Author: Duan Mingzhu

Source: Huxiu App

Produced by: Miaotou App

"Mango Excellent Media finally hit its 20cm limit-up," a user sighed in Miaotou's investment group on October 9.

After the National Day holiday, AI dramas led a wave of limit-up rallies in media stocks; leader Ciwen Media once printed six limit-ups in seven trading days.

The direct trigger was the dense catalysts around the holiday: Hongguo short dramas hit 168 million daily active users during National Day, surpassing the combined DAU of the four long-video platforms (iQiyi, Youku, Tencent Video, Mango TV). Roughly speaking, more than half of every 10 netizens had watched an AI drama — Hongguo covered in three years what long video took ten. Supply-side is even hotter: China's AI short-drama market reached RMB 22 billion in H1 2026, with full-year estimates exceeding RMB 40 billion.

Meanwhile, big players entered: Tencent launched "Lvpaopao Short Dramas"; AI content formats bloomed across the board, with Bona's AI theatrical film slated for an October 23 release, testing whether it can enter the cinema track.

These headlines turned "AI short dramas" from a track narrative into a yet-to-be-verified industry prosperity signal; with no stronger theme around, short-term capital needed an outlet.

And the choice of outlet follows A-share theme-trading's classic rules: a catchy label, no earnings, a small float, dispersed chips. Ciwen Media and similar names fit the bill.

All seems rosy — but there are real concerns.

At the industry level, booming supply and demand in AI dramas doesn't mean profits will necessarily grow. H1's hit rate was just 0.47% (DataEye data). Technology progress cut production costs 90%, but the flood of supply pushed traffic costs up 100% year over year, while monetization fell over 50% (Lin Qiwen, VP of DataEye).

On the investment side, the most "pure-play" AI drama names can't be picked. It's hard to find targets that simultaneously offer high business purity, proven profitability, and the ability to move the valuation anchor — so the rally lacks staying power.

So who pockets the biggest share of AI drama profits? How should one invest today? And if you work backwards from the endgame, where do you place your bets?

The zero-sum growth ledger

The trigger for today's AI drama boom: a cliff-drop in the "raw material" cost curve pushed supply-side barriers to nearly zero — but total costs didn't fall. Figure out where the cost cuts and price increases actually land and who holds pricing power, and the investment value hides in those nodes of cost redistribution.

AI dramas are like making a movie: by cost structure, money goes to three stages — making it (equipment), making it presentable (post-production), and getting it seen (screens, platforms).

Stage by stage, the price cuts today happen only in the first link, production: AI compresses per-title cost to 1/5–1/10 of a live-action short drama (RMB 500k–1M per title) — a genuine saving. But the more supply, the pricier scarce distribution and attention get.

The second link is post-production. Today, pricier models, more generation rounds, and cloud vendor hikes all raise AI dramas' compute costs.

On public pricing pages, Seedance 2.5 is roughly 20–35% pricier than 2.0; for a few-second shot in the hit "Fei Yao Zai," the creator re-rolled nearly 50 times before satisfied; Alibaba Cloud's compute card products rose up to 34%. Together these pushed per-episode compute cost of an AI drama from ~RMB 17 early on to RMB 80–100.

The third link is distribution. In 2026, the industry's daily ad-spend for short-drama acquisition runs RMB 150–170 million, with traffic costs up over 100% year over year (DataEye). It's as if 100,000 movies appeared overnight while the country's screens stayed in the tens of thousands — good release slots and good showtimes are finite, so marketing and scheduling fees naturally inflate.

On revenue structure, IAP (in-app purchases) and IAA (in-app ads) are the two dominant monetization models for short dramas today; IAA's scale share now far exceeds IAP's.

Under IAP, take an AI drama with RMB 10 million in gross revenue; costs mainly split into four parts (illustrative figures assembled from industry conventions): paid user acquisition (55–70%), platform and channel revenue share (10–15%), content production (0.8–1.5%), and IP costs (0–0.5%), among others.

That leaves roughly under 20% profit; if the acquisition or revenue-share coefficients shift, profit can fall further. Much of the remainder flows to platforms (as the acquisition channel and settlement party) and to the operators in the acquisition/distribution chain — producers end up with a tiny slice.

IAA works more like a theater joint venture sharing box office. IAA has no upfront paid-acquisition deduction, but platforms take the lion's share of ad revenue — reportedly 60–70% of total ad income goes to platforms, creators keep 30–40%, and many ordinary AI comic dramas end up with only about 20%.

Clearly, at a time when short dramas depend heavily on paid acquisition, under both IAA and IAP "the biggest chunk of profit ultimately flows to platforms."

A battle royale?

But even if platforms take the biggest share of AI drama profits, that doesn't equal net profit — nor does it make them great investment targets.

Among today's AI drama platforms, ByteDance's Hongguo, Tencent's Lvpaopao and Kuaishou represent three archetypes. In brief: Hongguo is the rent collector inside a closed loop, Lvpaopao the intermediary that trades content margins for ads, and Kuaishou the heavily funded infrastructure builder.

Let's unpack each.

ByteDance's Hongguo is the "closed-loop rent collector." It's the only platform holding IP, traffic, playback and monetization in one system: Fanqie Novel supplies IP, Seedance generates video, Douyin routes traffic, Qianchuan handles paid acquisition, and Hongguo closes the loop with "free + ads." What's more, Hongguo has enough sway to reshape the short-drama (including AI drama) ecosystem — minimum guarantees get cut at will.

On revenue, conservative estimates put Hongguo's 2026 full-year revenue at the RMB 20 billion scale, overwhelmingly ad-driven. Within the ByteDance system, AI drama revenue has never been disclosed, and no verifiable breakdown is available; under any inference, its contribution to ByteDance's total revenue is likely only a low-single-digit share; and ByteDance's 2025 net profit has not been disclosed either.

ByteDance's aggressive push into AI dramas is about user time, ecosystem control, and a bigger monetization entrance.

Tencent's Lvpaopao represents "conceding margins for ads." Its model gives away more content revenue share: in the up-to-100% revenue incentive, 90% is cash and 10% is ad credit usable only for paid acquisition. Read it this way: WeChat gives up content margins and earns from producers buying traffic; the more producers buy, the more intense the bidding, the pricier the ad slots. The 10% ad-credit design does indeed lock producers into its ad system.

So the real value in Lvpaopao isn't content settlement — it's acquisition ads. WeChat's short-drama mini-program channel had 508 million MAU, more than the 402 million of standalone short-drama apps (May 2026) — a huge traffic base. The catch: this ad revenue isn't broken out in Tencent's financials. In Tencent's 2025 earnings call, one line mentioned "mini-program ads."

Kuaishou is the "infrastructure builder in transition." IAA accounts for about 90% of its monetization; previously it ran a paid-acquisition ROI model — producers bought traffic, the platform took ad fees. Now it treats short dramas as part of its marketing and content ecosystem, shifting from an acquisition-ROI model toward GMV sharing, while committing RMB 800 million in funds, RMB 200 million in cash and RMB 1 billion in traffic to support content.

It lacks Hongguo's tight closed loop and Lvpaopao's WeChat-scale traffic pool, but its differentiator: shifting settlement rules from "acquisition-efficiency-dominated" to "shared commercial revenue," trading away some short-term retained margin for content supply quality and creator retention.

On revenue, Kuaishou's Q2 2026 platform-retained share from short dramas (settlement income) sits inside online marketing services revenue (RMB 20.6 billion), not broken out; gross margin fell from 55.7% to 51.6%, with brokerages attributing part of the decline to lower-margin short-drama IAA products and rising revenue-share costs. Separately, Kling AI revenue exceeded RMB 800 million.

The three platforms show three ways to play AI dramas: closed-loop rent (Hongguo), margins-for-ads (Lvpaopao), subsidies-for-ecosystem (Kuaishou). What they share: retained profit is diluted, hidden or commingled in each set of financials.

Which makes all three hard to invest in from an AI drama angle.

ByteDance isn't listed and Hongguo isn't spun out. Any A-share recommendation justified by "Hongguo beneficiary" is essentially a mapping exercise, not investing. However impressive Tencent's short-drama ad revenue, it's only part of the mini-program ad pool; the same goes for Kuaishou — AI drama revenue weighs too little in group profit, and Kuaishou's overall profit is being squeezed by both revenue-share costs and AI investment.

So platforms can't be bought. Settle for second best: look at whether other links in the chain deserve attention.

Who else is worth investing in?

To be clear, no AI drama name today simultaneously satisfies four conditions: high business purity, separable profit, sound unit economics, and the ability to move the valuation anchor. But plenty of links and names still deserve attention.

Producers are essentially working for platforms and channels — especially true in an oversupply phase; most such names are only worth a second look when they produce a breakout hit.

But producers that can bypass Hongguo or Lvpaopao in one or more of platform, distribution or monetization have higher value.

First worth watching: self-operated going-global.

ChineseAll's overseas short-drama platform FlareFlow is the standout: H1 revenue of RMB 253 million, book net profit of RMB 152 million, a net margin of about 60%; it reached monthly operating breakeven ahead of schedule, with 52 million cumulative registered users. Still, that doesn't prove the going-global model has fully worked.

First, peer data shows this isn't the industry norm. The general state of overseas short dramas is thin profit or losses: Inkeverse's Kuaichuang posted H1 revenue of RMB 948 million with an after-tax loss of RMB 6.18 million; Newborn Town's innovation business revenue grew 41.2%, but short-drama profit isn't separately disclosed, and its innovation business gross margin fell from 73.7% to 62.6% due to heavier short-drama investment.

Second, FlareFlow posted 2025 full-year revenue of RMB 352 million and a net loss of RMB 473 million. Swinging from huge losses to a 60% net margin within a year is too abrupt a turn.

Finally, current data can't verify standalone profitability. ChineseAll's H1 overall gross margin rose to 49.58%, up 17.71 percentage points year over year, and Q2 turned profitable — but these are consolidated figures; you can't reverse-engineer FlareFlow's standalone net margin from them.

Meanwhile, "building your own distribution platform" demands sustained capital and overseas localization operations capability — not replicable by most content companies.

From this angle, sustainable H-share watchlist names include Inkeverse, Newborn Town and China Ruyi — each with flaws: Inkeverse's Kuaichuang is swinging to losses, Newborn Town doesn't disclose short-drama profit, and China Ruyi's AI dramas are still in production.

A-share proxies: ChineseAll, IReader and Haikan. IReader's short-drama revenue share has risen to 66.44%, but losses persist, and between "short-drama revenue" and "AI drama profit" sit acquisition costs and platform revenue share; Haikan itself said on the interactive platform that "short-drama business accounts for a relatively small share of the company's total revenue and is not expected to materially affect current results."

The picks-and-shovels play is still worth watching. Seedance's annualized revenue is about USD 2 billion with a gross margin of roughly 70% — that's API revenue from a foundation video model; customers are enterprises, and what's sold is generation capability. It proves "people will pay for generated video," not "you can make money turning those videos into dramas." Wondershare's Juchang grew over 100%, with monthly compounded revenue growth from March to June.

On the IP/rights link, some attention is warranted.

China Literature is the only H-share name with both IP reserves and a short-drama adaptation chain: H1 2026 short-drama and AI comic-drama revenue of RMB 430 million, up 230% year over year, with 46 comic dramas crossing 100 million plays. But the figure is buried in "IP operations and others" — no separable profit. ChineseAll, IReader and others also book AI drama IP revenue.

The present, seen from the endgame

If AI dramas keep developing, which names in the track become important?

The AI drama industry will also go through a shakeout. Oversupply won't clear quickly through price mechanisms, but a slow shakeout is still a shakeout. DataEye counts active AI short-drama production houses collapsing from 1,216 to 698 in Q1 2026 — down 42% in a single quarter. What gets weeded out: producers without industrialized capabilities.

And content industrialization driven by AI is the broader trend. The last comparable shock to film and TV was streaming, and the winner was Netflix, not Disney+ — its edge: mastering both content production capability and a global distribution network.

Layering the two perspectives, the names worth watching fall into three buckets.

Bucket 1: IP libraries + matchmaking platforms — holding the "raw-material end" of industrialization.

Xiangcai Securities' research report ranks IP resources as the first investment theme, arguing "reader-market-validated mature IP lowers script development risk and raises the probability of hits." Other research institutions note that profit is shifting from production toward upstream IP: "AI hasn't weakened the importance of content — it amplifies the scarcity of quality IP."

Bucket 2: Own distribution platform + industrialized production — holding the "export end."

This bucket maps directly to the Netflix reference: a self-built distribution pipe is one of the few chain links that can retain profit. Watch platform-type companies with their own short-drama app matrix, an AI industrialized production system, and global paid-acquisition operations — they can enjoy overseas expansion dividends, but you must track marginal improvement toward breakeven.

Kunlun Tech is the name numerically closest to the "distribution + production" combo: H1 AI short-drama platform revenue of RMB 1.535 billion, up 163% year over year; June single-month gross billings exceeded USD 65 million; AI short-drama gross margin 67.49%. But its operating costs grew 406% year over year — clearly faster than revenue — and marketing expenses reached RMB 2.756 billion.

Bucket 3: The tool chain — holding the "production infrastructure."

The tool layer is, to a degree, drought-proof: when creators surge, tool call volume grows in sync; once a batch of production houses is cleared out, survivors may lean even more on refinement tools to raise per-title success rates.

To invest in the AI drama track right now, watch four things:

First, the direction of Hongguo's settlement rules matters most — it's the pricing anchor for the whole chain. If minimum guarantees keep falling and settlement keeps anchoring to real popularity, the platform is tightening its distribution power and content-side expected profit keeps compressing; if the rules shift to favor content creators, even partially, content-side logic needs re-rating.

Second, overseas platforms' operating data. Beyond profit, watch cash flow. ChineseAll's H1 2026 operating cash flow turned positive — check whether it's sustainable.

Third, third-party tools' ARPU and renewal rates. Wondershare Juchang's repurchase rate exceeds 85%, but the revenue base isn't disclosed; watch whether producers keep paying for these tools through the shakeout.

Fourth, content companies' segment profit disclosure for short dramas. If China Literature starts disclosing short/comic drama profit separately, the "IP-as-raw-material" logic gains a verifiable anchor.

Summing up: AI short dramas are on fire, but listed companies that can genuinely convert the industry heat into profit and support their valuations — there really are very few.

Hongguo's settlement rules are the pricing anchor; overseas cash flow is the real quality of going global; tool renewal rates are the stickiness; China Literature's segment disclosure is the verification anchor for IP monetization. Any directional change in these deserves re-weighing the affected names' valuation; otherwise, no matter how hot the track narrative, investment judgment should stay at "observe," not "confirm."

Disclaimer: This article is for reference only. The information or opinions expressed do not constitute investment advice. Please make investment decisions with caution.

The article comes from Huxiu App and does not represent the official position of Baijing. The content is for readers' reference and study only. Baijing assumes no liability for disputes or losses arising from this content. For reprints, please contact the original source.

Friendly reminder: Baijing currently has only WeChat and QQ groups. We have not created groups on Telegram or other social apps. Please beware of imposters asking our users and partners for fees or money.