The French translation pivot, the line closure, and now the Dashverse and Toonstar deals, are all consistent with a corporate direction that treats AI not as a threat to manage but as infrastructure to deploy.
As March concluded, Harlequin announced a multi-year agreement with Dashverse, an AI-native entertainment company, to co-produce 40 animated microdramas based on its romance titles.
Within hours, the literary internet had its verdict. “Add these new book-inspired microdramas to the piles and piles of AI-generated products that no one actually wants,” wrote Literary Hub, before dismissing the format as content that “authors didn’t ask for in the first place.”
Nobody asked? Eighty million paying monthly users globally would beg to differ. But we will get to the numbers.
What the snap reactions missed is that the Dashverse deal is not a standalone announcement. It is the fourth act of a strategic play that Harlequin and its parent company HarperCollins have been running quietly for the better part of twelve months.
Read in isolation, each move looks like a minor trade story. Read together, they form something considerably more significant: a blueprint for how a legacy publisher restructures around AI infrastructure, and a signal about where the rest of the Big Five are likely to follow.
HarperCollins of course owned by News Corp, which makes no secret of its position on AI. Expect PRH to be not too far behind. Both parent companies are clear that AI will be part of their business strategies.
In May 2024, News Corp signed a multi-year partnership with OpenAI worth c. $250 million over five years. In March 2026, they announced a three-year deal with Meta, allowing Meta to use News Corp content for AI training. Last year, PRH owner Bertelsmann described AI in a policy brief as “normal technology“.
Embracing AI publicly remans a bridge too far for most publishers, but as HarperCollins is demonstrating, that doesn’t mean it isn’t happening.
This same week, HarperCollins has announced a second AI-microdrama partnership, this time with Toonstar. Per Fast Company, the first series planned is Friendship List, based on YA books by Lisa Greenwald. Per FC, “the series will be accompanied by a graphic novel from HarperAlley – HarperCollins’s graphic novel imprint for kids and teens – based on the animation.”
In other words, a two way street where text begets AI begets new text. Not quite the AI apocalypse story being painted in some quarters. But let’s step back and look at the bigger picture.
Act One: The Translators (November 2025)
It began quietly in November 2025, when Harlequin contacted the translators working on its French ‘Azur’ series to inform them that their collaboration was ending. In their place, HarperCollins France confirmed it would work with Fluent Planet, a Bordeaux-based communications agency whose process runs texts through automated translation software before passing the output to freelance post-editors.
The French Literary Translators Association and the collective En Chair et en Os responded with a joint press release condemning what they called an “invisible social plan” – a mass removal of regular income from workers with no notice and no redundancy rights. The European Council of Literary Translators’ Associations noted that some of the affected translators had worked for Harlequin for years, often exclusively.
HarperCollins France’s public position was careful: it was “conducting tests”, no collection had been “translated solely using machine translation generated by artificial intelligence.” Yet. No, sorry, that was me. The yet was implied. Publish-Speak at its best.
The company’s stated rationale was equally candid: “Sales of our Harlequin collections have been declining in the French market in recent years. We want to continue offering readers as many publications as possible at the current very low retail price.”
This was cost restructuring, openly justified. The labour displacement was real and the human cost ought not to be minimised. But the direction of travel was unmistakable, and it would not stop at translation. But, just another publishing news story, and in France. Not our problem.
Act Two: The Line (February 2026)
Three months later, in February 2026, Harlequin announced the closure of its Historical Romance line, Harlequin Historical, with effect from September 2027 – ending a programme that had run since 1988. The stated reason was “evolving reader interests globally.”
An author familiar with the line described to Reactor a more granular picture: years of steady reductions – reduced retail presence, narrowed genre focus, fewer monthly releases. As recently as 2025, the line had been repositioned to Regency and Victorian titles only, presented as a stabilisation strategy. The final shutdown, when it came, was still a surprise.
The closure attracted more industry sympathy than the translator story, partly because the historical romance readership is a vocal community and partly because the Harlequin Historical brand had genuine cultural heft.
But strategically, the two moves were closely related. Both represent a rationalisation of operations that could not be made cost-effective under the existing model – and, implicitly, a reallocation of capacity towards formats that could.
Act Three: The Market Nobody in Western Publishing Was Watching (2021–2026)
This is the largely buried story – and the one that makes everything else make sense.
Four years ago, the microdrama industry as we know it did not exist. Today, according to research by Omdia presented at MIPCOM 2025, it generates $11 billion in global revenues annually – nearly double the entire global market for free ad-supported streaming television (FAST channels). It is projected to reach $26 billion by 2030.
The format originated in China, where it is known as duanju – short, serialised vertical dramas adapted primarily from web novels, consumed on mobile, monetised through micropayment per episode or subscription. Revenues in China alone surged from $500 million in 2021 to $7 billion in 2024, and in 2025 Chinese microdrama revenues surpassed the country’s entire domestic theatrical box office for the first time, reaching $9.4 billion. More than 830 million viewers consume microdramas in China, with nearly 60% paying or transacting.
Outside China, the market generated $1.4 billion in 2024 and is forecast to reach $9.5 billion by 2030, at a compound annual growth rate of 28.4%. The United States leads international markets, and the American audience profile is striking: affluent, urban women aged 30 to 60, gravitating towards romance, CEO storylines, and revenge narratives. That is not a peripheral demographic. That is Harlequin’s core readership.
For anyone who has watched the trajectory of online reading – where telephone-number user figures were being reported from China and Southeast Asia long before Western publishing paid attention – the shape of this will be familiar.
A format originates in East Asia, scales to billions of users on mobile infrastructure, and is dismissed by Western commentators as either non-existent or beneath notice, right up to the moment it cannot be ignored. And then, denial kicks in.
” Dear Harlequin: Nobody Asked For Your Weird, New AI Video ‘Microdramas’, ” ranted LitHub literally days ago. “Add these new book-inspired microdramas to the piles and piles of AI-generated products that no one actually wants”, it went on, while failing to ask the question, who asked for LitHub?
That’s not as snide a remark as may at first appear. Okay it is, but even so, it makes a valid point. No-one asked for ebooks, for digital audio, for the iPad, for music streaming, for Netflix, for Storytel, for television, for Wattpad, for the internet… Products create their own demand.
Microdramas are following the same arc as web fiction, as manhwa and webtoon, as the phone-reading boom that transformed publishing consumption across South and Southeast Asia. The question is not whether this format matters, or (just to keep LitHub happy) who wanted it. The question is why it took Western publishers this long to engage.
Act Four: Dashverse, and Why the Geography Matters (March 2026)
Which brings us to Dashverse – and to the most underreported dimension of the Harlequin announcement.
Most coverage treated Dashverse as an anonymous “AI entertainment company.” It is rather more interesting than that.
Dashverse was founded in 2022 by Sanidhya Narain, Lalith Gudipati, and Soumyadeep Mukherjee. It is headquartered in San Francisco, with operations in Bengaluru. Its $18 million in funding has come from Peak XV Partners – the firm formerly known as Sequoia Capital India & SEA – alongside Z47 and Stellaris Venture Partners. This is an Indian-origin, India-backed company, and its Frameo platform was not developed for Hollywood or Canary Wharf. It was built and proven in Bengaluru.
In September 2025, Dashverse launched Raftaar, described as India’s first AI-produced microdrama – 42 episodes centred on car racing, produced in three weeks at roughly 25% of traditional production costs. Within 15 days of launch, Raftaar had recorded one million views, with a completion rate above 75% – some 90% higher than other short dramas on the platform. DashReels, Dashverse’s distribution platform, recorded ten million downloads across iOS and Android within its first month.
Imagine, for a moment, that the AI revolution in generative media had first surfaced publicly not from OpenAI in San Francisco but from a lab in Shanghai or Bengaluru. The reaction from the Western literary establishment would have been immediate and reflexive: suspicion, dismissal, and a confident assumption that whatever was being produced could not possibly meet the standards that matter. The Dashverse story is a version of that counterfactual playing out in real time. A company built on Indian engineering and Indian venture capital, deploying infrastructure developed for South Asian content markets, has just struck a deal with one of the most recognisable publishing brands on the planet.
Ouch! That is not a footnote. That is a signal.
The Frameo platform itself is worth a closer look. Dashverse describes it as a “generative video studio for storytellers” that combines text prompting with automated editing protocols, capable of producing a complete short film with consistent characters and dynamic audio. The Harlequin microdramas will be developed by a team of illustrators assisted by Frameo – this is not purely algorithmic output, but an AI-assisted production pipeline with human creative involvement at key points.
Authors whose works are adapted will receive royalties, with monetisation via advertising and, on certain platforms, subscriptions. That authors are being compensated at all separates this from some of the more egregious AI-adjacent publishing practices of recent years.
The Audiobook Analogy – and the Back-Catalogue Question
There is a useful historical parallel here. In the early days of modern audiobooks, the format was widely regarded within publishing as a niche accommodation for the visually impaired, an afterthought sitting outside the core business.
What changed was not the technology – audio had always existed – but the delivery infrastructure. Downloadable audio via smartphone transformed what had been a physical format constrained by shelf space and production economics into something genuinely scalable. Publishers who recognised that shift early built audio revenue streams that now represent meaningful proportions of their total income. Publishers who dismissed audio as a sideshow spent years catching up.
Microdramas, assisted by AI and distributed via short-video platforms, represent a structurally analogous moment for video adaptation.
Until now, converting a novel to video required a film or television company to see commercial potential, commission a production, navigate rights negotiations, and commit budgets that made anything other than an established bestseller an unattractive proposition. The economics excluded the vast majority of a publisher’s catalogue. AI-assisted microdrama production changes those economics fundamentally. Dashverse reports a production timeline of approximately three weeks per series, at a fraction of conventional costs. That transforms the calculus entirely.
Every major publisher is sitting on a back-catalogue that is, by any assessment, vastly undermonetised. Rights exist. Audiences exist. What has been missing is an affordable mechanism for conversion at scale. Microdramas – short enough to be produced quickly, serialised enough to drive repeat engagement, mobile-first enough to reach the audiences actually growing – may be the first format that makes systematic back-catalogue adaptation economically viable.
Romance is the obvious starting point because the genre’s emotional architecture – tension, intimacy, resolution – maps naturally onto the short serialised episode format, and because Harlequin’s brand equity makes it a lower-risk proving ground.
But the opportunity is not confined to romance. Young adult fiction, whose readers are already living on short-form video platforms, is an obvious fit, and as we have just seen with the Toonstar partnership, HarperCollins is ahead of the game.
Crime and thriller fiction, which already travels across streaming platforms with ease, translates readily to a format built on cliffhangers and momentum. Fantasy and science fiction, whose visual imagination has never been financially accessible at the adaptation stage for mid-list titles, becomes a different proposition when AI can generate consistent worlds and characters at speed. Even narrative nonfiction and biography – storytelling driven by real events and real stakes – has potential in a format that privileges emotional immediacy over production spectacle.
HarperCollins, and What Comes Next
It is worth going full circle here to the opener. Harlequin does not make these decisions in isolation. HarperCollins, its parent, has been moving faster and more openly towards AI integration than any other member of the Big Five. The French translation pivot, the line closure, and now the Dashverse and Toonstar deals are all consistent with a corporate direction that treats AI not as a threat to manage but as infrastructure to deploy.
Like it or not, its coherence as a strategy is not in serious doubt.
The question for the rest of the industry is not whether to engage with this shift but how quickly to move and on whose terms. Harlequin has moved first, with a partner whose roots are in South Asia and whose funding comes from the venture capital infrastructure that has been backing the Global South’s content technology revolution for the better part of a decade.
That is not the origin story Western publishing expected for its next major format bet.
It should have been.
The microdrama market grew from nothing to $11 billion in four years, following a trajectory that anyone paying attention to East and South Asian content markets could have seen coming. The numbers are not small. The audiences are not niche. The technology is not experimental. And the publishers sitting on decades of under-monetised IP, watching a three-week production pipeline become commercially viable for the first time, would do well to ask themselves a simple question.
Harlequin just moved first. Who’s next?
This post first appeared in the TNPS LinkedIn Analysis Newsletter.
