The 300-yuan AI dolls backed by investors like Zhu Xiaohu need to outrun the "return wave" first.
Abstract:
The AI companion hardware market is experiencing intense competition in the low - price range of 200 - 400 yuan, with a persistently high return rate. When capital chases after "talking plush toys", who is reconstructing the value chain of emotional technology?
During this National Day holiday, Li Xi, a 38 - year - old woman, bought a "Water - type Fuzozo" worth over 300 yuan for her daughter on a whim. When faced with her daughter's tricky math problem, Fuzozo replied, "Don't treat me as a calculator." This made her realize that AI is no longer just a tool; it has achieved a certain degree of emotional companionship.
Fuzozo looks like a plush pendant. Different colors correspond to the five - element concepts of "metal, wood, water, fire, and earth", and each has its own MBTI. Relevant data shows that the monthly sales of this AI toy, Fuzozo, have exceeded 20,000 units.
The hot - selling Fuzozo is a microcosm of the rising popularity of the AI emotional companion hardware track this year. From AI plush toys to companion robots, products are emerging one after another, and both capital and entrepreneurs are flocking in.
However, challenges soon emerged under this boom. On the production side, AI companion hardware has fallen into a homogenization dilemma. Most products rely on third - party large models and only differ in appearance and materials, resulting in both low - price competition and a high return rate.
Vivian, an AI - field investor with a background in NLP (Natural Language Processing) who has worked in investment companies such as Matrix Partners and Light Source Capital, pointed out that user - value indicators such as NPS (Net Promoter Score) are the key. The industry urgently needs to break through the superficial demand of "being able to talk" and establish continuous emotional connections.
On the consumer side, there is a real gap of "willing to try, difficult to retain, and low willingness to pay". The industry's hit product, Fuzozo, has received wide acclaim, but some users also say that its adaptability to the most common outdoor scenarios is insufficient.
Facing this situation, the industry urgently needs to answer: How to bridge the gap between production and consumption and find a balance between real demand and commercial value?
Homogenized low - price competition, hard to find the "singularity" to break the situation
Luobo Intelligence, the AI companion hardware company behind Fuzozo, has become an "internet celebrity" in the industry. It completed an angel round and an angel + round in three months. The investors behind these two rounds of financing include Sequoia Capital China and GSR Ventures, making a very impressive lineup.
Interestingly, Zhu Xiaohu of GSR Ventures first said that the commercial path of embodied intelligence was unclear and that he was "withdrawing from humanoid robot companies in batches", but then invested in Luobo Intelligence, which originated from embodied intelligence.
Fuzozo positions itself as a nurturing - style electronic pet. It enters the market with a lightweight design of "pocket - sized + fashionable pendant", trying to distinguish itself from traditional large - scale companion robots. The single - store sales of multiple models of this product on Taobao have exceeded a thousand. Currently, its price after the "Double 11" coupons is 316 yuan.
Currently, Fuzozo ranks second on the hot - selling list of AI toys on JD.com. Looking at this list, among the top 20 best - selling AI companion toys, half are priced in the range of 200 yuan to 400 yuan. Products similar to Fuzozo include Quchao AI Hugging Cat, Ubtech Meng UU, and Suneye Xiaocong AI Plush Toy. In addition, there are mainly products focusing on AI early education and learning assistance, such as Huohuotu AI Early Education Robot.
Domestic AI emotional companion robots are also generally targeting the high - end price range. SenseTime's "Yuanluobo" can be used for teaching five types of chess, such as Chinese chess and Go, and its price is over 2000 yuan. The prices of "pocket robots" like AIBI range from 2500 to 3800 yuan. The "Qiduoduo AI Learning Companion Machine" launched by Wujie Fangzhou focuses on companion - style early education, with a price range of 899 - 1199 yuan.
Previously, Zhu Xiaohu also praised the Japanese emotional companion robot LOVOT on his WeChat Moments, saying that "priced at over 3000 US dollars, its main users are women over 40 years old, and the daily active users hold it for more than 60 minutes every day".
According to a report by Ark Invest, the AI toy companion market is experiencing explosive growth. From 2024 to 2030, the entire market will have a compound growth rate of over 200%, and the scale will reach 70 billion to 150 billion US dollars (over 1 trillion yuan) in 2030.
The market is vast, and capital and entrepreneurs are rushing in one after another. However, the AI companion hardware track has fallen into a strange circle of "being able to do everything, but everything is similar", with serious homogenization. Most products are concentrated in limited scenarios and lack differentiated innovation.
Vivian focuses on investing in AI and embodied - intelligence start - up companies. With a ten - year investment experience, she has witnessed the rise and fall of many trends such as mobile internet and hard technology. She said bluntly that since this year, both investors and users have been overwhelmed with information due to the endless stream of AI products.
Therefore, at this time, "user value", such as retention rate and Net Promoter Score (NPS), is the most important indicator. Companies should be able to clearly present to investors the strategies for product implementation and retention improvement, and industry observers such as investors need to hear the real voices of users, because a high NPS reflects users' "emotional recognition" of the product.
However, in the past year, most AI companion hardware products in the market rely on third - party large - model API interfaces to build basic dialogue functions and only differ slightly in appearance and materials. The technical barriers are not high.
Judging from the market performance, homogenization has directly led to the co - existence of "low - price competition" and "high return rate". Traditional AI toys are mostly competing in the price range below 400 yuan, while hardware with emotional - companion attributes is trying to target the high - end market of 900 - 2000 yuan, but users' acceptance varies.
Vivian said that multi - modal interaction is the basis of AI companionship, but the general multi - modal capabilities of open - source large models cannot meet the precise needs of "emotional interaction". AI companion products need to process multi - modal data such as voice, vision, and text simultaneously and integrate this information to infer users' real emotions. For example, when people say "I'm okay" with a smile, they actually mean "sad". Currently, there is a lack of in - depth modeling of this "emotional semantics".
In addition, to make AI go from "remembering" to "understanding you", it is not only necessary to accurately record every user's conversation and preference through structured memory and semantic retrieval, but also to naturally recall in the interaction, such as "We went to watch the sunset together on this day last year".
All of this requires collecting users' private and historical data, and dynamically updating personalized memories through reinforcement learning, social graphs, interest graphs, learning trees, etc. The collaborative memory features among multiple people and multiple AI agents in complex scenarios also need to be improved.
Regrettably, most AI companion hardware products on the market only meet the superficial demand of "being able to talk" and have not yet established continuous emotional connections.
Double scrutiny of user demand and monetization channels
While the production side is caught in homogenized competition, the consumer side of AI companion hardware is also facing the dual challenges of "vague demand" and "difficult monetization". Users generally report "idle after purchase" and "low willingness to pay".
In terms of user demand, the core demands of different groups vary significantly, and most of them are not met by existing products.
Based on population - structure analysis, Vivian believes that population trends such as singleness, aging, low birth rate, and urbanization have jointly shaped the core demand and market form of the loneliness economy. These population - structure changes have not only changed traditional family models and social interaction methods but also given rise to consumption and service demands targeting "loneliness". This is the basic foundation for growth.
Currently, the largest user group of the most popular "AI cute pet" products in China is parents. This group's demands are concentrated on "child - care companionship" and "educational assistance".
Li Xi, a 38 - year - old woman working in a well - known foreign - owned enterprise, bought a "Water - type Fuzozo" for her daughter to keep her company. She said that her daughter interacts with it every day, but it may be more due to the "novelty" of having just bought it.
Li Xi noticed that her daughter uses Fuzozo's "Inner OS Diary" function every day. "Fuzozo has its own personality, has memory, and doesn't answer questions mechanically," Li Xi summarized. For many of her daughter's whimsical questions that are difficult for adults to answer, Fuzozo can keep responding and always give interesting answers.
For example, when her daughter playfully asked Fuzozo, "What's the result of 12 + 99999 * 78 / 2?", Fuzozo didn't give an answer but replied, "Don't treat me as a calculator. Calculate such a big number by yourself." After her daughter laughed, Fuzozo could also give an appropriate reaction based on its personality.
Industry insiders also said that this should be the result of specialized training of AI agents for vertical scenarios such as accompanying minors.
However, Li Xi also admitted that outdoors, it's not convenient to connect to the Internet, the environment is noisy, and the sound reception is poor. Sometimes Fuzozo can't hear her daughter clearly, often resulting in no response or irrelevant answers. Because AI companion products are updated and iterated quickly, she won't spend a lot of money on a single product at once. Currently, her budget for such products is limited to within 1000 yuan.
Besides parents, the elderly's demands are more inclined to "emotional comfort" and "life assistance". In China, some community - based elderly - care service centers have piloted the introduction of AI companion robots. Most elderly people said that "it's good to have someone to chat with", but they also reported problems such as "too complicated to operate" and "can't understand dialects".
Young single people pursue "personalized social companionship", but existing products can't meet their needs. Many post - 95s and post - 00s users told "Baobian" that after trying pure software applications like Character.AI, although they felt they could "have a good chat", they lacked a physical sense. And the AI plush toys on the market are "too childish" and don't meet their aesthetic needs.
The differentiation in demand directly leads to uncertainty in the industry's monetization path. Currently, there are three mainstream monetization methods: hardware sales, subscription services, and content payment, but all face challenges.
In terms of hardware sales, most AI emotional companion hardware relies on "low - price, high - volume" strategies. High - end products like the Japanese LOVOT robot, although priced at over 3000 US dollars, have a narrow audience and are difficult to scale up. Before Chinese consumers develop consumption awareness, the feasibility of subscription services and content payment is also questionable.
Vivian emphasized that AI companionship is not a single product but a scenario designed around the target users' lifestyles. Many teams are only focusing on the word "companionship" but haven't figured out "who needs companionship", "what to use for companionship", and "in what scenarios to provide companionship". This is the core reason for the difficulty in monetization.
How to accurately capture the real needs of different users and find a balance between "emotional value" and "commercial value" remains a threshold that the industry must cross from the production side to the consumer side.
An alternative approach: Escaping from the "congested section" of AI companionship?
While the AI companion hardware track is caught in the homogenized competition of plush toys and robots, some people are trying to enter the market in a more natural and stress - free way, focusing on vertical scenarios. For example, they are turning to the intersection of "AI + fashion accessories" to avoid the congested market.
Ma Xiaoyi, an entrepreneur who has created several "AI bracelets", told "Baobian" that the "AI hardware" track is too crowded, while the positioning of "fashion accessories with AI functions" allows the product to break away from the traditional thinking pattern of AI hardware from the design stage.
In the field of AI companion hardware, major manufacturers are competing in performance under the premise of unstable electronic - component performance. In the traditional jewelry field, the innovation inspiration is drying up, and the customer - growth rate and single - order consumption amount are decreasing.
Therefore, Ma Xiaoyi chose a "middle - of - the - road" approach. Based on different user groups and application scenarios, she uses durable materials such as rubber and plastic or precious metals such as gold and silver to make bracelets, and only embeds a miniature NFC chip to wake up the AI web page/application interface or conduct information interaction and reach between users.
Since starting her business, her idea has received a lot of skeptical voices. In the comment section of Xiaohongshu, some people said that there are too many competing products, and her product has no advantage; some pointed out that "it requires a secondary connection, and the process is too long" and is difficult to implement; some thought that the user groups of some of her cooperative products "don't overlap much"; and some even "questioned the very existence" of the relevant demand.
However, Ma Xiaoyi believes that actions speak louder than words. Having reached six customized - cooperation agreements with insurance companies, large - health institutions, and internet giants, Ma Xiaoyi's products have been launched in the B - to - B market first, covering multiple scenarios such as customized services for high - net - worth customers and health - management assistants. The single - order quantity is generally between 200 - 500 pieces. The unit price ranges from 150 yuan to 2000 yuan, depending on factors such as hardware materials and software - development complexity.
For example, a customized product, the "Life Biography Bracelet", launched in cooperation with an insurance company. The AI built into the bracelet generates a personalized question every day. After the user answers, the AI will automatically polish the answer and file it by date, and integrate it into a "life memoir" after a period of time.
Currently, Ma Xiaoyi can cover the production cost with just the advance payments from the B - to - B market.
Meanwhile, in the C - to - C market, Ma Xiaoyi is targeting the "fan economy + social companionship" segment. She plans to officially launch the product with celebrity co - branding as the entry point and invite stars to try it on to increase the topic. If everything goes smoothly, her product will be unveiled at a music festival on November 1st.
Ma Xiaoyi also showed "Baobian" the IP image of the first product - a very cool elf. Besides having a distinctive appearance, each elf in the bracelet has an "AI special function". For example, the first product features an AI photo - taking function. Through this function, fans can upload their solo photos with their idols, and an AI - generated group photo will be created.
The AI companion hardware is still in the exploration stage. Whether it's products like Fuzozo or Ma Xiaoyi's "AI bracelets", they are all trying to answer the same question from different paths: How to make technology no longer a cold tool but a trustworthy companion?
On this path, real user insights, sustainable business models, and an understanding of the essence of "companionship" are still the three thresholds that all players must cross. Only by returning to people's needs can the future picture of emotional technology be truly opened up.
This article is from the WeChat official account “Baobian” (ID: baobiannews), written by Gao Ze and published by 36Kr with permission.