The global reading-and-listening population runs to over six billion people. Nearer 6.5 billion. Is four million new titles really a problem?

The Numbers We Prefer Not To Count

There are only three certainties in life: death, taxes, and TNPS doing the real arithmetic as the trade panics about a latest statistic – this week the four million new titles Bowker registered last year, which apparently heralds the end of civilisation as we know it.

Since January 2020, Amazon’s Kindle Unlimited has paid out a cumulative $2.58 billion directly to self-publishers and small presses. Not to the Big Five. Not to literary agents. Not to the assorted gatekeepers who are busy lamenting the sorry state of reading. To independent authors. In cash. Every month. Without fail. And growing.

The monthly payout in January 2020 stood at $28.2 million. In February 2026 it stood at $58.3 million. That is a 107% increase in just over six years. The format that plateaued. The format some in the industry buried at approximately the same time it buried the CD – prematurely, and for reasons that have always had more to do with narrative convenience than market reality.

Six Years. Six Records.

Let’s run through the annual scorecard for what Amazon has lately been paying to self-pub and small-pub, keeping in mind this is solely from the unlimited ebook subscription service Kinde Unlimited, not the main à la carte Kindle store or the Amazon books store istelf.

2020 delivered $379.8 million across the year – $31.65 million a month on average. That was the baseline. The floor. The pandemic year in which, incidentally, people were locked indoors and discovered (or rediscovered) that digital reading was rather convenient.

2021: $450.1 million. Up 18.5% on 2020.

2022: $511.4 million. Up 13.6%.

2023: $578.6 million. Up 13.1%.

Every single year, growth. Not explosive, not viral, not the kind of number that generates breathless LinkedIn posts about disruption. Just steady, relentless, compounding growth of the kind that actually sustains an industry.

2024 came in at $639.6 million – up 10.5% – a figure I reported on here at TNPS when it exceeded my own mid-year projection by a comfortable $10 million.

And 2025 pushed on further to $711.3 million, up 11.2%, at an average of $59.3 million per month.

Two months into 2026, and the monthly average has already ticked up to $60.25 million.

Six years. Six annual records. $2.58 billion in total.

A Measurement Error, Not A Market Truth

At this point in the annual ritual, someone in the trade press will inevitably remind me that KU is “just” self-publishing – Penny Dreadfuls! – and even worse, unlimited subscription. Heresy! This is therefore not remotely representative of the broader ebook market. Something to mutter and vent about in private, rather than risk giving oxygen to the notion of successful self-publishing. At the same time, the occasional outlier self-published author gets industry acknowledgement and applause.

Which actually makes the point. This is a category error dressed up as editorial discernment.

Kindle Unlimited is not a niche footnote. It is a subscription reading ecosystem with millions of members, paying Amazon monthly for access to a catalogue that dwarfs anything trade publishing has assembled.

The self-publishers and small presses it rewards are producing content readers are actively choosing to read, month after month, at volumes the mainstream industry’s data collection apparatus is institutionally incapable of measuring – because measuring it would complicate the story.

And the KU payout, it bears repeating, is only one slice of the invisible ebook economy. It covers KDP self-publishers and small presses operating through Kindle Direct Publishing. It does not include Amazon Publishing’s own imprints. It does not include the KU payout to larger publishers who participate in the programme on different terms. It does not include à la carte ebook sales from any of those players. Add OverDrive’s 379 million library ebook borrows in 2025 – digital text still outpacing audio by 20% in the zero-cost-friction environment of public libraries – and the picture becomes not just inconvenient for the plateau theorists, but embarrassing.

The ebook plateau is a measurement error. It always was. The market didn’t plateau; the industry’s collective willingness to count simply ran out at the point where the numbers stopped telling the right kind of story.

Four Million Books. The Horror.

Into this already uncomfortable picture arrives Jim Milliot‘s report in Publishers Weekly, noting that US book output topped four million titles in 2025 – a 32.5% jump over 2024. Self-published print and ebooks led the surge, rising 38.7% to 3.5 million titles. Traditionally published output, for the avoidance of doubt, grew a healthy 6.6% to 642,242 titles. Nobody, it seems, is actually publishing fewer books.

The industry response was, broadly, horror. “Terrifying,” said one commenter on LinkedIn. “Madness,” said another. The words “AI slop” appeared with the frequency of a nervous tic.

Let’s engage with this seriously for a moment, because it deserves more than an eye-roll – though it also definitely deserves an eye-roll.

Yes, AI-assisted content will be part of that 3.5 million. Probably a meaningful part. That is a legitimate conversation about quality signals, discoverability, and metadata integrity. And let’s keep in mind here that AI-assisted is not the same as AI-created.

But the hysteria goes well beyond that reasonable concern into something more revealing: the assumption that four million books is, in itself, a problem. That the right number of books is approximately however many traditional publishers decided to produce this year, plus a polite rounding error from indie authors who know their place.

Sorry, guys, but this assumption does not survive contact with arithmetic.

The United States alone has a population of over 330 million people. Is four million new titles really a problem? If so, for whom?

Let’s take a step back and revisit this from the other side.

The US has some 260 million people of reading or listening age. That is to say, of the 330+ million US population, roughly 260–265 million are aged 6 and above – old enough to read independently or follow an audiobook with comprehension.

If we apply the US literacy rate (around 79% of adults reading at a basic level or above, though functional literacy is higher among younger cohorts), we still land at well over 200 million capable readers.

For audio, the threshold is essentially anyone who can follow spoken English with comprehension – which takes us back up to that 260 million figure, and higher still if we include non-English speakers engaging with content in their own languages.

Now twenty, fifteen, even ten years ago, there was no meaningful audiobook market. Today, audiobooks are the new black. Almost every publisher has an audiobook strategy, and production costs are spiralling down as demand spirals up.

In 2026 we are, in the USA alone, looking at 260 million potential readers, potential audiobook listeners, potential subscribers – the vast majority of whom the traditional publishing industry is not reaching and has never seriously attempted to reach, because its commercial model was built around a much smaller, self-selecting audience.

The idea that this population is suffering from an excess of available books and will be confused by 4 million new titles is not a market observation. It is a category error from an industry that has consistently confused its own customer base with the full extent of human demand.

And then there’s the international market. The global reading-and-listening population runs to over six billion people. Nearer 6.5 billion. Four million new titles is not a problem. It’s not even a rounding error.

The idea that we are somehow drowning in too many books is a supply-side anxiety being projected onto a demand-side market that has given absolutely no indication it agrees.

KU’s millions of subscribers navigate a catalogue of millions of titles and somehow, month after month, manage to read something.

OverDrive’s library users borrowed 379 million ebooks last year and appear not to have collapsed under the weight of excessive choice.

Readers, it turns out, are rather good at finding books they want to read when the tools and the access are there.

The self-publishing ecosystem has built those tools. Bowker‘s own product marketing manager Andrew Kovacs notes that every stage of the publishing process – writing, editing, design, distribution, marketing – is now accessible to independent authors at a quality level previously reserved for those lucky enough to secure a traditional contract. That is not a crisis. That is democratisation. It is also, of course, precisely what is driving the KU payout numbers that the industry prefers not to count.

Who Benefits From The Myth?

The “ebooks are declining” narrative benefits those whose market position depends on print’s primacy: traditional publishers with physical distribution infrastructure, retailers who need footfall, industry bodies whose membership dues come from legacy players, and the conference circuit that services them.

None of which, I hasten to add, makes these players villains – just people acting predictably within structures that reward a particular version of the truth. Narratives that serve interests tend to outlast the evidence against them. The publishing trade is not unusual in this regard – it is simply more genteel about it than most.

What is less forgivable – what is the problem, to be blunt – is the cumulative effect on the broader conversation about who reading is for, and where it is going.

Every commentator that declares the ebook experiment over is a commentator that ignores the millions of readers – many of them in markets the Anglo-American trade has historically underserved – for whom digital access is not a lifestyle preference but a basic condition of literacy participation.

Every trade report that declines to cover the self-publishing numbers while parading the latest mainstream statistics as headline news is a report that renders invisible the reading lives of the people most likely to benefit from affordable digital access.

The Attention Span The Industry Forgot It Had

There is one final irony worth naming. Much of the alarm about four million titles is framed around discoverability – the fear that readers, their attention already fractured by competing digital stimuli, will simply give up and stare at their phones instead.

This is the industry that collectively wrings its hands about diminishing attention spans, while trade reportage column inches grow ever shorter. Meanwhile, readers – those poor, distracted, scroll-addicted readers – are watching prestige television series that run to twelve episodes of seventy minutes each, reading fantasy novels that top a thousand pages, and borrowing 379 million ebooks from digital libraries in a single year.

Attention has not diminished. It has fragmented and diversified. But it is as strong as ever. We’re just upset that their attention is not focussed on us.

The 2026 Trajectory

Going full circle to wind up this slightly more than industry length essay (too log? Try reading a book.), and based on the January and February 2026 data – averaging $60.25 million per month – a full-year Kindle Unlimited pot payout in the range of $720–740 million is a reasonable projection, assuming the pattern of recent years holds.

That would represent modest further growth on 2025’s $711.3 million: a slight deceleration in percentage terms, which is what you would expect from a larger base, but growth nonetheless.

The demand for ebooks – self-published, subscription-based, statistically invisible to the trade – keeps growing. The number of books keeps growing. The number of readers keeps growing. The industry keeps not noticing.

And TNPS keeps counting.

This post first appeared in the TNPS LinkedIn Analysis newsletter.