Viewing AI Companionship and Anime-Style Games Through Character.AI's Bankruptcy
On October 12, Character.AI filed for voluntary bankruptcy with the U.S. Bankruptcy Court for the District of Delaware and plans to sell a license to part of its intellectual property to The Walt Disney Company for $20 million. The author believes its failure stems from insufficient product strength: although the platform has 200 million registered users and 18 million AI characters, its underlying model logic is chaotic and lacks memory and worldview isolation, making it difficult to compete with anime-style games such as Genshin Impact and Arknights—which feature character art, voice actors, and affection systems—relying solely on text-based conversation. Based on this, the author concludes that the AI companionship sector is not over, and that the future direction is the integration of high-quality audiovisual presentation, deep worldbuilding, and more intelligent large-model interaction.
Why read this
AI score 62
Using the mature companionship mechanisms of anime-style games as a reference, the article analyzes the shortcomings of Character.AI's product experience and business model, and offers directions for integrating AI companionship.
Topics
Body · AI translation
[[Character.AI's bankruptcy filing does not mean the end of the AI companionship sector; rather, its product experience and business model have fundamental flaws. Anime-style games have repeatedly proven that emotional companionship is a strong demand, and AI companionship needs to integrate with mature content experiences.
**Key Points**
**1. The root cause of Character.AI's bankruptcy is its own insufficient product strength**
Although the platform had 200 million registered users and 18 million AI characters, its underlying model logic was chaotic and lacked memory and worldview isolation, so the quality of experience could not support its high valuation; this has nothing to do with "the end of the sector."
**2. Anime-style games are the truly mature form of AI companionship**
Popular games represented by Genshin Impact, Arknights, and Blue Archive strengthen player immersion by downplaying the protagonist's appearance and lines, letting players become "themselves" within emotional bonds—a model validated by the market for decades.
**3. Character.AI's companionship experience is far inferior to that of anime-style games**
Relying solely on a text dialog box, it lacks exquisite character art, voice actors, an affection system, and worldview support; its only strength, "conversation," is also full of holes, making it hard to compete with mature anime-style companionship systems.
**4. The "false premise of the sector" is high-EQ rhetoric that masks execution failure**
From the saturation of the Three Kingdoms SLG market to PC games forever being the core, and then to "large models don't matter," the industry often attributes individual product failures to the macro environment, which in essence is an excuse for its own slow strategy.
**5. The future of AI companionship requires a dual integration of technology and experience**
Truly promising AI companionship should combine high-quality audiovisual presentation, deep worldviews, and smarter large-model interaction, rather than blindly piling up text chatbots. 2026-10-16 18:18 From Character.AI's bankruptcy, let's talk about AI companionship and anime-style games He Beihang Quick Overview
On October 12 local time, Character.AI formally filed for voluntary bankruptcy with the U.S. Bankruptcy Court for the District of Delaware and plans to sell licenses to some intellectual property to The Walt Disney Company for a consideration of $20 million.
A bloated giant
In the AI social and chat companionship sector, Character.AI was once a star to watch. When it completed financing in 2023, its valuation reached $1 billion in one stroke; by the end of 2024, the company further announced that platform registered users had exceeded 200 million, monthly active users exceeded 20 million, and the total number of AI characters created by users had reached 18 million.
In my 2023 article "[Sushi Talk] Character.ai's current model has no future if unchanged," I once shared my hands-on experience.
Taking one highly popular character on the platform with 64.2 million interactions—Hu Tao, the "Director of the Wangsheng Funeral Parlor" in Genshin Impact—as an example, when I tried to ask her about information related to the "Wangsheng Funeral Parlor," the AI's answers were completely mechanical recitations piled together, with none of the lively, quirky soul of Hu Tao's character in the original work.
After experiencing product performance at that level, looking back at those grand numbers like "tens of millions, hundreds of millions" inevitably creates a strong sense of mismatch and doubt.
This phenomenon is exactly the same as low-quality mobile games in the gaming circle that "surpass ten million pre-registrations and quickly die immediately upon launch"—there is a huge gap between the user scale and buzz on paper and actual retention rates, monetization capability, and real reputation.
Some media outlets blame Character.AI's downfall on "the end of the AI emotional companionship sector," but a careful examination of its product quality makes it easy to see that this is nothing more than another classic play of "insufficient capability of one's own, yet dragging the macro environment in to take the blame."
Similar views have been repeated time and again in history.
Macro narratives and concrete execution
Whenever a company falls into trouble because of insufficient product strength, the market and internal decision-makers always tend to package it as "industry dividends have peaked" or "the sector itself is a false premise," using the collapse of a macro narrative to cover up the defeat of concrete execution.
An example I have cited many times before occurred in the SLG game sector.
When Archosaur Games' Under the Heaven performed below expectations in the market, many strategic analyses went along with the tide and blamed it on "the Three Kingdoms SLG market having already been carved up by leading products and being completely saturated." However, later rising stars such as Three Kingdoms: The Strategy of the World broke the deadlock in one stroke and achieved impressive results, and only then did everyone suddenly realize: so-called "market saturation" was merely a lie used to gloss over insufficient product strength and failure to hit user pain points.
Another example occurred on the eve of the rise of mobile games.
When mobile games had just begun to sprout, an internal analyst at a giant that had once dominated the PC game era, in order to explain the failure of the company's own mobile game trial product, solemnly wrote a research report asserting that "mobile game users are merely seeking novelty and will be fleeting like browser games; PC games remain the unshakable absolute core."
This report delighted the traditional PC game business bigwigs within the company and naturally delayed the entire company's strategic transformation. As everyone knows, the mobile game era arrived in full force with irresistible momentum, and this once-prosperous PC game overlord quickly became ordinary amid the self-deceiving slow boiling of the frog.
As a result, a large number of external media outlets and internal analysts began to find excuses for it with "high emotional intelligence," putting forward arguments such as "large models are only infrastructure; the key lies in who controls the final scenarios and traffic entrances." This kind of statement not only successfully soothed anxious executives but also provided perfect spiritual comfort for AI's slowness.
However, reality is very cruel.
Because they invested too slowly and stood on the sidelines during the early stages of AI infrastructure and foundational R&D, they later missed the head start when the computing power explosion arrived, saw the price of compute cards soar, and ultimately had to spend far more than before to rent computing power over the long term. Although the giants later learned from their painful lessons and belatedly mended the fold, the window period lost earlier to strategic justification ultimately turned into a cost paid in real money and the pressure of a passive catch-up.
From "the Three Kingdoms SLG market is saturated" to "PC games will always be the core" to "large models don't matter, the entrance is everything," the essence of these industry analyses and internal narratives is nothing more than the honeyed words analysts use to provide emotional value to their bosses when the giants are half a beat behind in the tide of the times.
The bankruptcy and demise of Character.AI is the same.
Yet the reality is that ACG (anime-style) games have long proven, through extremely mature commercial implementations, that "emotional companionship" is not only not a false proposition, but the strongest consumption engine.
If you look closely at the popular ACG games that dominate the market, you will find they all share one extremely strong commonality: players have an exclusive "avatar" in the game, and use this avatar as a medium to engage in direct communication and emotional interaction with the characters in the game.
Even more ingenious is that this "avatar" setting carries very little information, and is even a completely invisible design, for the purpose of making it easy for players to project themselves into it:
•Occlusion and blank space: For example, the Doctor in Arknights always wears a hooded trench coat with the hood covering the face, so the face cannot be seen;
•Scribbled figure: For example, the protagonist in Blue Archive is officially given only a scribbled figure image;
•Mute and weak presence: Even if characters like FGO's Gudako/Gudao have basic character art, they are basically "mute" in story dialogue, and the multiple-choice options are only used to express attitude rather than to define personality.
This kind of weak-protagonist setting that "spares every word" and deletes whatever can be deleted is by no means the writers being lazy; it is an extremely brilliant psychological mechanism—dissolving the interference of the "third party" and maximizing the player's own sense of immersion.
In the design logic of ACG games, player identity can be roughly divided into three categories:
1.Bystander: Observing the world from a god's-eye view, as in traditional JRPGs. Although this makes it convenient to tell grand narratives, the player's sense of participation and emotional attachment is weaker;
2.Actor: The player controls a specific protagonist, but the protagonist has a strong personality of their own, and the experience is middling;
3.Self (self-projection): The player is the protagonist, with the highest degree of freedom and the strongest sense of immersion and emotional engagement.
Looking across the ACG blockbusters that have truly topped the market and remained popular for a long time, without exception they all place the player in the "self" dimension. Even if some games have to give the protagonist a certain appearance due to anime adaptation or 3D modeling needs (such as Genshin Impact), they still exercise great restraint in dialogue and personality.
Because game companies are extremely clear: what players need is not to watch "someone else" fall in love on screen, but to have "themselves" sincerely chosen, accompanied, and loved by the characters on screen.
This "character-player" emotional companionship channel, which ACG games have verified for decades, is where the real gold mine lies. Compared with this deep immersion and emotional bond, Character.AI's text chat—with no worldview support, no image projection, and only a cold chat box—is simply like a crude paper toy.
Low tech, but high experience
You might say that the "companionship" in traditional ACG games is not intelligent enough, and is merely a set of preset program logic.
It is an extremely mature and highly coordinated experience system, with almost no obvious shortcomings:
•Visually: there are exquisite character art created by top illustrators, Live2D and even 3D dynamic modeling;
•Aurally: there is soul-giving empowerment from a luxurious voice actor lineup; a single exclusive birthday greeting voice line can immerse players in it;
•Mechanically: there is a series of mature companionship feedback such as the oath system, affection cultivation, and holiday-exclusive stories.
By contrast, Character.AI's "companionship" not only severely detaches from the support of these rich senses, but also bets everything on the single track of "text dialogue." Yet the result is obvious to all—this sole "text strength" is not actually that strong in terms of technology and experience.
The underlying large model that reveals flaws after a few more exchanges, frequent logical confusion, and a long-term sense of companionship lacking memory make this "text board," which should have been the core selling point, actually full of holes.
How could this one-dimensional and mediocre-quality experience possibly defeat mature products that have already pushed visual, auditory, and emotional resonance to the extreme?
Character.AI's so-called 18 million AI characters, in terms of real experience and value combined, are not even as good as the 130 meticulously crafted characters in Genshin Impact—because the former's experience is paper-thin bloated illusion, while the latter's companionship is genuine and solid.
Conclusion
Character.AI's bankruptcy filing is not only not the end of the AI companionship track, but rather a rational purge of "low-quality blind expansion." It pours a basin of cold water on the entire industry: merely throwing out a cold text chat box from a large model is not companionship at all; zombie bots piled up from a vast number of makeshift operations cannot support a high-valuation business empire either.
Users have never given up their obsession with and desire for virtual emotional companionship, and the commercial miracle of ACG games remaining popular for decades has long proven this.
The future "AI companionship" will eventually find its correct evolutionary posture—it should not be a lofty technical experiment, but must combine high-level visual and auditory presentation, deep worldview carrying capacity, and large-model interaction that better understands human nature and memory. Only when true "intelligence" fully merges with the mature "companionship mechanisms" of ACG games will the true era of AI companionship have just begun.
Has published 91 articles on Huxiu. This content comes from the internet and the original link; the views represent only the author and do not represent Huxiu's position. If there is any copyright issue, please contact hezuo@huxiu.com, and we will verify and handle it promptly. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]