AI comic dramas smashed 20 billion RMB in a year, yet 90% of players are bleeding: who actually banked it in 2026?

In 2025, AI comic dramas went from near zero to nearly 20 billion RMB in a single year. Total views hit 64.3 billion, transaction volume surged 12x. By early 2026, over 10,000 new titles were launching every month, with an active catalog close to 130,000. It is the most explosive corner of the entire short-drama market.

But here is the number that contradicts those flashy stats: 90% of non-top-tier companies are bleeding cash or just spinning their wheels. Burning thousands in tool subscriptions every month, hoarding dozens of titles, gambling on that one viral hit. That is the grim reality for most people in this wave. In Q1 2026, active AI comic drama production studios in China fell from 1,216 to 698, a 42% drop in a single quarter. Over 40% of teams tapped out.

The money is not in the act of “knowing how to generate AI video.” That barrier has collapsed to the floor. The real money sits in three overlooked places: scaling industrial production capacity, claiming platform revenue shares, and carving out uncontested consumer and overseas markets. This post will not teach you “how to start” (that is the beginner piece). It answers a sharper question: after you start, why are 90% of people losing money, and who is pocketing the rest?

First, understand the market: it is way more than 20 billion RMB

Lay the numbers out, and focus on the ones everyone ignores.

• The market doubled. In 2025 the AI comic drama market reached nearly 20 billion RMB (DataEye and Daily Economic News; iiMedia puts it at 18.98 billion, up 276.3% YoY), with 64.3 billion views and revenue jumping 12x. In 2026 it is projected to hit 22 to 24.36 billion RMB, half the entire short-drama sector’s growth, as users climb from 120 million to 280 million.
• Costs collapsed. AI swallowed the whole production pipeline. Efficiency is up over 80%, timelines cut by a third, and per-minute cost fell from about 2,000 RMB to roughly 400 RMB. An animated short drama used to need a 50-person crew, two months, and over a million RMB. Now one person can ship a final cut in days.
• But the hit rate is just 0.16%. Of 60,946 titles released in 2025, only 96 broke 100 million views. Content piles up in fantasy, suspense, time-travel, and rebirth, and audiences are burning out.

The rules changed too. Starting in 2026, tiered review for micro-short dramas tightened sharply: projects under about $140K (1 million RMB) are self-reviewed and filed by the platform; $140K to $420K (1 to 3 million RMB) goes to provincial broadcasting; anything over $420K (3 million RMB) goes to the National Radio and Television Administration, with a 7 to 15 business day review cycle. The Measures for the Administration of Micro-Short Drama Development took effect on September 1 on top of that. The old playbook, small teams churning low-quality trend-hopping junk to flood the feed, is dead.

Bottom line: the Wild West era is over. The opportunity is not gone; it changed hands. The game shifted from “knowing how to generate an AI video” to “knowing how to pick winning topics, run an industrial system, and choose the right platform.”

Position 1: build industrial-grade capacity, not just one video

The real money in this wave is not in “can generate an AI video.” It is in “find viral source material and scale it through continuous industrial production.” The first is a craft, the second is a factory. Craft gets commoditized by tools. A factory is your moat.

Industrialization means turning the pipeline of script → character → storyboard → video → compositing into a repeatable assembly line. Each step’s output feeds the next, and your character keeps the same face from the first frame to the last.

The one worth tearing down is Alibaba’s open-source LumenX Studio (MIT license, repo alibaba/lumenx). It takes the comic-drama industrial pipeline, from asset extraction, style locking, asset generation, storyboard script, and storyboard frames to storyboard video, and packs it into one platform. It uses Tongyi Qwen for script analysis, Tongyi Wanxiang for image and video, CosyVoice for voiceover, and lets you plug in mainstream video models like Kling, Vidu, and Seedance. It collapses the chaos of jumping between tabs and tools, from “starting with nothing” to “exporting a broadcast-ready cut,” into one production line.

The tool market’s heat tells the story. Per Huxiu, the one-stop creation platform “Youxi AI” launched in January 2026, pulled in 13,000 creators in 5 days, and booked over 36 million yuan in ARR in its first month. The shovel-sellers are printing cash, which means the gold rush is still on.

If you are running this yourself, memorize one battle-tested rule: lock the storyboard for every shot first, then feed it to the video model. Never go straight from script to video. Per multiple AI comic-drama creators, going script-to-video yields about 60% garbage; running it through a storyboard first drops that to about 15%. The trick is using the simplest-composition frame as your anchor for the first frame, the reference for every frame after. That one move cuts your material cost by more than half.

Top companies bet on this capacity to stack the odds of a hit. Jiangyou Culture scaled to over a thousand employees in a year, broke 100+ episodes a month, and hit roughly 50 million yuan in monthly revenue. Lingju Animation grew from 30 people at its May 2025 kickoff to nearly 800, lifting monthly output from 2-3 episodes to a projected 50-70 by year’s end, aiming at 150 a month. You cannot out-volume them, but you can beat them on per-episode precision. Which leads into the second position.

Position 2: claim platform revenue sharing, forget paid ads and driving your own traffic

In the paid short-drama era, most revenue went to the media buyers running paid traffic. In 2026 that flipped: free-to-watch plus revenue sharing became the standard, and platforms now compete to backstop premium content with guaranteed payouts. You no longer burn cash buying traffic, you just supply the drama. The catch is “premium content,” which is why 90% of low-quality mass production is getting wiped out.

The real scale of these revenue-sharing payouts, from public data:

• Hongguo Short Drama (ByteDance): From August to October 2025, script revenue-sharing payouts topped 189 million yuan in three months; top studios cashed out over 10 million each. Minimum guarantees run 40,000 to 200,000 yuan plus up to 40% lifetime share, on a “72-hour viral spike + 18-month long tail” model. The Hongguo app has about 240 million monthly active users (QuestMobile).
• Douyin: The “Starlight” and “Chenxing” plans hand comic-drama creators a 90-95% revenue split. Some IP-adaptation projects stack an extra 50% support bonus on top, and single titles have passed 6 million yuan.
• Yuewen: In six months it adapted over 1,000 web novels into comic dramas, and AI comic drama revenue passed 100 million yuan in the same window. Fanqie Novel has pushed over 6,000 titles into the adaptation pipeline.
• Kuaishou: For 2026 it is putting 800 million yuan into diversified revenue-sharing, 200 million yuan in cash to incubate premium content, and 10 billion in dedicated traffic incentives.

Here is what this means for you as an individual: you are no longer gambling on whether you can buy traffic. You are betting on whether your content can pass the platform’s quality bar. Fail, and 10 low-quality dramas equal zero. Pass, and one hit feeds you for an 18-month long tail.

Position 3: carve out uncontested niches: consumer creators, going global, micro-genres

Industry insiders say it themselves: the B2B market for lowbrow and motion comics is a bloodbath. But “consumer creators and premium production (S-tier series, AI + live-action) are completely untapped.” Hybrid talent will not face real competition for three years. Wherever the market is not saturated, that is where the money is.

Going global is the biggest open advantage. AI does not just cut production costs, it also solves the old bottleneck of taking short dramas overseas: translation. Localization time fell from 3 days to 4 to 5 hours, and cost dropped 90%. The market is scaling fast: per DataEye, the overseas AI short drama market was about $100 million in 2025 and is projected to hit $650 million in 2026, a 6x jump. The US is the biggest cash cow, over 40% of overseas short-drama in-app purchase revenue, with per-episode willingness to pay about 6x other markets. Southeast Asia, Latin America, and the Middle East are the fastest-growing incremental markets. Whoever first uses AI to localize a hit into dozens of languages and floods the market cashes in first.

One widely cited case: in April 2026, the AI short drama Revenge of Persia, made by Chinese creators, hit the overseas platform YourChannel. Within 72 hours it pulled about $500,000 in GMV (about 3.6 million RMB), and the platform pays creators a 90% share. Both the title and platform check out (multiple media reports, self-reported platform figures, no independent audit), and production cost was reportedly about 3,000 RMB. Take the numbers with a grain of salt, but the leverage of “low cost plus multilingual global expansion” is real, and the direction is certain.

Niche genres are another open gap. Hongguo’s “Guoran Plan” gives up to a 30% revenue-share boost to scarce categories like realistic themes, regional culture, and innovative historical romance. The platform is paying to pull you out of the saturated red ocean of fantasy, suspense, time-travel, and rebirth.

These three positions are not either/or; they stack. Industrial production is your foundation, revenue splits are your monetization, uncrowded niches are your positioning. But when you start, nail one down first. Do not grab all three at once.

An honest starter checklist

No motivational fluff. The exact steps:

1. Stop hoarding low-quality output. The hit rate is 0.16%. Ten rushed episodes are worth less than one carefully made one. Get one episode through platform review and earning revenue splits before you think about scale.
2. Start with open-source tools. LumenX Studio, Huobao Drama, and Toonflow are all open-source and commercial-use. Beginners, skip the command line. Start with cloud-hosted or desktop versions; Toonflow has a desktop installer that is a double-click away.
3. Obsess over character consistency. Use DeepSeek to generate the script structurally (the character profile layer matters most, it has to be specific enough to feed image generation). Generate and lock your character reference shots first. Map the full storyboard before going image-to-video (this drops wasted footage from 60% to 15%). Assemble in CapCut and add AI voiceover.
4. Position away from red oceans. Prioritize going global (Southeast Asia, Latin America) or tap Hongguo revenue-sharing. Pick niches with favorable splits like realistic, regional, or historical romance. Stop crowding into fantasy and time-travel.
5. Calculate your monthly tool costs. High-quality video and image generation burns through Jimeng and Kling credits or Tongyi Wanxiang API fees. Once you run at capacity, a few hundred to a few thousand a month is normal. Do not get played by the “one-click AI video for $10k a month” hustle bros. Get one episode profitable end-to-end first, then talk scale.

The Wild West era is over, and the opportunity has found a new owner

The 20 billion RMB market is real. The fact that 90% of players are bleeding is just as real. Both being true tells you one thing: this wave is no longer “free money for anyone who knows the tools.” The ones eating now are the operators who pick winning topics, build industrial systems, choose the right revenue-sharing platforms, and carve out niches where no one else is fighting.

The window is still open. It belongs to the ones willing to polish a single project until it qualifies for revenue share, not to the gamblers hoarding dozens of low-quality titles and praying for a viral hit.