Before the spring of 2025 arrives, the first batch of children's AI hardware has already become obsolete | Focus Analysis
Text | Lin Qingqing
Editor | Yuan Silai
One of the groups most in need of liberation by AI must include parents who have to tell the same story ten times over.
At the CES in early January this year, the AI - enabled educational companion market was crowded with Chinese manufacturers seeking better solutions. A new product called “Ling!” from Lingyuuzhou attracted a lot of attention. According to the official introduction, “Ling!” focuses on the concept of “study companion”. Through its built - in wise sages, subject experts, and virtual IP characters, it covers multiple disciplines such as science, biology, English, history, geography, and literature. By means of character performances, interactive stories, and light games, it creates an immersive interactive learning experience for children.
The founder of Lingyuuzhou, Gu Jiawei, is also the founder of Wuling Technology. The latter's AI picture - book machine, Luka Hero, was launched in July 2018. It has multi - sensory interaction capabilities including vision, hearing, and touch, and contains English enlightenment picture - book reading and English learning content.
Reports say that the launch of this product created a new category of picture - book reading robots. In just half a year after its launch, it achieved the top sales on the JD platform among picture - book reading robots for smart devices, with a user satisfaction rate of over 99.8%.
After that, the demand for “AI + children's education” was gradually verified by the market and came into the view of Chinese families.
From smart watches to early - education robots, from storytellers to learning tablets, various types of children's smart hardware products such as smart wearables, smart educational robots, and smart interactive picture books have flooded the market. Among large companies, Amazon has launched products like Echo Pop Kids, Fire HD 10 Kids Tablet, and Echo Dot Kids smart speaker. There are also star startups that have received substantial financing.
However, beyond those inspiring stories, the most successful case comes from a company not overly enthusiastic about changing the world. The Xiaotiancai watch under BBK has been on the market for 10 years and has entered more than 100 countries around the world. By the seventh year of its release, over 20 million units had been sold. Although it only has basic functions such as social networking and calling, the price of its flagship model can rival that of the iWatch SE. It ranks second in the global smart - watch market, and has a dominant market share among users under 14 years old.
Smart watches are among the few surviving children's smart hardware products. More products have been submerged in the wave. According to a recent report by CSDN News, the American AI hardware star company Embodied announced the closure of its operations due to failed financing. Its flagship product, the children's emotional robot Moxie, with a single - unit cost of over $500 and a selling price as high as $800 (about 5,785 yuan as of March 17), and a repurchase rate of less than 5%, had to exit the market. Moxie is a robot designed to provide emotional companionship and social interaction for children. It has functions such as face recognition and voice interaction and can have simple conversations and play interactive games with children. However, its high cost and low repurchase rate made it unsustainable.
With the shutdown of its cloud services, its hardware will become “electronic bricks”. This news has, to some extent, awakened the industry from its crazy fantasy about “emotional companionship”.
“The core reason for Embodied's failure lies in the cultural differences in the market. The demand for emotional companionship is higher in East Asia, but the market in Europe and the United States lacks a fertile ground. In addition, the technology is not mature enough. From AI dialogue and sensors to hardware implementation, there are many setbacks. Coupled with the high cost, it is doomed to be unsustainable,” Wang Gefei from Shouhua Technology told 36Kr. Their company was founded in 2017. They have developed AI robots through cooperation with global top IPs and self - developed AI voice models. In recent years, they have gradually expanded from being a B - end technology service provider to a C - end product provider.
Although there are always new and exciting stories in the smart hardware market, the children's smart hardware market has been underwhelming. According to data from iiMedia Research, the market size of China's children's smart hardware reached 52.2 billion yuan in 2023, but the growth rate has dropped from 35% in 2019 to 12%. The market is shifting from a frenzy to rational competition.
Wang Gefei told 36Kr, “Large companies like iFlytek and Xueersi focus on educational features but lack emotional design. Guangdong's supply - chain manufacturers rely on low - price strategies to increase sales volume, resulting in serious product homogenization. The real differentiation lies in binding IPs for emotional companionship instead of crowding into the highly competitive education market.”
In the marketing field, the purchasing power of children and their parents has always been at the top of the pyramid. However, children's smart interactive hardware has always been a tough nut to crack. The buyers are parents, but the end - users are children who cannot accurately express their needs. High cost and useless functions are the fatal flaws of such products. Apparently, no company has yet produced a product that can fully meet children's companionship needs.
Entrepreneurs are still flocking to this market. Perhaps they should think about simpler and more straightforward functions instead of being carried away by the trend.
Overhyped Technology and a Demand Gap
Parents' insistence on “educational necessities” and children's preference for “entertainment features” create an irreconcilable gap.
The “2024 China Children's Smart Hardware User Research Report” by iResearch shows that nearly 60% of parents complain about the “lack of interactive experience” in products. Moreover, the frequency of children's use drops precipitously with age - children aged 3 - 6 use the products for an average of 45 minutes per day, while those over 7 years old use them for only 12 minutes.
On the market, many smart early - education machines claim to offer a vast amount of educational resources, but in fact, their content is mostly a hodgepodge of simple nursery rhymes and stories, lacking systematicness and pertinence, and unable to meet children's learning needs at different growth stages. For example, the “AI Smart and Fun Pleasant Goat” developed by Alpha Group (a smart toy product integrating plastic plush, IP, and AI) sold out on its first day of launch on Tmall. However, although it combines IP and AI technology, some parents reported “insufficient depth of interaction”. Moreover, the interaction between the product and children is too mechanical, and children tend to lose interest after using it a few times. From the perspective of children's behavior, there is a “three - minute enthusiasm” phenomenon in product use.
In response, Wang Gefei said, “Many product failures are due to 'designing children's needs with adult logic'. For example, they over - stack functions while ignoring the fact that children have short attention spans. Therefore, we try to improve the experience through dynamic sensors and IP - based emotional design - for example, when a child is fidgety, the hardware automatically switches to 'game mode'.”
Ultimately, such products fail to understand children's behavior and psychological logic. Child psychologists point out that children have relatively short attention spans. Although they are highly curious about new things, if a product cannot continuously provide interesting and attractive interactive experiences, they will soon abandon it. For example, some smart toys initially excite children with their novel appearance and simple operation, but after getting familiar with them, children find the gameplay monotonous and lacking depth, so they stop using them. For instance, Wonder Workshop in the United States over - emphasized programming functions while ignoring the operation threshold for young children, and finally, its products became “geek toys”.
On the technical level, there are also many bottlenecks. Voice recognition is an important function of children's smart interactive hardware. Voice - recognition technology is being increasingly used in children's smart hardware products. However, research shows that the accuracy of children's voice recognition is about 20% lower than that of adults, and the misjudgment rate in scenarios with a mix of Chinese and English exceeds 40%, greatly reducing the interactive experience.
The American consumer - embedded voice - recognition manufacturer Sensory Inc. once emphasized “developing a dedicated voice - recognition algorithm for children” to better understand and process children's voice input. The children's educational robot developed by Zero Second Technology in China also faced difficulties in recognizing mixed Chinese - English voices, especially in the common mixed - language scenarios during early - childhood English education. Subsequently, by introducing a voice - recognition engine powered by Microsoft's artificial - intelligence technology, the semantic - recognition accuracy was increased to over 90%.
Children's cognition is still developing, and their language and behavior patterns are completely different from those of adults. Products for children are bound to be more complex and difficult to control. Many designers of children's products are stuck in adult - centric thinking, and such arrogance is doomed to failure. As the CEO of Embodied reflected in the shutdown statement, “We were obsessed with making the robot'more human - like' and forgot that what children really need is'more like a buddy'.”
Natural Supply - Chain Advantages but Content Shortcomings
Compared with trying to meet complex emotional connections, Chinese companies' pragmatism is smarter and more expedient.
According to People's Daily, in Jakarta, Indonesia, a mother came to buy an imoo watch (imoo is the same as the Xiaotiancai phone watch in China) with a bucket of coins. The Xiaotiancai phone watch is very popular among children. Its unique social function allows children to easily add friends and share fun moments through the watch. In school social activities, students with Xiaotiancai watches can better integrate into their peer groups.
Specifically, the Xiaotiancai watch uses the “tap - to - socialize” function to build a closed social network, transforming the hardware into a “communication necessity” and creating a children's social circle to meet children's social needs and form strong user stickiness. Its Southeast Asian version is priced at $50 (compared to Apple Watch SE's $249), supports 12 local languages, and integrates scenario - based functions such as “emergency call” and “homework reminder”. Its repurchase rate exceeds 20%, far higher than the industry average. The essence of this strategy is to solve the most basic social and guardianship problems with extreme cost - effectiveness instead of piling up flashy technical parameters.
The case of Xiaotiancai also provides a simple lesson for Chinese children's smart hardware manufacturers. It never flaunted “AI” but solved the social necessity with the simplest interaction.
In addition, regarding market selection, Wang Gefei also pointed out, “Emotional - companionship products in the Japanese market are more substantial. For example, dynamic sensors are used to simulate the 'cute' behavior of pets. This kind of design is well - received in East Asia but may not work in Europe and the United States. Therefore, our product definition must consider the market characteristics.”
Another approach is vertical - scenario binding. Wuling Technology gave up general - purpose robots and focused on the differentiated path of “AI + picture books”. By scanning picture books to automatically generate interactive stories, it increases the frequency of children's weekly use. This “in - depth experience in limited scenarios” has become an effective way to solve the “three - minute enthusiasm” problem.
For older school - age children, the “hardware + content subscription” model usually dominates. Some large companies, such as BBK's tutoring machine and NetEase Youdao's dictionary pen, have performed well in the market, with continuous growth in user activity and content - subscription volume. Alpha Egg's children's dictionary pen has ranked first in sales in its category on Tmall in the past few years. By including the copyright content of “Oxford Advanced Learner's Dictionary” and dozens of other primary - and secondary - school reference books, it can comprehensively query nearly 70 types of knowledge points.
There will also be some formal innovations in children's products in the future. Gartner predicts that multi - modal interaction and screen - less design will become mainstream in 2025. In recent years, some manufacturers have started new attempts based on this. Some manufacturers (such as Wuling Technology) have launched screen - less storytellers that simulate the “book - turning experience” through tactile feedback. BrainCo in China uses brain - computer interfaces to optimize attention training, increasing learning efficiency by 40%. However, no matter how the technology evolves, the essence is still to return to children's cognitive logic.
When hardware stops chasing technical parameters and focuses on “making children willing to use it”, perhaps this game between supply - chain advantages and content ecosystem can truly find a balance.
Although new players are still entering the market, some manufacturers that once focused on children's smart hardware are looking for new ways out and are transitioning to the female emotional - companionship market. Wang Gefei admitted, “(In the past,) the good sales of our children's products were essentially driven by IP, and the emotional functions have not been verified. However, the demand from adults definitely exists. If the authenticity of AI dialogue reaches 90%, the hardware will be irreplaceable. We are betting that the technological breakthrough will be much faster than expected.”