Bona Ventures into AI, Failed Three Times

By Short Drama Eight and a Half

All three of these fields also claimed to be the "first."

On October 23, Bona released "Sanxingdui: Future Past," billed as China's first AI-native theatrical film to obtain a public screening license.

Maoyan real-time box office shows a first-day gross of 300,000 yuan.

I also roughly checked nearby cinemas; there were very few ticket buyers, but prices still ranged from 25 to 40 yuan each. The most screenings were at Bona's own cinemas.

Judging by this box office trend, it is still a long way from recouping its costs (the tens-of-millions-level budget Bona claimed).

But its value seems to have already been realized.

A month ago, when Bona announced the release date for "Sanxingdui: Future Past," Bona's stock price hit the daily limit-up at the open, with nearly 700,000 lots in sealed orders and net main-force capital inflows of 137 million yuan.

In July 2024, "Sanxingdui: Future Revelation" landed on Douyin, billed as China's first AIGC sci-fi short drama.

At the time, Sora had been out for only half a year, and generative video technology was still at the stage of being watched by the whole public.

Looking back now, this AI short-drama series actually did not perform badly in views; at a time when the hit rate for AI short dramas is below 1%, it achieved over 200 million views on Douyin.

But those views were based more on the audience's curiosity about the "first." It was more like a PR event: it validated the topicality of "AI + Sanxingdui."

After that, Bona neither built up industrialized short-drama production capacity nor caught the commercialization boom of live-action short dramas in 2024-2025. Nor did it release any works amid this year's wave of AI short dramas.

"Legend," released in July 2024, used AI technology to recreate a 27-year-old Jackie Chan, whose scenes accounted for over 70% of the film, at a cost as high as 330 million yuan. The final box office was about 80 million yuan, a loss of over 250 million.

Before this, the film industry had used digital technology to recreate actors.

In "Furious 7," Paul Walker unfortunately passed away during filming. The production team completed the character through existing footage, stunt double filming, and digital technology—both to allow the film to be finished and as a farewell to the actor and the character.

Audiences like the 27-year-old Jackie Chan because they like his spirit of risking his life for films and innovating the Jackie Chan-style action, not just that face.

In Jackie Chan's classic films, it is his action rhythm, physical reactions, and tumbles and collisions that make the Jackie Chan audiences know.

A "Pure" Film Salesman

Chen Kaige once commented on Bona founder Yu Dong, calling him a "film salesman." Not only that, Yu Dong is also a "pure" film salesman.

A salesman needs a precise market vision. Actually, Yu Dong still has this. After all, that is how Bona's two successes came about.

The first was in the late 1990s, when Bona began doing film distribution.

At the time, mainland cinemas were starved for content, Hong Kong film production capacity was mature, and Yu Dong, holding the country's first private distribution license, sold a batch of Hong Kong films such as "New Police Story," "Confession of Pain," and "Overheard" into the mainland, moving from pure distribution to "investment + distribution" and earning his first pot of gold through information asymmetry and channel monopoly.

The common thread behind these Hong Kong films was a mature Hong Kong film industry, star lineup, and genre-film experience. Bona did not need to build a content production system from scratch; what it had to do was bring mature film products into the mainland market and then expand returns through investment and distribution.

The second was in the 2010s, when Hong Kong films declined and Bona grafted Hong Kong director teams onto mainland main-melody themes, creating "new mainstream blockbusters."

This path constituted Bona's most successful film business model in the past:

In 2014, Tsui Hark directed "The Taking of Tiger Mountain," grossing 880 million yuan;

In 2016, Dante Lam directed "Operation Mekong," grossing 1.18 billion yuan;

In 2018, Dante Lam directed "Operation Red Sea," grossing 3.65 billion yuan;

In 2021 and 2022, the two films in the "The Battle at Lake Changjin" series grossed nearly 10 billion yuan, setting the box office record at the time.

In contrast, for "Sanxingdui: Future Past," Bona still used the film logic of a big film and television company selling blockbusters—"big production + big cast + big subject."

It generated more than 1.2 million base images, carried out hundreds of full-film iterations, set up 17 professional divisions including cinematography, lighting, and art according to a traditional live-action crew, involved nearly 100 people, and so on.

It is already 2026, and they are still doing grand narratives for the film market.

Let's review the National Day holiday period that just ended. Total box office was 1.165 billion yuan, a new low for the same period in the past 12 years.

The film market probably can never return to the era when billions in box office came easily; even the most successful film "product manager" Chen Sicheng's "Traces of the Detective" only took in more than 400 million yuan, which a few years ago would not even have made the top ten.

In terms of film and television genres and even forms of entertainment, audiences have more and more choices. In terms of film and television content, audiences' demands for films are also getting higher and higher.

Audiences will not pay just because you are a "big production, big scene, big IP"; instead, those small productions and small-scale works that move people have won the market.

This year's most typical example is "A Love Letter to Grandma," with a cost of 14 million yuan, a Chaoshan dialect film, and an all-amateur cast.

Its final box office exceeded 2 billion yuan, with an input-output ratio of 140 times.

This film relied only on the simplest word-of-mouth praise from audiences.

Bona can say that making "Sanxingdui" with traditional live-action production might cost 800-900 million yuan. But the logic that AI cost-reduction narratives can hold exists only on the premise of good content.

One can only say that Yu Dong is "pure" to the point of being somewhat stubborn; he seems to recognize only films.

He once said that short dramas are just a supplement to their film and television content.

You may look down on short dramas, but you cannot look down on this market, or on the content preferences of audiences. In the film Infernal Affairs II, Ni Yongxiao in his transitional period can also be patriotic.

In fact, another company that entered short dramas around the same time as Bona was Stephen Chow's Bingo Group. By comparison, Stephen Chow is much more honest and never shies away from making money.

Starting with Gold Pig and Jade Leaves, Bingo Group established a connection with the short drama market. Since then, Bingo Group has continued to extend into new content formats.

In 2026, Stephen Chow, together with Bingo Group, made a strategic investment in Suzhou Interactive Star, publicly announcing that they would cooperate around AI dramas, interactive film-games, live-action film and television, and full-industry-chain IP operations, with a focus on AI animated dramas and AI simulated live-action dramas, and plans to jointly develop related content with Interactive Star.

This year, Stephen Chow also announced a partnership with Volcano Engine to directly monetize film IP licensing.

Three moves in three years, each at the starting point of a technology trend, yet each giving the impression of "rising early but arriving late to the fair."

All three share a similar capital narrative: after securing a "first," a stock price rise always makes the news. Whether the content itself can be accepted by audiences and whether the business model can work seem, by contrast, less important.

Yu Dong's "purity" toward film is probably not thorough enough. The simplest purity of all is a sense of faith in film content.