Understand the original / Produced by

Author | Yang Yang

AI hard tech has had its ups and downs, but AI short dramas have become a safe haven in the market.

Recently, AI short-drama concept stocks with consecutive limit-up streaks have been anything but rare. The once-unloved film, television, and animation sector was also lifted to the top gainer in the A-share market over the past 20 trading days.

AI short dramas are sought after because they represent a scarce incremental market.

After AI short dramas cut production costs to just 2%–10% of traditional short dramas, supply has fully opened up. In the first nine months of this year, the number of short dramas released was 13 times that of last year. The supply explosion has brought a surge in traffic. The short-drama user base has reached 851 million, and the audience has expanded from small towns to inside the Fifth Ring Road—to the workstation next to you.

Where the people are, the money follows. As a new business format, the short-drama market is already nearly twice the size of the film market. An explosion of this scale, coming amid a pullback in hard tech and a lack of stronger themes in the market, finally gave capital an outlet.

But industry heat and corporate profit are two different things. According to DataEye, a drama needs 50 million views to break even—and only 1.3% ever get there.

Production costs have fallen by more than 90%, yet producers still aren't making money, because AI short dramas have become a paid-traffic game dressed in content clothing. Lin Qiwen, vice president of DataEye Ju Chacha, offered a comparison: AI short-drama production-side costs fell by about 90%, while traffic-placement costs rose by more than 100% year-on-year, and industry revenue shrank by more than 50%.

This situation isn't hard to understand.

AI made dramas cheaper to produce—and also made competition more brutal. In nine months, AI short-drama supply expanded 13-fold, but users and viewing time on the demand side clearly will not expand proportionally. At this point, for short dramas of similar quality, whether users ever see them depends on paid traffic.

After going around in circles, the money in AI short dramas has once again ended up in the hands of the shovel sellers and the rent collectors.

/ 01 / A New Traffic Entrance

Film, television, and animation is a tough industry: project-based, capital-intensive, with unstable returns and no repeatable formula for success. If one drama flops, a year's profit is gone. That's why it has long been unloved by the market, and valuations could never be pushed up.

But recently, film, television, and animation have turned things around. Choice data shows that over the past 20 trading days, the sector rose 20%, ranking first among more than 100 Eastmoney secondary industries—ahead of even robotics, which is regarded as a new quality productive force.

The turnaround in film, television, and animation can be said to have been driven by AI short dramas alone.

AI short-drama concept companies are leading the sector into a wave of limit-up gains. One leading AI short-drama concept stock once notched a sharp six limit-ups in seven days. The offensive in Hong Kong stocks was also significant, with one concept stock once rising as much as 336% over nine trading days.

The outbreak of AI short dramas in the capital market occurred around October, catalyzed intensively by a series of triggers:

During the National Day holiday, Hongguo Short Drama surpassed 200 million daily active users—more than the combined DAU of the four major long-video platforms iQIYI, Youku, Tencent Video, and Mango TV. Monthly per-user usage time on short-drama apps reached 26.9 hours, nearly double the 14.7 hours of the traditional online video industry. At the same time, giants entered the fray: Tencent launched "Lüpaopao Short Drama"; Bona's AI theatrical film was scheduled for release on October 23.

These dense news items formed a signal: AI short dramas are becoming a scarce incremental track and the next internet traffic entrance.

Today, among segmented internet business formats, the landscapes of e-commerce, short video, and social platforms have long been solidified, and user bases have successively hit ceilings. AI short dramas are the only incremental market in the current internet traffic business.

AI's cost-reduction and efficiency-improvement effect has made AI short dramas cost only 2%–10% of traditional short dramas. After the links in the content industry originally constrained by cost were fully opened up, the first nine months of 2026 saw 490,000 titles—13 times last year's number.

The supply explosion has brought a surge in traffic. As of May this year, the deduplicated user scale of the short-drama industry had reached 851 million. A user scale of 800 million shows that the short-drama audience has moved from small-town youth to inside the Fifth Ring Road.

Where the people are, the money follows. Through ad displays and membership subscriptions, AI short dramas are converting users' attention and consumer demand into revenue. In 2025, China's micro short-drama market exceeded 100 billion yuan, while nationwide cinema ticket sales for the entire year were only 51.8 billion yuan.

Dense supply, surging traffic, and soaring revenue are turning "AI short dramas" from a track narrative into an industry prosperity signal. When hard tech pulled back and there were no stronger themes in the market, capital finally found an outlet.

But industry heat and corporate profit are two different things.

/ 02 / A Paid-Traffic Game Dressed in Content Clothing

AI short dramas are popular—but are practitioners making money?

According to LatePost, more than 90% of participants in the industry failed to recoup their investments. This matches DataEye's drama-level data: a drama needs 50 million views to break even, and only 1.3% can do so.

AI has already reduced content costs to 2%–10% of traditional short dramas, so why is it still so hard for practitioners to make money? It starts with the business model.

Short-drama revenue comes from converting a large amount of wish-fulfillment fiction into video content, which users watch by viewing ads or paying. Paying users are therefore the foundation of revenue—and also where costs lie. To get more users to pay to watch, short dramas generally require feed advertising placement, equivalent to game promotion.

Under this model, paid-traffic costs already account for the bulk. At present, AI short-drama production costs account for less than 10% of full-chain costs (about 7.5%), while paid traffic and platform revenue sharing account for more than 80%.

From this perspective, AI short dramas are more like a paid-traffic game dressed in content clothing.

But they differ from games: hit games have life cycles lasting years, while hit short dramas last only a few months. Short dramas therefore don't involve long-term operations, and companies can only keep looking for the next hit.

The problem is that in a "paid-traffic business" that is always hunting for the next hit, the dividend from lower AI short-drama costs cannot catch up with continuously rising traffic costs.

Lin Qiwen, vice president of DataEye Ju Chacha, offered a comparison: AI short-drama production-side costs fell by about 90%, traffic-placement costs rose by more than 100% year-on-year, and industry revenue shrank by more than 50%.

It's not hard to understand why paid traffic eats profits. AI made dramas cheaper to produce—and also made competition more brutal. In nine months, AI short-drama supply expanded 13-fold, but users and viewing time on the demand side clearly will not expand proportionally. At this point, for short dramas of similar quality, whether users ever see them depends on paid traffic.

Of course, this supply-demand imbalance is also related to the large amount of similar, rough products brought by the pulse-like growth of AI content supply. When not only plots are similar, but even lines like "don't burn your heart" and "only by setting the rules first can the road become wider and wider" appear across different dramas, the short-drama industry is highly likely to see a round of elimination of low-quality capacity.

After capacity is cleared, the industry may have a better competitive environment. And the industry's revenue structure is the signal to watch for that clearing.

Advertising revenue depends on supply and algorithms. As long as total volume is large enough and algorithms are accurate enough, user time can be retained and then sold to advertisers—whether an individual work is good matters less. Subscriptions, however, require users to pay for quality content. When subscription revenue growth begins to exceed advertising revenue growth, that is the signal of industry clearing.

But no matter how the industry clears, the characteristics of high paid traffic and chasing hits are likely to remain unchanged, and it will be hard for AI short-drama producers' profits to stabilize. By contrast, the industry's most certain opportunities lie elsewhere.

/ 03 / The Biggest Beneficiaries Are Still the Shovel Sellers and Rent Collectors

In an industry with fragmented and excessive supply, producers are all working for others—and AI short dramas are no exception.

AI short-drama producers are largely lightweight production companies and project-based teams, with screenwriters mainly individual creators or small studios. The entities are highly fragmented, as hard to observe and count as hardware stores on the street. Most of them therefore face intense competition, and can only catch their breath when hit content appears.

But the upstream (shovel sellers) and downstream (rent collectors) of AI short dramas are doing well.

Among them, the most typical upstream shovel sellers are Seedance, Kling AI, Shengshu Technology, and Aishi Technology. What they provide is the underlying capability that "makes content producible"—the indispensable infrastructure business of AI short dramas.

These foundational video models charge through APIs. Their customers are enterprises, and what they sell is generation capability. Leading models have annualized revenue of about $2 billion and gross margins of about 70%. As long as short-drama supply continues to increase, shovel sellers' revenue will continue to rise.

Further upstream are IP suppliers. IP and copyright are important content sources for short dramas. Multiple platforms such as Yuewen, Jinjiang, Chinese Online, and Fanqie Novel will benefit, but the degree of benefit depends on the industry's level of premiumization and copyright policy control.

Downstream of AI short dramas, distribution platforms play the role of rent collectors. The more fragmented the supply, the more dependent it is on platforms. Domestic social media platforms and overseas Meta are rent collectors that AI short dramas cannot bypass. They bear no content risk—they only set rules and take commissions—yet they can take the bulk of industry profits.

Before making money themselves, they first let infrastructure providers make a fortune. This problem not only troubles AI short dramas, but most new AI businesses as well.