跳到正文
非凡资本

UNIQUE RESEARCH / ENGLISH ARTICLE

Inside StoReel: Making AI Short Dramas That Move American Audiences for US$20,000–40,000

Original · Unique Research · 2026-03-19 · Shanghai

English edition note: This complete translation retains the source author's first-person reporting, Audrey's interview statements, the original promotional lines and the author's explicit doubts. Production budgets, industry statistics, financing and team histories are reported source claims, not independently verified results or price guarantees. The source's cost range and percentage-reduction wording are preserved without assuming identical production conditions. References to StoReel Canvas being forthcoming, future capabilities and the company's ambitions reflect the original publication date, not confirmation of a current launch or proven business model.

Unique Research · Guest Interview

100 Ways to Slap Someone—and an AI Short-Drama Business Built Around Virtual People

US$200,000 for one live-action short drama made for Europe and North America.

That is a figure Audrey, StoReel's head of global business, mentioned repeatedly. In the business of taking short dramas overseas, it stands like a wall in front of countless creators hoping to bring Chinese hits to international audiences.

But Audrey immediately offered another figure: US$20,000–40,000.

A reduction of more than 80%. The same drama, still intended to move audiences in America's central states, goes from costing US$200,000 to US$20,000. Not by squeezing quality, but through AI.

I initially thought this was another familiar story about using AI to cut costs and improve efficiency. Then Audrey said something that made me stop and think for a long time—

"

We encourage creators to use AI to generate 100 ways of slapping someone, but ultimately a person chooses which one best fits the personality of this domineering-CEO character.

100 ways to slap someone.

The image is extraordinarily specific. It points not to a grand narrative about AI replacing humans, but to a very practical question: when AI can generate content in bulk at extremely low cost, choosing becomes the most valuable skill. StoReel is building a business model around that ability that is entirely different from ReelShort and its peers.

Not Selling Dramas, but Selling "People"

In early 2026, the one thing the overseas short-drama market did not lack was new entrants.

ReelShort's monthly gross receipts exceeded RMB300 million, with DramaBox and ShortMax close behind. Even Disney and Netflix had begun testing vertical short dramas. But Audrey was not especially interested in those figures. The metrics she watched were entirely different.

"What we define as effective IP value is not views or daily active users, but recognizable distinctiveness," she said. "We measure it through the rate at which fans follow a character across platforms and the rate of fan-created derivative content. If AI-generated, factory-made sweetness leaves fans remembering the plot devices but not the characters, we have failed."

There is a great deal packed into that statement.

The dominant overseas short-drama strategy today is a content pool plus paid user acquisition: produce large quantities of series, test them through advertising to identify hits, then spend more on the winners. It is essentially a traffic business. ReelShort has reached 2 in-house series releases per week, while ShortMax has more than 1,000 episodes in its catalog. Plenty to keep viewers fed.

StoReel takes a different approach. Audrey summarized its model in three terms: turning characters into IP, putting writers and directors at the center, and the forthcoming StoReel Canvas.

Start with turning characters into IP.

Traditional short-drama platforms make disposable consumer products: a user finishes one series and moves to the next. The characters have different names, but underneath they are the same domineering CEO and the same Cinderella. StoReel is different: it uses AI to incubate and manage virtual performers.

"We are not just distributing content. We use AI to incubate and manage virtual performers, turning these characters into assets owned by the platform," Audrey explained. "That not only reduces costs, but also builds long-term user stickiness through ongoing character interaction—such as social-media management and fan engagement—creating IP value beyond an individual series."

What does that mean? A virtual character does not live only inside a drama. He can post selfies on Instagram, reply to fans on TikTok, appear across different series, and even be licensed for co-branded collaborations.

The character becomes an asset. A drama is just one vehicle for that asset.

My first reaction was: isn't this just a virtual idol? Luo Tianyi on Bilibili, also known as B-site, and Japan's Hatsune Miku have already traveled this road.

But Audrey's explanation that followed made me feel the logic really was different this time.

"AI Puts Directors Back in the Director's Seat"

"Our AI tools always put the creator at the center," Audrey said firmly. "Creators can shape characters directly from their imagination, instead of having their creativity worn down layer by layer through traditional filming and production processes."

This is what StoReel calls its writer/director-centered approach.

The traditional short-drama production process roughly runs: a writer creates the script → a director interprets it → casting → live-action filming → post-production. Every step can drift away from the creator's original vision. The domineering CEO in your head is cold-faced and ruthless, but casting finds you a sunny, warm-hearted guy. Much of your creative idea is lost in execution.

AI compresses that chain. Creators can directly control characters' appearance, expressions, and body language, moving seamlessly from script generation to storyboard previews. Audrey's metaphor of 100 ways to slap someone is about precisely this: AI does not decide for you, but gives you 100 options. You only need to choose the one that feels right.

"Creators can oversee the whole picture like directors, while AI handles efficient execution."

This suggested a counterintuitive logic to me: the more powerful AI becomes, the more valuable human judgment is. When generation costs approach zero, what to make is no longer scarce; what to choose is.

The underlying tools StoReel is building are intended to put that power of choice back in creators' hands.

It is about to launch a platform called StoReel Canvas. Audrey described it as "an AI creative-collaboration platform for creators worldwide that breaks through the limits of a single model, allowing creators to freely access all the world's mainstream large models."

Whether designing characters or drawing storyboards, creators would no longer be confined to a closed toolset. Instead, they could combine cutting-edge AI capabilities in one unified interface. In Audrey's words, it is a creative hub.

Put plainly, StoReel does not want to be just a short-drama platform. It wants to become infrastructure for content creators.

Ammunition and a Home Base

When we discussed monetization, Audrey offered a metaphor that I found clearer than any business analysis—

"Short-term traffic is ammunition; long-term IP is a home base."

Too many players in this field are stockpiling ammunition.

The industry figures are there: by early 2026, AI short dramas featuring realistic human characters had surged from 7% of the top 100 animated/comic-drama rankings a year earlier to nearly 60%. More than 80% of leading short-drama companies were already using AI in production. Costs were plunging: a 100-minute AI short drama cost between RMB100,000 and RMB200,000 in total, while a live-action production of the same scale cost at least three or four times as much.

Costs have fallen and output has risen. What happens next?

"We have seen too many models that do nothing but pile up gratification beats. Traffic arrives quickly and leaves just as quickly, and users remember none of the characters." Audrey paused after saying this.

There is an open secret in the industry: 80%–90% of overseas short-drama projects struggle to recoup their costs. User-acquisition costs are high, retention is low, and users bought through heavy spending leave after watching. If retention does not improve, paid acquisition must continue, creating a vicious cycle.

StoReel wants to break that cycle with virtual performers. If a character can keep interacting with users outside the drama, users stay for the person, not for a particular plot device. Fans follow the character rather than the series.

This is much like the logic of short-video influencers: you follow a creator not because one particular video is good, but because you connect with that person.

"Technology can be bought and models can be copied," Audrey said. "But directing an AI actor's gaze, adjusting narrative pacing to the habits of users in America's central states, and getting people to keep interacting with virtual performers outside the series—accumulating that experience and data takes time."

Hidden in that statement is StoReel's real claim to a moat. It is not technology itself, but AI operating experience grounded in audiovisual storytelling, combined with the DNA of social products.

Speaking of DNA, StoReel's core team has a substantial pedigree. Its CEO came from Linmon Pictures, led the construction of its short-drama business system, and helped drive Linmon's stock-market listing. Its CTO has a technical background at Momo and ByteDance. In early 2025, the team secured several million US dollars in strategic financing from a well-known overseas gaming company.

"Our core team combines Momo's social DNA with the production capabilities of leading film and television companies. We know how to keep users engaged, not just push content at them."

Momo's social DNA. I sat with that phrase for a moment.

What does Momo do? It creates sticky interactions between strangers. The virtual-performer management StoReel wants to do is essentially the same—except that one side of the interaction changes from a real person to an AI character.

AI Short Dramas Are Only the Starting Point

When I asked Audrey about the most important commercialization breakthrough she wanted in 2026, she did not name a specific market or series. Instead, she pointed toward something larger—

"AI short dramas are only a starting point. We are more focused on how AI can substantially improve content-production efficiency and evolve content beyond traditional video into IP assets that are interactive, extensible, and capable of sustained operation."

Interactive. Extensible. Capable of sustained operation. Unpack those three terms: audiences can influence the direction of a story; a character can extend from short dramas into games, social platforms, and even e-commerce; a virtual actor does not die just because a series ends.

This is no longer the logic of making short dramas. It is about building a content operating system around virtual characters.

"We want StoReel to become an AI-native content company, not merely a short-drama company."

Audrey said this quite naturally. But put it in the industry's context, and its meaning becomes clear: while ReelShort and its peers are still competing over catalog size and paid-acquisition ROI, StoReel has bet on a different arena—competing over assets rather than traffic.

Not who has more ammunition, but who has the stronger home base.

A Story That Has Yet to Be Proven

Honestly, it is a beautifully told story. But I have to pour a little cold water on it.

Whether virtual performers can actually work in European and North American markets remains an open question. Japan has successful precedents in virtual streamers, or VTubers, but its cultural environment is entirely different. Will American users interact with an AI character on Instagram? Will they follow a virtual actor across different series? Actual data is needed to answer these questions.

AI technology for realistic human characters is also still improving. In early 2026, the industry was still discussing the uncanny valley intensely: the more realistic an AI character becomes, the more uncomfortable audiences can feel. Microexpressions and physical responses are already said to be infinitely close to those of real people, but there is still a gap between approaching and arriving. That gap may be the dividing line between retaining users and losing them.

StoReel Canvas has not officially launched yet, and freely accessing all the world's mainstream large models is a major promise. Large-model API costs, stylistic consistency across models, and creators' learning curves are all engineering problems still to be solved.

But I have to admit that Audrey's metaphor gets at a real industry problem.

100 ways to slap someone. When AI can produce 100 versions, whoever can choose the right one wins.

Ultimately, AI has made doing cheaper, but choosing more valuable. That is what StoReel is betting on.

Will the bet pay off? I do not know. But the question itself is worth considering for everyone who makes content.

StoReel Canvas Creative-Collaboration Platform

Freely access the world's mainstream large models and usher in a new era of AI short dramas

Selected Interview Q&A

Q: AI can quickly generate vast amounts of content, but creators worry about diluting the distinctiveness of their IP. How do you balance this?

Audrey: "We will not let AI replace creativity; we make it a copilot. What we define as effective IP value is recognizable distinctiveness, measured by the rate at which fans follow a character across platforms and the rate of fan-created derivative content. If AI-generated, factory-made sweetness leaves fans remembering plot devices but not characters, we have failed. We encourage creators to use AI to generate 100 ways of slapping someone, but ultimately a person chooses which one best fits the personality of this domineering-CEO character."

Q: Creators commonly chase short-term viral traffic and may overlook long-term IP value. How do you view this trade-off?

Audrey: "Short-term traffic is ammunition; long-term IP is a home base. We have seen too many models that do nothing but pile up gratification beats. Traffic arrives quickly and leaves just as quickly, and users remember none of the characters."

Q: What capability is hardest to replicate in AI-driven global content?

Audrey: "It is the combination of an AI creator ecosystem that understands audiovisual storytelling and the DNA of social products for overseas users. Technology can be bought and models can be copied, but directing an AI actor's gaze, adjusting narrative pacing to the habits of users in America's central states, and getting people to keep interacting with virtual performers outside the series—accumulating that experience and data takes time. Our core team combines Momo's social DNA with the production capabilities of leading film and television companies. We know how to keep users engaged, not just push content at them."

Q: What commercialization breakthrough matters most in 2026?

Audrey: "AI short dramas are only a starting point. We are more focused on how AI can substantially improve content-production efficiency and evolve content beyond traditional video into IP assets that are interactive, extensible, and capable of sustained operation. In doing so, we will keep improving our AI content-production tools, creator ecosystem, and global distribution system."

Q: What kind of team will StoReel become by the end of 2026?

Audrey: "We want StoReel to become an AI-native content company, not merely a short-drama company. The team will have a clear structure: a product team for AI technology and creative tools, a creative team focused on content and IP, and a content ecosystem serving creators around the world. StoReel's role is not just to produce content, but to enable more creators to participate in creating the next generation of content."

Material sourced from Unique Awards guest interviews

Originally published by Unique Research on Unique Research Substack on March 19, 2026. This page preserves the public article for reading on UniqueCapital.

View the original publication ↗
← Back to English research