The comic drama with 6 billion views fails to earn even 1 million yuan.

When producing AI manhua dramas, content producers have to bet on hit works, while platforms do not need to stake on any single drama.

If a drama flops, the content producer may not even recoup its costs. From IP purchase to production completion, funds are paid out bit by bit, and if no one watches the released work, all that money will be sunk in the content library.

Platforms are in a different situation. As long as users remain active, they can launch another drama immediately, and advertising slots can still be sold as usual.

Therefore, the most critical point of this business is who controls the revenue distribution rules.

01

The two core assets of ByteDance's ecosystem.

Fanqie has opened more than 60,000 original IPs for manhua drama adaptation. As for Hongguo, its daily active users reached 168 million in July 2026. According to the standard of QuestMobile, this DAU figure is larger than the sum of iQiyi, Tencent Video, Youku and Mango TV.

One side holds a massive library of stories, while the other controls the traffic exit.

Hongguo's core revenue in the first half of 2026 was about 200 billion yuan, 97% of which came from advertising, representing a year-on-year increase of around 87%. It should be specially noted that this figure is calculated by media estimates.

On the advertiser side: Hongguo's quotation for brands is 25 yuan per thousand impressions, with mid-roll ads forced to play for 5 seconds.

However, the quoted price is not the actual transaction price. The actual revenue for real-person short dramas per thousand impressions has dropped from around 30 yuan in 2024 to less than 20 yuan in 2026, and the figure for AI dramas is even lower, at about 14 yuan.

If the advertising slots cannot be fully sold, the platform has to fill the vacancies with lower-priced ads, which will also affect the actual transaction price. Advertisers pay for impressions, and which drama the ad is played in is not the top priority.

Turning to the content producer side, the algorithm has been updated again.

The free revenue share of an AI drama is calculated by multiplying four indicators: newly added effective viewing duration in the current month, unit price of duration, AI drama type coefficient and copyright coefficient.

I have checked several revenue sharing documents and found that the unit price of duration is not public, and the two coefficients are adjusted by the platform, which has been adjusted for several rounds in 2026. The only value that content producers can calculate is their own effective viewing duration.

Every time the platform adjusts the coefficients, the revenue of the same drama may change accordingly.

The quality of the script and the cast will eventually be reflected in the effective viewing duration. There is no separate item in the formula to measure the actual value of the work itself. If content producers want to get more revenue, they have to keep audiences watching for longer.

Even for the same 100 million views, the revenue may vary if the viewing occurs in different months.

The widely circulated revenue sharing standard in the industry is:

The platform takes 60% to 70% of the advertising revenue, while content producers take 30% to 40%. For ordinary AI manhua drama content producers, the share may only be around 20%, which will be further split between the copyright owner and the production contractor.

These proportions do not have unified statistical samples, and the calculation base is not public, so they cannot be regarded as the industry average.

Some people online have calculated an illustrative account based on 10 million yuan of advertising revenue: 6 million yuan for the platform, 2.5 million yuan for the production team, 800,000 yuan for the screenwriter, 600,000 yuan for the cast, and 100,000 yuan for the IP owner.

The IP owner only accounts for 1%. Of course, this is not the actual revenue sharing situation.

In reality, copyright contracts include both guaranteed minimum payment plus revenue sharing and full buyout. For full buyout contracts, no matter the drama gets 1 billion or 10 billion views, the revenue of the copyright owner will not increase.

The 6 million yuan for the platform is not net profit either; costs such as bandwidth, user acquisition and coin subsidies all require capital expenditure. Hongguo does not disclose its profits, so no external party can calculate its exact net profit figure.

For advertisers, Hongguo sells impressions; for content producers, it purchases content. The pricing rules on both sides are completely controlled by the platform.

Content producers have limited choices of top-tier traffic exits. If their works are not released on Hongguo, they will lose a large portion of the audience.

Therefore, what the platform really fears is that users stop coming. Works and content producers can be replaced, and as long as the platform retains users, its advertising business can continue to operate.

Works have a life cycle, but traffic can be monetized repeatedly. The same group of users can watch one drama today and another tomorrow, and the advertising slots are always available.

The party that truly has pricing power is the one that controls the traffic exit. Fanqie has the largest number of IPs, but it does not control its own traffic exit.

Indeed, this is the actual situation.

02

Fanqie owns more than 60,000 IPs, why does the bargaining power of a single work still remain limited?

The Legend of Ten Thousand Demons is a typical example.

This drama is adapted from Fanqie Novel That Demon is Lord Jiang!, and the total views of the whole series have exceeded 6 billion. The author stated that he only got 20,000 yuan for the first season which had 2 million collections, and the total revenue of all 12 seasons was less than 1 million yuan.

More than 6 billion views for less than 1 million yuan. At first glance, this figure seems extremely disproportionate, but view count never equals copyright revenue in the first place.

There are about 7,000 original novels on Fanqie that have been authorized for short drama adaptation, and Fanqie plans to invest 2 billion yuan in IP in 2026. Production teams can get up to 90% of the net revenue, while the author gets a separate payment.

The minimum guaranteed copyright fee for short drama adaptation is 120,000 yuan, and additional revenue will be shared after exceeding the guaranteed amount. There are also individual cases for new creators who received 50,000 yuan of authorization fee plus 200,000 yuan of shared revenue.

These cases show that authors are not unable to get paid, but how much they can get depends entirely on how the authorization contract is signed.

Fanqie has a huge IP library, so production teams have a large number of scripts to choose from. Unless the work is extremely scarce, it is almost impossible for a single work to negotiate a high price.

Fanqie and Hongguo both belong to ByteDance, but they sit at two different tables in the manhua drama industry chain: one provides content, the other operates traffic. When the IP is transferred from one table to the other, the pricing method changes completely.

Authors get paid according to the authorization contract, production teams get content revenue share, and Hongguo earns advertising revenue. For the same work, the way to make profit is completely different at different links.

Jinjiang Literature City presents a different situation.

Its copyright agency fee is usually 20%, and the author gets 80%, with specific details subject to the contract. The film and television copyright of top-tier IPs can be sold at the 10 million yuan level. The 2018 bundled film and television authorization of Tian Guan Ci Fu was traded at about 40 million yuan.

Of course, film and television copyright cannot be directly compared with manhua drama copyright, but this example at least proves that extremely scarce IPs can negotiate high prices in the authorization link even without their own playback platform.

However, no matter how high the authorization price is, it does not mean that the copyright owner can continuously share the subsequent traffic revenue.

China Literature wants to go a step further in this direction.

In the first half of 2026, China Literature's revenue from short dramas and AI manhua dramas reached 430 million yuan, up 230% year-on-year, accounting for 27% of its total copyright operation revenue. It owns IP resources, production capabilities, and has also launched its own Qidian Theater.

But according to the analysis of Kaijuan, more than half of the short dramas and AI manhua dramas still need to be distributed through external platforms such as Hongguo. Its own traffic exit can retain part of the users, but the rest of the revenue still has to be split according to other platforms' rules.

So does building a self-owned traffic exit solve all the problems?

The financial data of iReader just shows that things are not that simple. In the first half of 2026, iReader's revenue was 1.784 billion yuan, of which derivative businesses such as short dramas contributed 1.185 billion yuan, accounting for 66.44%. It owns IP resources, self-operated App, and short drama business.

But in the same period, its attributable net loss reached 50.1418 million yuan, sales expenses reached 1.109 billion yuan, accounting for 62% of total revenue, and operating cash flow was negative 117 million yuan.

For every 100 yuan of revenue, about 62 yuan is spent on sales expenses. Not all of this expense is for user acquisition, but it at least shows that maintaining user growth is not cheap. Its self-operated manhua drama platform Paoman also has works released on Hongguo.

Owning a self-owned entry is one thing, and being able to retain users at a sufficiently low cost is another.

ChineseAll took another path, building its traffic exit overseas.

Its subsidiary FlareFlow achieved a revenue of 253 million yuan in the first half of the year, with a net profit of 152 million yuan and 52 million registered users. It has launched nearly 7,000 episodes, of which AI dramas account for more than 10%.

However, this is the profit of FlareFlow, not the group's overall profit.

In the same period, ChineseAll's revenue from short dramas and IP derivatives reached 411 million yuan, up 108.72% year-on-year. Its attributable net loss narrowed by about 81% compared with the same period of last year, but still did not turn positive.

The third quarterly report released at the end of October added more details: the total revenue in the first three quarters was 833 million yuan, down 17.64% year-on-year. The revenue of the third quarter alone dropped by about 44%, and sales expenses were cut by more than 40%. The narrowing of loss was partially achieved by reducing investment.

Overseas operation also requires capital investment for user acquisition, translation and localization. Once this investment is reduced, the profit will recover slightly, but the growth rate will slow down. The difference is that if users can stay on the self-owned platform, the platform will have more autonomy in subsequent drama release and revenue collection.

Putting these companies together, the differences become obvious.

Fanqie has a large number of IPs, but a single work is not necessarily scarce. Jinjiang raises the authorization price by relying on top-tier IPs. China Literature tries to connect content with its own traffic exit. iReader has its own entry, but faces high sales expenses. ChineseAll is trying to retain traffic on its own platform in overseas markets.

Therefore, having a large number of IPs does not necessarily mean having pricing power, and owning a traffic exit does not necessarily mean making profit. It also depends on where users come from and how much it costs to retain them.

03

This is where Hongguo's advantage lies. It already has a huge user base, content can be replaced continuously, and advertising can be sold sustainably.

But the traffic exit is not a permanent property. Hongguo holds the pricing power, but other players also want to take a share of the profit.

As calculated earlier, the actual revenue per thousand impressions has been falling continuously, but Hongguo's total revenue is still growing. This shows that the growth of advertising scale has at least offset part of the impact of the decline in unit price.

Mid-roll ads cannot be skipped for 5 seconds, and there is still room to increase the number of advertising slots, and empty slots can continue to be filled. However, users' daily available time is limited, so ads cannot be added infinitely.

According to media estimates, Hongguo's annual revenue can reach 420 billion to 450 billion yuan. At the end of October, Hongguo specially issued a statement saying that this figure is untrue. No external party can access its real account, only the platform itself knows the exact number.

In addition to advertising revenue, Hongguo is also adjusting the cost of retaining users.

Coins are the subsidies Hongguo gives to users, users can earn coins by watching dramas and exchange them for cash. On October 21, Hongguo announced that it will phase out the coin system. From November 1, new users will no longer have access to the coin entry, and the system will be completely shut down by the end of the year.

Why did it stop investing in this subsidy?

The specific cost is not public, and no external party can calculate the exact figure. But at least it shows that Hongguo is adjusting the way of exchanging user viewing duration for cash incentives.

The 62% sales expense of iReader mentioned earlier also reminds us that attracting users is not cheap, and retaining them also costs money.

What is more troublesome is that peers are openly competing for high-quality content.

WeChat offers up to 100% revenue incentive for high-quality new dramas, iQiyi offers up to 100% revenue sharing for exclusive first-release works, Youku offers a single work reward of up to 1.35 million yuan for AI story films, and the top 10 copyright owners in Kuaishou's July revenue sharing list received a total of more than 13 million yuan in a single month.

On Hongguo's side, the high-quality AI drama incentive launched by Volcano Engine offers a maximum reward of 200,000 yuan per work, half of which is coupons that can only be used to exchange ByteDance's own tools. TikTok also allocated about 6.5 million US dollars in May to distribute to AI short drama content producers.

The applicable conditions of these policies are different, and the amounts cannot be directly compared, but one point is clear: content producers have more and more platform choices, and Hongguo also has to face the quotations offered by other competitors.

WeChat has another core asset in hand.

In February this year, the monthly active users of WeChat short drama mini-programs reached 427 million, up 59.9% year-on-year, larger than the sum of all independent short drama apps, and users aged 51 and above account for 30% of the total.

This group of users already exists on WeChat. Tencent's self-operated Lvpaopao Short Drama had 143 million MAUs in August, and the number was still rising rapidly in October. For content producers, in addition to Hongguo, there is another huge traffic entry. As for how much revenue they can get, it still depends on the respective rules of the platform.

Regulation is also changing this business.

From April 1, AI manhua dramas must be filed for record before release. From July, they are classified and managed according to two cost levels of 300,000 yuan and 800,000 yuan. In June, Hongguo removed 14,357 non-compliant manhua dramas in a single month. On September 1, the Administrative Measures for the Development of Micro Dramas came into effect.

The rules are getting more and more detailed. What content the platform can launch and retain is restricted. Controlling the traffic exit also means undertaking corresponding review and governance costs.

The advertising unit price is falling, and there is an upper limit for the increase of advertising volume. The coin subsidy is being phased out, and it remains to be seen whether users will stay. Peers are competing for content with capital, and regulation is also raising operational requirements.