上线仅8个月 米哈游联合创始人做的AnuNeko停运 AI陪伴为什么这么难？

This article covers the industry event of AnuNeko, an AI companion product launched by HoYoVerse co-founder Cai Haoyu, which was shut down just 8 months after its launch. It summarizes the current overall development status of the AI companion track, with key takeaways as follows:

1. AnuNeko was a small-scale experiment in the AI companion space launched after Cai stepped back from daily management at HoYoVerse. Positioned as a sarcastic, humanized chat companion with no productivity features, it gained initial attention for its differentiated character design. However, it ultimately shut down due to its limited single-function design: only offering chat with no extended interactive features to retain users long-term.

2. The entire AI companion industry is still in an exploratory phase. Multiple AI companion apps in China have already shut down; even leading standalone products struggle to retain users long-term despite initial traffic traction, while also facing content security and compliance challenges. That said, authentic user demand for emotional AI companionship persists, and the track continues to attract sustained capital attention, with funding increasingly flowing into AI companion hardware.

This article analyzes the current development landscape, user demand and industry trends of the AI companion track, offering actionable insights for brands planning to enter the AI space. Key takeaways are as follows:

1. On user demand: There is authentic consumer demand for emotional AI companionship, and many users find emotional release with AI products that have distinct, memorable characters. However, pure chat-only product formats cannot retain users: consumers expect more diverse interactive experiences and extended content, and "being good at chatting" alone is no longer sufficient to build core competitive advantage.

2. On industry trends: Purely software-based AI companions currently lack clear commercialization paths and face high content compliance risks, with multiple well-known products already shutting down. Capital has already begun shifting its focus toward AI companion hardware such as AI toys and companion robots.

3. On competitive intelligence: Most leading current players follow a character + story + user-generated content model, while Tencent has taken a lightweight approach by reviving the digital pet concept. Products across all positioning segments are still exploring viable, scalable business models.

This article summarizes the latest developments in the AI companion track, and clearly outlines the core risks and opportunities in the space. Key takeaways are as follows:

1. Risk warnings: Purely software-only AI companion products are increasingly difficult to scale successfully, and face three core challenges: First, chat functionality has become a standard feature of large language models, so distinct character design alone cannot drive long-term user retention. Second, tightening content security regulation has pushed compliance costs sharply higher, and over-blocking of legitimate content is a common issue that harms user experience. Third, traffic acquired from external sources has extremely low retention rates; viral hype fades quickly, and it is difficult to convert fleeting attention into long-term stable users.

2. Opportunity outlook: The AI companion track as a whole is still growing, with funding deal volume in 2026 already exceeding the full year of 2025. AI companion hardware and lightweight products that integrate into existing social scenarios have already produced successful use cases. Compliant standalone AI companion products also have opportunities to capture overflow traffic from general-purpose platforms, and players can carve out niches in differentiated segments based on their unique strengths.

This article outlines the evolving landscape of the AI companion track, offering clear commercial opportunities and transformation insights for factories developing AI-related products. Key takeaways are as follows:

1. Product design and manufacturing opportunities: The AI companion track has seen a clear split between software and hardware. As pure software products shut down one after another, both capital and consumers favor AI companion hardware, which has higher user acceptance for emotional companionship. Factories can leverage their existing production capacity to develop and manufacture AI-focused products such as AI toys and companion robots.

2. Commercial upside: The AI companion track continues to grow, with funding volume increasing year over year, and market demand remains largely unmet. There is still a significant gap in the market for combined software-hardware AI companion products, so factories that enter the space early can capture significant first-mover advantage.

3. Insights for digital and AI transformation: The development of the AI companion track demonstrates that AI commercialization cannot rely on hype alone; products must align with actual user needs. When advancing digital and AI transformation, factories should focus on tangible, real-world use cases instead of blindly chasing trends, and adjust product strategy to match user demand.

This article analyzes the core pain points and trends of the AI companion track, helping AI-focused service providers identify high-potential business directions. Key takeaways are as follows:

1. Industry development trends: The AI companion track has evolved from early hype around concepts into a phase of differentiated commercialization. Growth has stalled for purely software-based companion products, and AI companion hardware has emerged as the new segment favored by both capital and consumers. Service providers can adjust their business focus to develop technologies and services tailored for AI companion hardware manufacturers.

2. Core client pain points: The most pressing challenge for AI companion products today is content security and compliance. Following the implementation of new regulatory requirements, the entire industry has strong demand for content moderation and character review services, as well as technical solutions to reduce false blocking of legitimate content. The second key pain point is low user retention: all players struggle to consistently produce new content that gives users reasons to keep engaging. The third is low traffic conversion: external traffic has very low retention rates, so brands need targeted operational solutions to address this.

Service providers can build tailored offerings around these pain points to accurately meet client demand.

This article outlines user demand and industry challenges in the AI companion track, offering operational and risk mitigation insights for platform operators building AI-focused businesses. Key takeaways are as follows:

1. User expectations for platforms: Users have genuine demand for AI companionship, but are not satisfied with basic chat functionality alone. To retain users, platforms need to offer a rich library of pre-built characters and story content, alongside supporting features for user content creation and social interaction.

2. Core risks to avoid: Content security and compliance is the biggest risk for AI companion platforms today. A large number of products have already shut down or pivoted due to compliance issues. Platforms need to build robust content moderation and character review systems in advance, while balancing compliance requirements with user experience to minimize false blocking of legitimate content.

3. Operational guidance: AI companion platforms typically see high traffic volatility, and external inbound traffic generally has very low retention rates. Platforms should design retention-focused operating frameworks in advance, and keep users engaged with continuously updated content and features. They can also pursue differentiated positioning—such as lightweight digital pets or combined software-hardware models—to avoid cutthroat homogeneous competition.

This paper documents the latest industrial developments in the AI companion track, offering high-value reference for researchers studying large language model commercialization. Key insights are as follows:

1. Latest industry dynamics: The AI companion track is undergoing clear segmentation. Purely software-only companion products have shut down consecutively, and multiple leading standalone products have seen initial hype fade into traffic declines. Capital allocation has shifted from software to AI companion hardware, and the entire industry is still exploring sustainable commercial models, with no widely accepted successful path established to date.

2. Unresolved industry challenges: Two core problems remain to be addressed in the AI companion track. First is balancing content security compliance with user experience: following regulatory tightening, false blocking of legitimate content is an industry-wide issue, and no mature solution for balancing the two has emerged. Second, after conversational capability became a standard feature of large language models, the industry is still searching for answers on how AI companion products can differentiate and improve long-term user retention.

3. Case study resources: This article covers multiple cases of products across different sizes and positioning, including celebrity founder experiment AnuNeko, internet giant offering Maoxiang Xingye, independent developer hit TheOne, and Tencent's revived QQ Pet. These diverse cases provide resources for researchers to analyze the viability of different business models.