When AI flattens the production threshold: What exactly are short dramas competing for when they go global?
In September this year, the estimated in-app purchase revenue of overseas short drama apps on both platforms reached about 285 million US dollars, with the top 3 products accounting for 45% of the total revenue of the top 20. Among them, NetShort recorded an estimated revenue of 29.22 million US dollars, a month-on-month increase of 8.1%, officially surpassing ReelShort and rising to the second place on the ranking list.
In the same month, a landmark change appeared in the top 100 overseas AI drama list: NetShort returned to the top of the list with a short drama themed on "overweight wife's counterattack", and took more than half of the seats on the list.
Judging only by the ranking, this is a story of a company's rise. But when placed in the current cost structure of the industry, the meaning we read out is quite different.
Data from DataEye Drama Check shows that in the first half of 2026, the total number of placed materials for overseas short dramas reached 19.61 million groups, a year-on-year increase of 324%, exceeding the 13.957 million groups for the whole year of 2025; during the same period, the industry's in-app purchase revenue on both platforms was 1.27 billion US dollars, a year-on-year increase of 13%. According to the "2026 China Short Drama Overseas Market Research Report" by Diandian Data, user acquisition and traffic investment have accounted for 50% of the total cost of short dramas going overseas, and the cost per user acquisition in the North American market has reached 5.28 US dollars.
Supply has more than tripled, revenue has increased by more than 10%, and customer acquisition costs are still rising. When production capacity and materials can be replicated almost infinitely, scarcity has shifted from "whether it can be produced" to "whether the produced content has been verified".
The change in NetShort's position this time just falls at the node of this shift.
I. Two Curves and the Depth of the Ranking List
Data from DataEye shows that its estimated revenue on both platforms has continued to rise over the past year.
From 14.2 million US dollars in September last year to 29.22 million US dollars in September this year, the revenue increased by about 15 million US dollars in one year, with a growth rate of 105.8%. The estimated revenue in September hit a new high, approaching 30 million US dollars. In the first three quarters of this year, its total estimated revenue exceeded 200 million US dollars.
The curve on the placement side is equally steep. According to the data of ADX Overseas Short Drama Edition, from January to September this year, NetShort's placement intensity showed a fluctuating upward trend, rising from about 34,000 groups at the beginning to about 180,000 groups at the end of September. The total number of placed materials in September reached 933,000 groups, more than 230,000 groups higher than the second place, ranking first in overseas market placement.
The superposition of the two curves shows that placement and revenue are moving in the same direction, which is not common in the current industry environment. According to Sensor Tower's "State of Short Drama Apps 2026", in the first quarter of 2026, the global downloads of short drama apps exceeded 850 million times, a year-on-year increase of 140%; the in-app purchase revenue was about 750 million US dollars, a year-on-year increase of 20%. The download scale has expanded to 2.4 times of the same period last year, while the revenue has only expanded to 1.2 times. The whole industry is facing the same problem: users are growing faster than revenue.
In this context, the fact that it ranks first in the scale of material placement and second in revenue ranking indicates that the investment on the placement side has indeed been converted into revenue and ranking position, rather than just staying at the download volume level.
The figures on the content side also confirm this point. In the top 100 overseas short drama list in September, 3 NetShort dramas entered the top 10, with the highest ranking being third, and the heat value of a single drama reaching 12.677 million; in the top 100 overseas AI drama list in September, 6 dramas entered the top 10, with the highest ranking being the top of the list, and the number of placed materials for a single drama is close to 38,000 groups. NetShort took 54 seats and 53 seats in the two top 100 lists respectively.
Having both height and depth points to the breadth of content reserves, rather than the accidental success of a single hit drama.
II. Content: Treat Themes as Assumptions, Not Judgments
The market has fully shifted to AI dramas, and NetShort is no exception. The question is, in a market where everyone can make dramas with AI, why can your dramas be selected?
The drama "Overweight Wife Turns Hot, President Begs for Reconciliation" which topped the AI drama top 100 list in September is a suitable example. It is a typical combination of the core of "counterattack" and the shell of "plus-size". Plus-size itself is not a popular track, but two of the top 3 dramas in the September top 100 list are similar themes on the same platform.
One drama topping the list can be regarded as luck, but two dramas entering the top 3 at the same time indicates that there is indeed unmet demand in the market during this window. What is more noteworthy is the way it achieves this result: this combination is more like a very low-cost trial and error. The topic selection is treated as an assumption to be verified, quickly produced with a small budget, put into the market, and the data decides whether to increase investment.
The "One Move God Mode" at the beginning of the year took another path. The son of the sea god, with the identity of a humble farmer, participated in the knight selection holding a rusty pitchfork, and revealed his identity all the way to complete the counterattack. This work, regarded by the industry as the first real AI drama hit overseas, seized a time difference: the Western fantasy theme was restricted by cost in the era of live-action short dramas, and it was difficult to present grand world views and scenes, which AI just made up for, while the market supply at that time was still blank.
The common point of the two cases is not the sense of theme, but the verification path. The former verifies "whether a niche label can be accepted", while the latter verifies "whether a demand locked by cost will be released after the cost disappears". Neither of them is a judgment question, but an experimental question.
III. Localization: Rewrite the Shell, Not Just Translate Subtitles
After the theme is determined, the second question is what kind of shell can fit into the target market.
The approach of "Horror Game: Seduce My Ghost Teacher or Die" is to put the most evergreen core of the female-oriented genre "sweet romance" into the shell of "horror game" with a strong contrast. The heroine is trapped in a campus horror survival game, bound to a romance survival system, and must capture the dangerous ghost teacher to save her life, creating a sense of danger by setting "you will die if you fail the strategy".
The same romance, placed in a safe urban area and on a dangerous death line, has completely different dramatic pressure. The user psychology mined by this path is not just the love for watching romance, but the love for romance with costs.
What better illustrates the awareness of localization is its visual choice: the horror of this drama is not traditional Chinese horror, but Western-style horror that focuses on blood and monsters, and the characters and environment also fit Western aesthetics. This is not the result that can be obtained by adding foreign language dubbing to a Chinese script, but a rewrite from the setting level according to the fears and desires of the target market.
The difficulty of this set of methods lies in the proportion. The proportion of the two emotions of romance and horror needs to be precisely controlled, which is also the place where similar themes are not easy to replicate.
IV. AI: From Production Capacity Tool to Verification Tool
From script analysis, content production, language translation, to creative materials and traffic investment, AI almost runs through the whole process of the short drama industry chain. The way and efficiency of platforms using AI largely determine their position in the AI era.
Wang Li, CEO of NetShort, made a public statement: AI is not only an equal rights tool, but also a gap amplifier, and the gap between top creators and ordinary creators will be further widened.
This sentence has a direct correspondence in the current market. While the production capacity has increased explosively, the high-quality content rate has not increased equally. Most manufacturers use AI to quickly make dramas to increase the quantity, the production capacity goes up, but the proportion of high-quality content does not, and a large number of drama supplies instead distract users' attention. Those who really make good use of the efficiency improvement do not make more dramas, but make more good dramas.
Taking the September ranking as the benchmark, the difference can be seen very clearly. In the top 100 overseas AI drama list, NetShort takes 53 seats, 6 of which are in the top 10, and the themes cover labels such as plus-size, sports, werewolf, racing, horror game, cute dragon baby and so on. A large number of dramas of different themes on the same platform enter the top of the list at the same time, behind which AI makes low-cost verification of theme combinations possible: conduct small-budget trial production for the favored combinations, let the market give the answer directly, and then iterate quickly according to the feedback.
Under the logic of high-frequency testing and market orientation, it has the underlying support for a single platform to take more than half of the seats in the top 100 list.
The market side is the other side of the same logic. According to the data of ADX Overseas Short Drama Edition, NetShort's placement trend has risen linearly since the second quarter of this year, and now it ranks first in overseas app placement. DataEye Research Institute observed that it uses AI to produce materials for placement on most of its dramas: on the one hand, AI produces materials in batches to quickly verify high-performing materials, on the other hand, it uses AI to produce scenes that do not appear in the drama to guide downloads, and then superimposes traffic investment optimization to expand reach.
Multilingual layout also relies heavily on AI. Taking the data of September as an example, in addition to the core English market, NetShort takes more than half of the seats in the top 10 placed short dramas in markets such as Japanese, Korean, Arabic, Thai, Hindi, Vietnamese, Indonesian, Turkish and so on.
First, use AI to translate market-verified dramas into multiple languages at low cost, then use AI to produce materials to complete multi-market coverage, and finally use placement data to collect feedback from various markets. In this link, massive materials are not only customer acquisition tools, but also a set of extremely low-cost market research systems.
V. The Threshold of This Method and Its Truly Difficult Part to Learn
If you only look at the results, it is easy to have an illusion that you can replicate the results by copying the process. Three thresholds need to be clarified, which are also the truly difficult parts of this method to learn.
The first is the cost threshold. The 933,000 groups of materials in September correspond to a considerable scale of purchase expenditure, superimposed with multi-market translation and localization, which is not a structure that small and medium-sized manufacturers can afford on a regular basis. This also explains why the same methodology will produce different results in different teams: more investment does not mean correct investment, and this "correctness" often requires paying the tuition fee of wrong investment first.
The second is the window period. Once a theme combination is verified by the ranking list, the supply will be filled within one or two quarters. At the beginning of the year, Western fantasy was still a blank track, but now it is already crowded. The leading edge brought by verification efficiency is a time difference, not a barrier, so it requires continuous verification, not just one-time verification.
The third is the ability to make trade-offs. The ranking list only presents the parts that have succeeded. What high-frequency testing really tests is not how skilled the team is in using AI tools, but whether the team can stop in time when the data is unfavorable and concentrate resources on a few upward directions. The core action of this method actually takes place after the test.
The superposition of the three points points to the same thing: what can be transferred is not the process, but judgment.
Treat topic selection as an assumption, not a conclusion. The most common mistake in the industry is not making wrong judgments, but not having judgments that can be overthrown at all. "I think the audience will like it" cannot be verified, while "I can use two weeks and a small budget to prove whether the audience will like it" can be verified.
Treat localization as rewriting, not translation. The shell should be replaced with the fears, desires and class symbols familiar to local audiences. Translation solves the problem of being understood, while rewriting solves the problem of being willing to watch.
Treat materials as tests, not consumption. In the environment of material inflation, the amount of materials is a cost indicator, not a growth indicator. What we should really focus on is how much effective feedback every dollar invested in materials can get in return.
From a latecomer in the overseas market to taking more than half of the seats in the two top 100 lists and rising to the second place in revenue, NetShort's path can be summarized as exerting efforts on both content and AI lines, and using AI from a production capacity tool to a verification tool. What AI really changes is not the upper limit of production capacity, but the cost of trial and error. After the cost of trial and error is reduced, the winner is not the team that has the most trial and error times, but the team that can quickly converge to the correct answer from failures.
The anxiety of the industry is real: supply inflation, rising CPI, rigid upward cost. But the way out is also in the same place. In a market where almost everyone can make dramas, the only thing that has not been inflated is judgment.