China's AI short-drama boom has solved the production problem. It may have created a much bigger economic one.
China has discovered something remarkable about artificial intelligence.
You no longer need a large production company, a film crew, expensive locations or months of shooting to make a drama.
A small team with a few computers, a script and access to generative-AI tools can now produce episodes at a speed that would have seemed absurd only a few years ago.
And China is discovering the consequence.
When production becomes almost frictionless, everybody produces.
And when everybody produces, the scarce resource is no longer content.
It is attention.
That is the story behind China's extraordinary AI micro-drama explosion.
430,000 dramas in eight months
On September 17, China's National Radio and Television Administration gave the industry a number that captures just how rapidly the technology has changed entertainment.
Approximately 430,000 micro-dramas were released in China during the first eight months of 2026.
That is roughly 13 times the number released during all of 2025.
More than 90% were produced using AI technology.
China's micro-drama audience now exceeds 800 million people, while the domestic market surpassed RMB100 billion in 2025, according to the regulator.
Those numbers are extraordinary.
But there is another number hiding underneath them.
More content does not mean more money.
In fact, China's AI-drama industry is discovering the uncomfortable economics of abundance.
AI won the production war
The first advantage of AI is obvious:
cost.
Traditional short dramas require people.
Writers.
Actors.
Directors.
Cinematographers.
Editors.
Make-up artists.
Set designers.
Production assistants.
Locations.
Equipment.
Then there is post-production, marketing and distribution.
AI attacks many of those costs simultaneously.
A Chinese industry report cited by local media describes teams producing dozens of episodes with only a handful of people. One report says a five-person team can produce a 60-episode AI short drama within roughly a week, at less than one-tenth the cost of conventional live-action production.
That changes the economics of the entire industry.
The barrier to entry collapses.
You no longer need to raise a large production budget before discovering whether an audience exists.
You can make the content first.
Then test it.
Then make more.
And that is exactly what creators are doing.
The problem with making content almost free
There is a paradox at the centre of generative AI.
When something becomes cheaper to produce, people initially assume the industry becomes more profitable.
Sometimes it does.
But when the cost falls dramatically enough, competitors can flood the market.
Imagine that producing one drama costs $50,000.
Only a limited number of producers can afford to make one.
Now imagine the cost falls to $5,000.
Suddenly, ten times as many people can enter the market.
The audience, however, has not become ten times larger.
There are still only 24 hours in a day.
People still have only so much attention.
The result is predictable:
the supply of entertainment explodes faster than the supply of attention.
That is what appears to be happening in China's AI-drama market.
221,900 dramas chasing the same viewers
DataEye figures reported by 36Kr show that 221,900 AI-generated dramas and animation dramas were launched on Douyin alone during the first half of 2026.
That works out to more than 1,200 titles every day.
But only 2,886 crossed the reported 50-million-view threshold used as a break-even benchmark—about 1.3%.
Only 1,055 exceeded 100 million views, equivalent to approximately 0.48% of releases.
This is the part of the AI revolution that gets lost in the excitement about cheap production.
**AI can make a drama.
It cannot make people watch it.**
And that distinction is becoming economically brutal.
The million-dollar problem is attention
The traditional entertainment industry has always had a supply problem.
But AI is turning it into something much larger.
Before generative AI, making a bad drama was expensive.
Now making a bad drama is cheap.
That sounds like good news until millions of people start doing it simultaneously.
Then the market begins to resemble a casino.
Thousands of creators produce.
A small number hit.
Everyone else pays the cost of trying.
Industry reporting in August found companies producing hundreds of AI dramas per month while still struggling to generate meaningful profits. Some operators told 36Kr that they were essentially breaking even despite enormous production capacity.
The technology is therefore creating an unusual situation:
productivity is rising while profitability is becoming harder.
That is not a contradiction.
It is what happens when supply grows faster than demand.
The economics are getting worse
The pressure is also showing up in advertising and platform monetisation.
Industry reporting says the return per 10,000 views for AI short dramas has fallen sharply, with some reports putting current returns at only a few yuan per 10,000 views.
At the same time, platforms are reducing guaranteed payments.
Hongguo Short Drama reportedly changed its script cooperation policy in August, cutting the guaranteed payment for an S+ script from previously reported levels of around RMB40,000 to RMB50,000 to RMB10,000. The new policy also removed guarantees for lower A+/A categories.
That is an important signal.
Platforms are not simply encouraging more supply indefinitely.
They are beginning to ask:
Which supply is actually worth paying for?
That is a very different market.
The regulator's message is revealing
China's broadcasting regulator has not responded to the AI boom by simply saying:
“Stop AI.”
Quite the opposite.
It recognises that AI can reduce production costs and expand creative possibilities.
But the regulator is simultaneously saying that human-made productions remain important.
At the September 17 press conference, regulator Wang Xiaoliang said:
“AI is a tool. Humans are the creators.”
The regulator said it would continue supporting live-action productions and encourage platforms to establish dedicated sections and financial support for high-quality human-made dramas.
That is more significant than it sounds.
China is effectively trying to prevent its entertainment ecosystem from becoming a factory producing millions of interchangeable AI stories.
Because once the marginal cost of creating content approaches zero, quality becomes the new bottleneck.
AI has won the volume war
This is where the original observation becomes particularly interesting.
When Chinese regulators say they will vigorously support human-made dramas, it can reasonably be read as recognition that AI has already won one part of the competition:
volume.
AI does not need lunch breaks.
It does not need actors to arrive on set.
It does not need to rent a location.
It does not need to reshoot an expensive scene because an actor made a mistake.
It can generate alternatives.
Again and again.
And again.
That means human productions are increasingly competing against an opponent whose marginal production cost can be dramatically lower.
But that does not mean AI has won the entertainment war.
It means the competition has moved.
From:
Who can produce?
to:
Who can produce something people actually care about?
The human advantage becomes more valuable
This is where the argument becomes more complicated than “AI is replacing filmmakers.”
AI is already replacing or reducing demand for some production tasks.
CNA reported that some Chinese production teams have reduced editing staff because AI can generate video with parts of the editing process already built in. It also reported declining work for some actors and entry-level production workers.
But the same report highlights something important.
Creative judgement may become more valuable.
A story still needs structure.
Characters still need motivation.
Dialogue still needs rhythm.
A dramatic moment still needs emotional meaning.
Someone has to decide whether a scene is actually good.
AI can generate 100 versions.
That does not mean any of them deserve to exist.
This could produce an unexpected labour-market shift:
AI reduces the value of production labour while increasing the relative value of taste.
The person who knows what to make may become more important than the person who knows how to manufacture it.
The next battle is not China versus Hollywood
It is China versus the world for attention.
Chinese micro-dramas were already going international before the current AI explosion.
Chinese short-drama companies have expanded aggressively into overseas markets, including North America and Southeast Asia.
CCTV reported in February that more than 300 Chinese-linked micro-drama applications had appeared in overseas markets and that Chinese companies were increasingly exporting the entire production and commercial model. AI is being used for scripting, translation, editing and marketing.
The United States has become an especially attractive market.
The Financial Times reported that Chinese-backed micro-drama apps generated nearly $1 billion in net in-app revenue in the U.S. during 2025, according to Sensor Tower data, while Chinese-backed platforms accumulated tens of millions of downloads.
That means China's AI-drama oversupply has an obvious escape route:
export the content.
But there is a problem.
Everyone else has noticed the opportunity too.
Going global does not solve the quality problem
AI makes translation easier.
It can generate subtitles.
It can dub voices.
It can modify scripts.
It can localise expressions.
It can even generate marketing material for different countries.
That makes international distribution dramatically easier.
But localization is not translation.
A joke that works in Beijing may fail in Lagos.
A romantic fantasy built around Chinese social assumptions may not resonate in Los Angeles.
A family conflict that feels emotionally obvious in one culture may feel artificial in another.
The industry's own experience is already demonstrating this.
The Paper reported that North America, Southeast Asia and East Asia are major targets for Chinese micro-drama companies, but industry participants emphasised that localisation and business models remain critical.
AI can translate language.
It cannot automatically translate culture.
At least not reliably.
And this is where the market could get really interesting
The biggest opportunity may not belong to companies that can generate the most AI dramas.
It may belong to companies that can generate fewer but better dramas.
Imagine two studios.
Studio A produces 10,000 AI dramas.
Studio B produces 100.
Studio A wins on volume.
Studio B wins on storytelling.
If audiences become overwhelmed by endless AI content, Studio B may have the stronger business.
That is the paradox China is now entering.
AI destroys scarcity in production.
That makes scarcity elsewhere more valuable.
Story.
Trust.
Brand.
Characters.
Taste.
Originality.
Cultural relevance.
Distribution.
Audience relationships.
These become the new competitive advantages.
The coming AI content crash may not look like a crash
There may be no dramatic collapse.
Instead, something quieter could happen.
Thousands of small creators enter.
They produce hundreds of thousands of titles.
Platforms initially reward volume.
Then audiences become saturated.
Monetisation falls.
Platform subsidies shrink.
Low-quality content gets removed.
Creators discover that viral views do not necessarily equal profit.
Many leave.
A smaller number build professional studios.
The industry consolidates.
And AI becomes infrastructure rather than the product itself.
That pattern would not be unique to entertainment.
It is what happens whenever technology dramatically reduces production costs.
The first phase is abundance.
The second is saturation.
The third is selection.
China's regulator is already preparing for phase three
The regulatory response is revealing.
Since the beginning of 2026, Chinese authorities say they have removed 68,000 problematic micro-dramas and directed platforms to act against more than 1,200 accounts. More than 90% of the content involved in those enforcement actions was AI-produced.
AI-generated dramas must also carry appropriate labels, while platforms are expected to strengthen checks around the use of people's faces and voices.
At the same time, China is pushing for higher-quality productions and wants to release 1,000 high-quality micro-dramas during 2026.
That combination tells us where the industry is heading.
Not:
AI versus humans.
But:
AI-generated abundance versus human-directed quality.
The real revolution is bigger than entertainment
The Chinese micro-drama market offers an early glimpse of what generative AI may do to many other industries.
AI will make it dramatically cheaper to produce:
Articles.
Videos.
Advertisements.
Music.
Software.
Games.
Images.
Courses.
Marketing campaigns.
Customer-support content.
The immediate assumption will be:
More production = more opportunity.
But that is only half the equation.
If everyone gets the same production advantage, the advantage disappears.
When everyone can make a video, video becomes less valuable.
When everyone can write an article, writing becomes less scarce.
When everyone can create an advertisement, advertising becomes harder to distinguish.
When everyone can make a film, having something worth watching becomes the competitive advantage.
That is the economic lesson emerging from China's AI-drama experiment.
AI did not destroy the entertainment industry
It destroyed one of its old assumptions.
The assumption was that production capacity was scarce.
It no longer is.
A handful of people can now produce what once required an entire studio.
That is an extraordinary technological achievement.
But it creates a second-order problem that technology alone cannot solve.
Who gets watched?
Because when 430,000 dramas can be created in eight months, the scarce resource is no longer cameras.
It is human attention.
And attention does not scale just because AI does.
That may ultimately be the most important lesson from China's AI short-drama explosion:
AI has made creating content cheap.
It has not made caring about content cheap.
The next generation of entertainment companies will therefore not necessarily be the ones with the biggest AI models or the largest content factories.
They may be the ones that understand something much older than artificial intelligence:
people still need a reason to care.
Comments
Post a Comment