What the chart for 8 to 15 November shows
ShorStory put together the weekly GASP chart for the week of 8 to 15 November. The ranking measures view growth over seven days across 4 288 dramas. First place is Roaming the World Untamed: 24 million added views, 47 episodes, 12 of them free, about 21 dollars to watch the season to the end. Second is The Dragon King's Forbidden Maid: 20 million added, 63 episodes, 11 free, about 31 dollars. Third is The Triplets' Final Regret: 19 million added, 51 episodes, 8 free.
All three leaders sit on ReelShort and carry an AI label on their page. Last week's leader, Surf God's Revenge, dropped to fifth with 14 million added.
One caveat about the data. Chart figures come from the apps themselves and there is no outside audit behind them. Everything below is my reading of those numbers.
Demand fell, the AI share grew
The top 40 pulled about 251 million views together. A week earlier it was roughly 325 million, so about 23% down. Against that drop in total demand, AI-labelled titles went from 16 to 19, and ReelShort took 25 of the 40 slots, up from 21.
My reading is simple. When weekly demand falls, a platform cuts the expensive end of production first and keeps the cheap end. Nineteen of forty is close to half the chart, and at that scale the label stops being an experiment.

What the chart says about vertical series
Method matters more than positions here. The ranking counts growth over seven days, so a title with a large base loses to a fresh one: Love Is A Dangerous Game sits fourth with 119.8 million views in total and 14 million added in a week. For a client, that means the chart measures how fresh the demand is. A series that spent a month building an audience looks weaker in that table, even though it brought in more money.
There is a second catch. The ranking pool is 4 288 dramas and the table shows 40 rows. Forty rows tell you where the wind blows, they do not describe the whole market.
Genres this week: mafia, dragons, mind reading
Titles with Mafia, Don or Godfather in the name went from 4 to 7. DramaBox shipped three new series about reading minds. Fantasy built on dragons, sea gods and werewolves now accounts for about 10 titles.
The pattern reads like a shop window of expectations. Viewers arrive for a familiar conflict and for the promise of power, money or secret knowledge. For a client, genre is also a budget frame. A mafia story needs suits, interiors and grown-up faces on screen. A dragon and a sea god need a world that can be built entirely in generation. The second route is cheaper to produce, and the chart shows viewers accepting it.

Free episodes and the paywall: the numbers do not match the ranking
Love Is A Dangerous Game gives away 8 of 43 episodes, about 19% of the season. The leader gives away 12 of 47, about 26%. The runner-up gives away 11 of 63. The free share and the unlock price do not track chart position, and that is the most useful news of the week for anyone planning to sell a series episode by episode.
The practical conclusion: copying the leader's free episode count is pointless. You count from the hook. Some stories need three episodes before the viewer understands the conflict, some need eight because the story moves slowly. The paywall belongs at the sharpest point, and that point belongs to the story.
The AI label and what sits behind it
On ReelShort the title carries a note that it was made with AI. What share of the frames actually went through a model is not visible from the label: it can be every shot, it can be backgrounds, it can be boards assembled by hand. For a budget the difference is enormous. Animated series built by hand alongside generation are priced per second of finished runtime, and the AI label will read exactly the same as on a fully generated title.
So the chart cannot tell you how much real generation is on this market. It counts views, and production chains stay invisible.
Context: China treats this market as its own
Global Times on 15 November, citing DataEye: the share of locally made titles on the overseas short drama market grew from 32% to 53% in 2026. The same piece cites China Newsweek: more than 150 million active overseas viewers watch AI animated dramas. These are third-party figures and we have no way to verify them.
For a client the conclusion is this: the fight for attention runs against a producer whose episode costs less than your graphics budget. Winning then happens on story and on the point where the viewer pays.
What a client should do with this
Three corrections from the chart. First, do not chase episode count. A 63-episode series and a 43-episode series both hold the top because of the hook. Second, pick the genre by the cost of its world. Third, plan the paywall together with the season finale, because a wall in a weak spot kills retention.
Then the tools. If the series is really an ad format, vertical video ads for an app work best as short episodes, and the chart is a good source of references. When an episode is built entirely in a model, that is AI video production from a script, priced per second. When a character needs hand control, handcrafted AI videos give you that control, and the share of manual work still will not show on the card.
We have a separate breakdown of the Chinese market with 550 thousand microdramas, and the numbers there are about production volume. We also wrote about Meta's Series test, which covers format and monetisation.
My conclusion
This week showed that AI in vertical drama has stopped being a separate niche. Demand fell 23%, the AI share grew: the market is shrinking and getting cheaper at the same time. For a client that is leverage, because the cost of production falls faster than the cost of attention.
And a dose of sobriety. The chart is weekly app data. In a month the picture can turn, the AI label can become a viewer filter, the paywall can move into a subscription. You do not build a yearly plan on one week, and testing one episode against a table like this costs little.




