China’s AI microdramas are already half the top shows

Roughly 40 to 50 per cent of the top 100 microdramas streaming in China in early 2026 are already AI-generated or heavily AI-assisted, according to Media Partners Asia. The format — vertical episodes of one to three minutes, 60 to 90 to a season — is no longer testing AI at the margins. It is built on AI at the core, and audiences in China, Indonesia, Thailand and the Philippines are watching it win.

Media Partners Asia forecasts China’s short-drama revenue at USD 11.5 billion in 2026. The cheaper the production, the more bets a platform can place — and that economics, not prestige, is reshaping who leads.

The competition is not about who spends the most anymore. It is about who can fail the fastest and cheapest. A conventional Chinese web drama costs several million yuan to produce. An AI-assisted microdrama can be made for between CNY 1–2 million, and a platform can spin up dozens for the price of a single mid-budget series.

That changes what kind of company wins. When a title is cheap, a platform stops betting on one expensive show and starts running many small experiments at once. The recommendation system finds the ones that hold attention, and the platform pours resources into those.

ReelShort, owned by Beijing-based COL Group, doubled its AI engineering team after rapid growth in 2025. Viu and iQiyi are moving the same direction. The surprise is not that they are exploring AI. It is that, in China at least, the exploring phase is over.

Half the top titles are already synthetic

The headline number is the one most coverage missed. Media Partners Asia estimates that AI now generates or heavily assists 40 to 50 per cent of China’s highest-ranked short dramas. That is not a pilot. That is a production standard.

Joey Jia, founder and CEO of ReelShort, frames the shift bluntly. AI is “completely changing the competition landscape,” he says, and the company doubled its AI team to build better content faster. Jia traces his conviction to watching his ten-year-old daughter stream AI-made clips on YouTube. “I realized, at the end of the day, our [audiences] are all looking for great content,” he said. “Doesn’t matter what way you” make them.

Yang Xianghua, president of iQiyi’s overseas business, makes the cost case. He argues AI-native production will cut mid-budget drama costs by 20 to 30 per cent within a few years, mostly through virtual production and automated editing. He also notes most viewers cannot tell AI-made content from human-made work — animation already crossed that line.

Here is the honest limit on that claim. Yang concedes audiences still resist AI-produced high-end films today, and predicts the gap closes in roughly three years. That is a forecast, not a demonstrated result — the technology has shown it can make cheap content fast, not that it can make a feature people choose over a human one.

The twelve-month implication sits underneath all of this. Once AI handles scripting, localization and editing, the constraint stops being budget and becomes catalogue size and ranking. Vivek Couto of Media Partners Asia puts it plainly: the winners “won’t be the heavy spenders” but the operators who use AI to make more bets and new genres. The advantage shifts to whoever can industrialise story testing — and the open question is whether anyone outside China can match that pace.

The pipeline is the real machine

The mechanism matters more than the marketing. An AI microdrama is not one tool but a chain of them. A large language model drafts the plot and tightens scripts for a cliffhanger every minute or two. Text-to-video models build storyboards and simple scenes. Voice-cloning systems turn one performance into dialogue in Indonesian, Thai or Japanese.

Automated editors then assemble the cut and tune pacing against viewer-retention data. The whole pipeline produces many versions of one story, fast and cheap. That is what lets a platform double down on whatever holds attention.

China sits ahead of this race for a reason. Its platforms — ReelShort, Mango TV, Kuaishou-backed apps — pair a dense creator base with in-house AI teams wired straight into production. Janice Lee, CEO of Viu, describes the appeal of AI as doing “things faster, be more responsive to market, perhaps at a lesser cost.” That is the language of a follower, not a leader.

So the contest is settled on speed of iteration, not budget. A Western platform refining a few prestige shows is playing a different game from one that can launch fifty AI titles a month and let the algorithm sort them. The exploring phase is over in China because the economics already rewarded the operators who stopped exploring and started industrialising. The question for everyone else is whether they can copy the pipeline before the catalogue gap becomes permanent.

Beyond the headline

The money trail

The most important cash flow here is not subscription revenue. It is the arbitrage between ultra-low-cost AI production and the high engagement microdramas pull on mobile. Operators can launch dozens of AI titles for the cost of one mid-budget series, then earn through ads, in-app purchases and overseas licensing — letting capital chase volume and data rather than prestige.

The reach

The actor easiest to miss is the recommendation algorithm itself. As production costs fall, catalogues explode, and ranking systems effectively decide which titles become phenomena. Tuned for watch time rather than quality or diversity, those systems mean viewers reaching these formats through global feeds may see a narrow slice of hyper-addictive tropes — reshaping pacing and storytelling expectations well beyond Asia.

What isn’t being said

The upbeat industry talk skips the labour reconfiguration underneath. Writers’ rooms, storyboard artists and junior editors risk being broken into task-based, algorithm-monitored gigs with fewer routes to develop as long-form storytellers. There is also little public debate in Asia over residuals, moral rights or likeness consent when cloned voices and synthetic actors carry a show — leaving Western unions to set norms that may later clash with Asian practice.

What to track as the rules catch up

With China’s labeling rules expected late in 2026 and the EU’s disclosure regime phasing in, the gap between cheap AI output and the law governing it is where the next decisions sit.

• Streaming subscribers in North America, Europe and Australia

  The microdramas now dominating Chinese apps will reach your feeds through ReelShort and similar exports. If you want to know how the AI behind them is governed at source, read the Cyberspace Administration of China’s Interim Measures for Generative AI Services — the baseline rules for platforms shipping AI-heavy content abroad.

• Media and entertainment investors

  Watch the next Media Partners Asia short-drama update in early 2027. If ex-China revenue is clearly tracking toward tripling the USD 3.2 billion 2026 baseline, AI microdramas are going global. The fuller picture sits in the Daxue Consulting analysis of China’s short-drama market, which sets out the production-cost economics driving the bet.

• Policy and compliance teams at Western platforms

  Read the EU AI Act on disclosure of AI-generated audiovisual content. Its transparency duties could decide how you label or acquire Asian AI microdramas — and where they collide with looser Southeast Asian rules.

Explainer

Microdrama

A short, vertical-format serialized drama built for smartphones, with episodes of one to three minutes. Chinese seasons typically run 60 to 90 episodes, often romance-themed and engineered for binge viewing. The format’s economics flipped once AI cut production to roughly CNY 1–2 million per title, making volume rather than budget the competitive edge.

ReelShort

A consumer-facing microdrama streaming app owned by Beijing-based COL Group. It became a breakout export hit by localizing Chinese short-drama formats for overseas audiences, particularly in the United States. After rapid 2025 growth, it doubled its AI engineering team to automate scripting, editing and localized versions.

iQiyi International

The overseas arm of iQiyi, one of China’s largest video platforms. It distributes Chinese dramas and local originals across Southeast Asian markets including Indonesia, Thailand, the Philippines and Malaysia, often via bundled subscriptions. Its leadership argues in-house AI teams can cut mid-budget drama costs by 20 to 30 per cent within a few years.