Recently, an AI short drama titled Huo Qubing, which was allegedly "produced by three people in 48 hours at a cost of only 3,000 RMB and garnered 500 million views," went viral on social media platforms, sparking heated public discussion. Beyond the content of the drama itself, the market is more eager to understand whether AI has truly achieved such high efficiency in production.
However, Yang Hanhan, the director of Huo Qubing, subsequently clarified these claims: The short drama does not have an 80-episode volume; there are only two versions, one lasting just over 4 minutes and another lasting 6 minutes. The 3,000 RMB cost refers only to computing power expenses and does not include any labor costs. The "500 million views" figure originated from media reports; since the company could not accurately verify this data, it adopted the media's numbers in its publicity without rigorous statistical confirmation.
Although the figures are not as exaggerated, video generation models represented by Seedance 2.0 are rewriting the fundamental logic of content production through low production costs and generation effects that closely mimic physical reality. With technological support, AI comic dramas are rapidly capturing the content landscape, and major platforms are swiftly moving to seize this new opportunity in the short drama sector.
According to Yang's personal account, her team is the sole team behind Huo Qubing, with its main business encompassing films, short dramas, AI digital humans, and full-process AI creation. The team, she noted, utilized "Namistory" to complete the entire production process.
Information shows that "Namistory" is the first industrial-grade AI manhua intelligent agent production platform launched in China by 360 Group. The actual controller of the operating entity behind it is Chinese entrepreneur Zhou Hongyi.
Zhou commented on social media, saying, "Seeing everyone discussing the 'Namistory' and the AI short drama Huo Qubing produced by Yang, I believe that as AI technology matures, more high-quality works will emerge in the future."
According to relevant reports, the "Namistory" is positioned as a B-end productivity tool for professional creators and film and television companies. Its core revenue model is based on token consumption for tool services (i.e., the computing power costs incurred when creators call upon AI models to generate content on the platform). The platform officially launched its public testing on Jan. 30.
With the advent of models like Seedance 2.0, AI short drama production has undergone a revolutionary change in the production phase. According to the pricing standards for its video generation large model Seedance 2.0 disclosed on the official website of Volcano Engine, the price, including video input, is 28 RMB per million tokens, while the price excluding video input (pure video generation) is 46 RMB per million tokens. A simple estimate suggests that for pure video generation scenarios, the equivalent cost is approximately 1 RMB per second.
Currently, AI manhua dramas are mainly categorized into motion comics, 2D/3D manhua, and AI-driven digital-human dramas. Liu Zun, research director of DataEye, stated that the costs for AI-driven digital-human dramas primarily include computing power costs, labor costs, and copyright and compliance costs. Among these, computing power costs account for the highest proportion, roughly 30% to 50%.
"Computing power costs are transparent now, with mainstream models essentially charging based on duration," Liu noted. According to him, computing power costs are approximately 1-3 RMB per second, meaning the computing cost for a 1-minute video ranges from 60 to 180 RMB, plus an additional 2 to 5 rounds of iterative modifications. Overall, the cost of an AI-driven digital-human short drama is roughly between 2,000 and 6,000 RMB per minute, with premium content potentially exceeding 10,000 RMB per minute.
Lyu Shaolong, deputy general manager of Lingem Anime, said that currently, the cost for AI-driven digital-human dramas is approximately 3,000 RMB per minute, with a typical total duration of around 100 minutes. Based on this estimate, the cost of a single drama is about 300,000 RMB. Regarding the production cycle, from receiving the script to delivering the final product, AI-driven digital-human dramas generally require 35 to 40 days.
Chen Junjie, head of a media technology company in China, said that with the development of AI technology, public demand for AI-driven digital-human dramas has significantly increased. Unlike dynamic comics or 2D/3D manhua, the production requirements for these dramas are higher, involving a process known as "card drawing" (Note: inputting prompts, keyframes, and other instructions to the AI, then waiting for the generation of the final result).
"For example, if we want to generate a scene: a young man in Han Dynasty armor facing the enemy in the desert. The model's generated result might deviate from our expectations, requiring repeated 'card drawing' until it satisfies us," Chen said.
As Seedance 2.0 accelerates technological progress, the production of AI dramas has become more convenient and efficient, rapidly capturing user attention. In February this year, Tencent launched its first standalone manhua app, "HuoLong," and entered into a manhua copyright licensing cooperation with COL Group. Additionally, major companies like ByteDance and Baidu have launched their own standalone manhua apps, while platforms such as Kuaishou, Bilibili, iQiyi, Youku, and Mango TV have also introduced manhua channels.
In DateEye's January 2026 Top 100 manhua list, AI-driven digital-human dramas saw the most significant surge, with their representation in the top 100 increasing from 7% last year to 38%, accumulating a total of 2.548 billion views. In the February Top 100 manhua list, it accounted for nearly 60%, becoming the absolute mainstream.
Liu mentioned that following the emergence of models like Seedance 2.0, the possibility of one person producing a short drama has become a reality, representing a democratization of creation in the industry. The barrier for ordinary people is extremely low; anyone can get involved with creativity. Side hustles and small teams can produce content, and with more opportunities for trial and error, it's easier to create a hit. In the long run, a group of super individual creators will emerge.
However, he also noted that current technological capabilities still have limitations since long videos and complex actions are not yet generated naturally enough, and content can easily become homogenized and templated. Ultimately, the competition will still hinge on the story and creativity. Furthermore, in terms of compliance and copyright, issues such as filing for the use of portraits, music, and deep synthesis are also important considerations that need attention.
"Individuals might create a hit, but they cannot sustain mass production," Chen argued. He believes that short drama production companies operate differently from individuals; they often integrate multiple models to form customized intelligent agents, achieving the best results at the lowest cost, rather than just relying on a single model.
On the future of AI manhua dramas, Chen emphasized that content creativity is paramount. He believes that the biggest bottleneck for AI in the future will still be talent, because the industry needs individuals with thoughts and aesthetic judgment.
(Source: CLS)
Related News:
Deepline | Beyond the algorithm: Why AI can master music but not touch souls
Deepline | OpenClaw's tale: Two paths of AI diffusion in China and the US
Comment