Why Did China's First Fully AI-Produced Theatrical Film "Sanxingdui: Future Past" Fail at the Box Office?

A film that, when its release date was announced, once sent its producer's stock price surging to the daily limit. It features no live actors at all—every character is an AI-generated digital figure—and its story forges a sci-fi link between Sanxingdui civilization (the bronze sacred tree, the protruding-eye masks) and a future AI crisis (binary codes, a superintelligent AI). It sounds cool and trendy. But after release, as the first domestic AI feature-length theatrical film to obtain the Dragon Seal, "Sanxingdui: Future Past" collapsed at the box office, taking in only 1.246 million yuan nationwide in its first four days, with an average attendance rate of less than 1%. Its projected total gross falls far short of its ten-million-yuan-scale budget, and the icy box-office figures even quickly cooled the capital market. Only 1.246 million yuan in four days—just why? First, the "fragmented" viewing experience is quite severe. Although AI performs impressively in handling grand scenes and artistic texture—the film's Sanxingdui bronze sacred tree, ancient Shu sacrificial scenes, and the lighting and shadow effects of its sci-fi scenes are stunning, rich in detail, and demonstrate AI's powerful computing capacity in image generation—once it comes to dynamic performance, AI's shortcomings are fully exposed: characters' expressions are dull (a "frozen-face feel"), micro-expressions are missing, lip movements don't match, and there are even modeling flaws like characters who "don't age for decades," drawing audience complaints that it looks like "a high-end game CG" or "a short drama stretched out." Moreover, perhaps to avoid the complex staging that AI is not good at, the film relies heavily on collages of short shots, causing scenes to jump and leaving many hard flaws in narrative logic (such as poorly paid-off foreshadowing and thinly drawn villains); audience feedback includes "so tiring to watch" and "like several different group assignments stitched together." Second, although the "AI" gimmick was previously popular in the capital market, it is hard for ordinary viewers to convert into willingness to pay. First, free AI short films are flooding short-video platforms; audiences are long accustomed to them and have even formed the fixed impression that "AI = short video," making them unwilling to pay high movie ticket prices for a "long AI video." Second, the big screen amplifies AI's coldness; compared with the individualized charm and emotional tension of live actors (especially stars), AI characters seem cookie-cutter and struggle to move ordinary viewers. AI characters lack personality and tension. It can thus be seen that although "Sanxingdui: Future Past" obtained the Dragon Seal (the film's filing category is animated film; the Dragon Seal prefix is "Dian Shen Dong Zi," following the animated film review path), the Dragon Seal only resolves whether it can enter theaters, not whether audiences are willing to buy tickets. Theatrical films must face hard metrics such as ticket price, screening slots, star effect, story completeness, and emotional resonance. For the industry, this Dragon Seal for "Sanxingdui: Future Past" has value for exploration; for the producer Bona, it has value for its technology roadmap; but for the commercial return of this single film, it offers almost no direct help. It is more like an "AI feature-length theatrical stress test": the compliance channel has been cleared, but audience acceptance has not. In other words, the market does not buy the selling point of "full-process AI"; audiences pay only for story and experience, not for a technology label. When AI technology still cannot support the delicate emotions and rigorous narrative required by a theatrical film, it is very difficult to gain a foothold in a fiercely competitive market on the name of "the first" alone.