withdrawal. Scarcity of investment produces scarcity of return, and the two are easily – dangerously – conflated.

The Mechanism: Supply Creates Demand

Nobody in Sweden listened to audiobooks until Storytel made audiobooks available in Sweden. Nobody wanted a smartphone until smartphones existed. Nobody wanted television before television was invented. These are not controversial observations – they are the foundational logic of product innovation. And yet publishing, an industry that prides itself on cultural adventurism and the defence of ideas, has a striking institutional resistance to applying this logic to its own market development.

Amazon’s Kindle rollout followed a similar arc to Storytel’s – one market, rapid expansion, then consolidation around the most commercially productive territories. These are rational corporate decisions. These are commercial enterprises with shareholders, and they take risks, reap rewards or absorb losses, and ultimately serve their own interests – which do not always coincide with consumer interests, and rarely with the interests of a genuinely global publishing economy.

The problem is not that companies make rational commercial choices. The problem is when those choices get reframed as evidence about markets rather than evidence about strategy.

The Pattern Is Industry-Wide

Storytel’s geographic revenue concentration is one instance of a much broader phenomenon.

Territorial rights carve the world into zones of permitted sale. The standard framework runs roughly as follows: North American English rights cover the US, Canada, and the Philippines. UK and Commonwealth rights cover Britain, Australia, New Zealand, South Africa, and sometimes India. The remaining fifty-plus Commonwealth nations – across Africa, the Caribbean, the Pacific – are typically not named, which means they fall into rights limbo: neither clearly included nor clearly excluded, and therefore, in practice, unserved.

“World rights” sounds expansive. It is frequently anything but. A publisher acquiring world rights to prevent a competitor doing so has no obligation – and often no infrastructure or intention – to serve most of those territories. The rights claim is defensive. The market development is optional.

Publishers typically push for worldwide ebook rights specifically because digital distribution makes geographic restriction difficult to enforce, and the push for world English ebook rights has at times functioned as a mechanism for consolidating control rather than a genuine commitment to global reach. The result is a structure where a reader’s ability to access a book in Ghana or Jamaica or Bhutan is determined not by whether they want to read it, but by whether the contracting publisher has chosen to serve those markets – which, in most cases, they have not.

Amazon’s Kindle store illustrates the gap between apparent universality and functional access. The store nominally covers a very large list of countries. But coverage and access are different things. For countries without a dedicated Amazon storefront, readers are directed toward the US store as the default international option – which requires a compatible payment method, the ability to transact in US dollars, and in some territories, navigating geographic blocks that prevent access to the store at all.

The Format Grab: Reserving Rights No One Intends to Use

The territorial issue is compounded by standard contract practice around format rights. A typical overreaching clause grants the publisher “an exclusive worldwide license to print, publish, and sell in all book, ebook, audio book, or in any form or format, in whole or in part, now known or hereafter invented in any language throughout the world.”

The phrase “now known or hereafter invented” is the tell. It exists not to enable the publisher to serve readers in formats not yet imagined, but to prevent anyone else from doing so. The practical consequence is that publishers secure rights to formats that don’t yet exist – as happened with the digital transition, when legacy contracts became contested territory precisely because no one had anticipated ebooks when the original agreements were signed.

The irony is complete: many publishers resisted ebooks vigorously while simultaneously having acquired the rights to them. Now the same logic is being applied to AI.

The Authors Guild has been explicit: AI training is not a publishing right. A publishing contract grants a license to publish. AI training is neither a new format nor a new distribution mechanism – it is an entirely separate right that must be separately negotiated.

And yet the industry’s kneejerk response to AI has been either to resist it entirely or to reach for it contractually, while making no corresponding commitment to deploy it for what it could actually do: translate books into languages publishers have never served, narrate audiobooks in markets where human narration would be economically unviable, make back-catalogue titles accessible across formats and territories that have been neglected for decades.

The Exclusivity Trap

Platform exclusivity adds a further layer. The arithmetic is seductive: ACX offers authors a 40% royalty rate for exclusive distribution, dropping to 25% for non-exclusive. Audible controls around 90% of the audiobook market in the US and UK, making the ecosystem difficult to ignore.

But ACX exclusivity runs for seven years, during which the author cannot sell the audiobook anywhere else in audio format. Seven years. In a market where Spotify’s entry transformed the competitive landscape within three, that is an eternity of foreclosed options.

The consumer consequence follows directly: word of mouth, a TikTok recommendation, a friend’s enthusiasm – and then nothing. The book isn’t on the platform they use. It isn’t available in their country. The format they prefer doesn’t exist for that title. The sale doesn’t happen. The reader doesn’t become a fan. The backlist doesn’t get discovered.

Non-exclusive distribution opens access to over twenty additional platforms globally – and increasingly the calculation favours wider distribution, particularly for authors in genres where discovery drives long-term growth, or in non-English markets where regional platforms significantly outperform Audible or Kindle.

The exclusivity trap operates as its own self-fulfilling prophecy. Authors choose exclusivity for the higher platform royalty. The dominant platform’s catalogue deepens, drawing more listeners. Rival platforms have thinner content, so fewer listeners join them, making them less attractive to authors. The dominant player’s dominance compounds – not because listeners and readers don’t want alternatives, but because the rights architecture makes alternatives harder to build.

This is exactly the dynamic that enabled Audible’s complacency for so long. The Scandinavian audiobook challengers were geographically contained; the competitive threat seemed manageable. Then Spotify arrived, and the industry – with notable exceptions – treated it as a punchline. A music streamer dabbling in spoken word. Charming, really.

TNPS was not laughing. In August 2020, when industry consensus was amusement, TNPS argued that Spotify’s audiobook move was a seismic shift whose ripples would take time to be felt.

Those ripples are now waves, and Audible and others are scrambling in ways that would have been unnecessary had the initial signal been read correctly.

AI and the Audiobook Gap: The Tool That Could Break the Loop

Here is a statistic that should stop every publishing professional in their tracks if somehow they have previously missed it: fewer than 5% of published books currently exist as audiobooks.

Not because demand for audio doesn’t exist. Not because the stories aren’t worth hearing.

Because the economics of traditional production have made it impossible for most titles to ever reach that format.

A professional narrator in the US charges between $200 and $400 per finished hour. A ten-hour book costs $2,000 to $4,000 to produce. For a midlist title in a major Western market, that may be commercially marginal. For a small publisher in Dhaka, Laos or Lagos, it is simply not viable. The 95% of books that never become audiobooks are not silent because no one would listen to them. They are silent because the production cost model was designed for a market that assumed professional narrators, recording studios, and publishing advances.

AI narration changes that calculation categorically, not incrementally.

The platform landscape shifted meaningfully during 2025. Spotify announced it would accept audiobooks narrated using ElevenLabs’ AI voice technology, enabling authors to narrate in 29 languages and distribute via Findaway Voices. Spotify explicitly acknowledged that digital voices could help smaller authors produce audiobooks more cost-effectively, and facilitate creation for older backlist titles that publishers would otherwise ignore.

Google Play Books has supported AI narration longer still. Kobo Writing Life accepts it without restriction. As of 2026, ACX still rejects AI-assisted narration for its users, despite Audible itself exploring AI assisted narration and translation. The direction of travel across the industry is now unambiguous, whatever the pace of individual platforms and the self-serving limitations.

The economics are not marginal. They are transformative. An ElevenLabs Creator Plan costs $22 per month. Even for a long book requiring extra credits, total production typically comes in under $80. That is not a 10% cost reduction. It is the difference between a production model that excludes the vast majority of the world’s authors and one that opens up to many of them.

ElevenLabs already enables audiobook creation in over 90 languages – not the handful of European languages that traditional production economics made viable, but ninety. And counting.

A book published in Yoruba, Twi, or Mandinka has never had a realistic path to audio distribution through traditional means. AI narration and AI translation together represent the first time in publishing history that the gap between “written in a language” and “available in audio across platforms” can be closed across countless languages without requiring significant capital or a major publisher’s backing.

This is precisely why the industry’s dominant response – treating AI primarily as a threat to be resisted, a rights question to be litigated, or an aesthetic compromise to be disclosed and apologised for – misses the larger story entirely.

The rights framework examined in this piece was built to control access in a world of physical scarcity. Printing, distribution, and territory made exclusion the default, and inclusion the exception that required economic justification. Digital eliminated the scarcity of distribution. AI is now eliminating the scarcity of production. When both barriers fall simultaneously, the self-fulfilling prophecies that have sustained publishing’s geographic and linguistic concentration lose their structural support.

The question is whether the industry’s established players with vested interests in the status quo will recognise this before the vacuum is filled by those who are moving with the times, not on the shore trying to hold back the tide.

That question is not rhetorical. Spotify was laughed at in 2020. It is not being laughed at now.

The publishers and platforms that treat AI narration as a concession to cost-cutting rather than as infrastructure for a genuinely global audio publishing economy are making the same category error that has run through this entire piece: measuring present supply and calling it the ceiling of demand.

Markets are not found. They are made. The 95% of books that have never been heard are not a data point about listener appetite. They are an opportunity of historic proportions – and for the first time, the tools to act on it are affordable to almost all.

This post first appeared in the TNPS LinkedIn Analysis Newsletter.