AI-Spielzeuge im Trend: Startup-Strategie oder Kapital-Wettlauf?

Image source: Haivivi Official Flagship Store

Preferred for entrepreneurship by executives from large companies, with good market development prospects, favored by capital, no technical barriers, and easy product implementation... With a series of "halos", AI toys have become a highly - anticipated entrepreneurial hotspot.

From the cooperation between Volcengine (under ByteDance) and Espressif Systems to launch the AI toy development kit "Miao Ban", to the cooperation between Baidu Smart Cloud and Shifeng Culture to launch "AI Magic Star", etc. As a new entrepreneurial and capital - chased trend, according to IT Juzi data, this track has attracted 96 investment institutions, including leading investors such as ByteDance, Lenovo Capital and Incubator Group, and JD Technology.

Seeing this situation, even in a relatively tight economic environment, there are still many individual entrepreneurs who see the development potential of this track and resolutely plunge into the entrepreneurial wave of AI toys.

However, success and failure can both be attributed to AI. Since the beginning of this spring, free and open - source large models like DeepSeek have accelerated the popularization and iteration of AI toys, further lowering the technological threshold for entrepreneurship. But at the same time, they have also brought problems such as product homogenization, lack of innovation, and uneven quality to such products.

Behind this seemingly booming new consumer market, there are actually countless struggles and dilemmas of start - up companies.

When AI Toys Become an Entrepreneurial Hit

"AI toys have completely exploded! A company founded by a former partner of Tmall Genie shipped 200,000 AI dolls crazily in a year and received 200 million yuan in financing from top VCs such as Sequoia Capital and CICC."

"Leading brands in the market rely on the 'AI + plush toy' emotional companionship model, with monthly sales exceeding 10 million yuan, annual sales on Douyin exceeding 20 million yuan, a premium of 300%, and an average of nearly 200 daily interactions per user. The market's explosive power is amazing. As an upcoming market worth hundreds of billions, the industry's future is promising."

"Looking for entrepreneurial partners: Venturing into the AI hardware entrepreneurial track. The first product is a smart cotton doll movement, supporting AI voice dialogue and touch action feedback, targeting fans of self - insert fanfiction, otome games, anime, and star merchandise."

AI toy entrepreneurship investment promotion. Image source: Xiaohongshu

As described in the above investment promotion information, in the current relatively tight economic environment, the seemingly emerging track of AI toys has quietly become a high - quality entrepreneurial project in the eyes of many.

When it comes to AI toys, we have to mention the executives who left large Internet companies to start their own businesses and made AI toys popular. From the AI start - up company YueRan Innovation founded by Li Yong, a former partner of Tmall Genie, to Beijing Singularity Lingzhi Technology Co., Ltd. founded by Bao Ta, a former vice - president of Meituan, and then to Shenzhen Yijiayiban Intelligent Technology Co., Ltd. founded by Guo Zhi, a former general manager of the Global Solution Innovation Department of Microsoft DPS... There are countless such cases.

There is indeed a certain practical basis for executives to focus on this market. Platform data shows that in the first half of 2025, the sales volume of AI toy products increased by 6 times compared with the previous period, and the year - on - year growth rate of AI toy sales exceeded 200%. Founder Securities estimates that if the penetration rate of domestic AI toys reaches 20% to 25% in 2028, the market scale will exceed 30 billion yuan.

The key is that in 2025, the investment and financing in the AI toy track was extremely hot, and it has become a new trend highly chased by capital: Xuanyuan Technology completed two rounds of financing totaling over 100 million yuan in less than half a year; Luobo Intelligence completed angel - round financing of tens of millions of yuan; YueRan Innovation completed a Series A financing of 200 million yuan; Beipei Technology's cumulative financing reached nearly 10 million US dollars...

Since the beginning of this spring, thanks to the empowerment of large - model technologies such as DeepSeek, the functions and interactivity of AI toys have been improved unprecedentedly, making this market full of unprecedented vitality and potential.

The application of large - model technologies such as DeepSeek not only enables AI toys to understand more complex instructions but also allows them to show more delicate and personalized emotional interactions. This technological innovation has greatly enhanced the user experience, making AI toys no longer just cold electronic devices but partners that can perceive, understand, and respond to human emotions.

As a result, many innovative products have emerged in multiple fields, including emotional companionship, educational interaction, and IP linkage. For example, the Tom Cat robot is equipped with an "AI emotional engine", which can recognize the emotional state by analyzing children's voice intonation, conversation content, and facial expression changes and dynamically adjust the response strategy; the chess - playing robot launched by Yuanluobo SenseRobot has built - in functions for Chinese chess, Go, chess, Gomoku, and fun card programming, constructing a "low - frustration growth system"; the second - generation products of the CocoMate series launched by YueRan Innovation have deeply cooperated with the Ultraman IP, making the classic plots and gameplay of Ultraman into "NFC theme cards".

Image source: Yuanluobo official website

Of course, for entrepreneurs attracted by this market boom, the consumer group is not limited to children. The emotional needs of female and elderly users provide a broader imagination space for the AI toy market.

It is even simpler to realize these entrepreneurial ideas. Many AI toys use the method of connecting to external large models, eliminating the need for independent research and development of complex underlying models, which greatly reduces the R & D cost; start - up companies can also cooperate with IP owners to quickly open up the market with the influence of mature IPs, reducing the time and cost of building their own brands... In other words, this is a light - asset entrepreneurial project that is relatively easy to implement. Entrepreneurs only need to focus on the creative design of products, the precise positioning of market strategies, and the optimization of user experience.

As an entrepreneur described, the definition of an AI toy is: "Something that can talk, has self - awareness, and can meet people's emotional needs."

Low Threshold, but Entering Means Becoming Cannon Fodder

The question is, has the golden age when one can easily get rich through AI toys really arrived? The answer may not be so optimistic.

Looking at the current AI toys on the market, most of them are priced below 1,000 yuan. The prices of some high - end customized or specially - functional AI toys can rise to several thousand yuan or even higher. However, compared with high - end intelligent devices like AI robots, which can cost tens of thousands or even hundreds of thousands of yuan, AI toys still maintain a relatively affordable price.

Image source: AI Research Institute

This price advantage allows consumers to experience the wonder and fun of future technology at a very low cost, which is the key for AI toys to open up the consumer market.

From the perspective of consumers' usage experience, it is unrealistic to regard AI toys as intelligent playmates. Although manufacturers promote them vigorously, their products are far from being able to completely replace real playmates or carry deep emotional sustenance.

On the Black Cat Complaint platform, a consumer accurately pointed out many pain points of AI toys: "I bought a teddy bear for my child, which was advertised as an intelligent AI toy. I set it to use DeepSeek for answers, but the answers really didn't seem to be from DeepSeek. They were very monotonous and incomplete. Some questions were said to be unanswerable. When I asked it to tell a 20 - minute story, it could only tell one and then stopped. It was not intelligent at all. The seller said it had a long - term memory function, but after my child learned new English words two days ago, the teddy bear forgot them two days later. I think the functions are not as powerful as the seller claimed. I applied for a return on the 8th day, but they said I was overdue. The merchant really exaggerated the functions, and the bear also shed fur. What if my child inhales it into the nose while sleeping? Recently, it sometimes couldn't connect to the Internet and didn't respond..."

From the perspective of entrepreneurs, all this is predictable. Zinc Scale found that in many posts looking for AI toy entrepreneurial partners, some entrepreneurs admitted that making AI toys is actually equivalent to "miniaturizing devices like Xiaodu, Tmall Genie, and Xiaomi AI Speaker and connecting them to DeepSeek". With this as the core and adding any unique appearance, it becomes a so - called "intelligent toy".

"This is not a good project, and it will definitely be extremely competitive. If a start - up brand has no sales volume, no breakthrough innovative technology, no brand, and cannot reduce costs, it simply won't survive." Some industry insiders directly pointed out the core problem faced by the AI toy industry at this stage: "The main problem with general - purpose AI intelligent hardware is the lack of technical barriers. Everyone connects to open - source large models like DeepSeek and then conducts simple customized development, resulting in serious product homogenization in the market and a lack of real innovation."

Feedback from some users

On this basis, one option is to build a brand and attract and retain consumers through a strong brand effect, enabling consumers to remember and trust the brand among numerous homogeneous products. However, this development path is costly and time - consuming, difficult for most start - up companies to afford, and not necessarily successful.

Another option is to engage in a price war and attract consumers with ultra - low prices. However, this often comes at the expense of product quality and user experience. Moreover, even if one wants to engage in a price war, they can't compete with the products made by the original manufacturers in Huaqiangbei.

Therefore, some people also say that when it comes to AI toys, selling services is better than selling products, and selling IP is better than selling services. "If your product has no technical barriers, it is easy to be copied. Even if you quickly expand the market, the first - wave sales will be the limit, because in the second wave, you may have to compete with hundreds of competitors."

It's Better to Tell a Good Story to Capital First

Of course, for consumers, the most core question is why they should choose such AI toys, especially when they fail to fully meet expectations and are even disappointing in some aspects.

AI toy products are often evaluated from three dimensions: value perception, practical utility, and emotional satisfaction.

First, when consumers find that the so - called "intelligent" answers are monotonous and incomplete, and even the promised basic functions cannot be stably implemented, their value perception of AI toys will be greatly reduced.

Second, intelligent hardware products such as smart speakers, mobile phones, and tablets that people are accustomed to already have relatively mature voice - interaction functions. If AI toys cannot perform better or provide a unique experience in related functions, their practicality seems useless.

Finally, some start - up companies hope to obtain authorization for mature IPs to endow their AI toy products with special emotional value. However, what really touches users is not only the appearance image carrying the IP but also the story core strongly associated with the IP. Relying solely on the external IP label, it is difficult to establish a long - term emotional connection with consumers.

The consumer market needs more cultivation time. Image source: Xiaohongshu

Moreover, some consumers are also worried about the potential safety hazards of AI toys after being connected to the Internet, especially before the industry has relatively mature safety standards and effective regulatory mechanisms. The awareness of privacy protection is relatively weak. When AI toys collect and process user data (especially children's voice and behavior data), if the safety measures are not in place, it is easy to lead to information leakage or malicious use.

It is not difficult to see that the current AI toy industry is still in the early exploration stage, facing many challenges and uncertainties. Some people who left the industry after setbacks even sighed that this industry is actually a "pseudo - demand". They believe that although the concept of AI toys is hyped up in the market, consumers' actual demand for such products is not as strong as expected.

More importantly, if an individual entrepreneur wants to enter this already crowded and highly competitive track, instead of becoming "cannon fodder" accelerating the industry's growth during the market cultivation process, it is better to come up with a good story that can impress capital before product development.

After all, the momentum of capital entry is indeed very strong. Those investment cases may indicate that this is really the "first pot of gold" that can be earned at present.

This article is from the WeChat public account "Zinc Scale", author: Meng Huiyuan. Republished by 36Kr with permission.