Priced at tens of thousands and with a gross profit margin of over 90%, can AI toys produce the next Pop Mart?
Investors collectively misjudged the situation, and missing the mythical pop culture giant Pop Mart has become a classic negative case in the venture capital circle. This time, capital seems to be adopting the strategy of "it's better to invest wrongly than to miss out."
On August 25th, AI toy startup Haivivi officially announced the completion of a Series A financing round worth 200 million yuan. This round of financing was led by funds under CICC Capital, Sequoia China Seed Fund, Huashan Capital, and Joy Capital, with participation from CMBC International and others. Existing shareholders such as Brizan Ventures also continued to increase their investments.
Notably, this is the second round of financing Haivivi has received this year, just four months after its Series A2 financing. Moreover, it has set a new record for the largest financing amount in the AI toy sector.
Haivivi is not the only company favored by capital. The AI toy sector, which also sells products based on the appeal of emotional value, has attracted nearly a hundred investment institutions so far. According to IT Juzi data, 96 investment institutions have participated in the AI toy sector, including top - tier investment institutions and internet giants such as Sequoia, Shunwei, Wuyuan, ByteDance, and JD.com. Since 2024, both the number and amount of financings in the AI toy sector have increased significantly.
Capital is desperately trying to secure a position, betting on the next Pop Mart - like legendary era.
Capital Collectively Bets on AI Toys
If 2024 was the first year of AI toys, then 2025 is the year when AI toys are booming. AI toy players range from startups to listed companies and internet giants, with financing, mass production, and sales in full swing.
Former executives from large companies have ignited the AI toy startup scene. The founders of several startups that received financing this year all come from top - tier companies.
Li Yong, the founder of Haivivi, has served as the first marketing director of Smartisan Technology, was involved in 360 Marketing, and was a partner at Alibaba's Tmall Genie. In 2021, Li Yong left Alibaba to found Haivivi and received the first - round financing just one month later.
Data from Tianyancha shows that since its establishment in 2021, Haivivi has completed a total of six rounds of financing worth hundreds of millions of yuan. The investors include top - tier investment institutions such as CICC Capital, CMBC International, Hong Kong Science and Technology Fund, Sequoia China, Huashan Capital, Joy Capital, and BlueRun Ventures.
Another company, VitaPower, which received two rounds of financing in three months, has Yu Yinan, its founder and CEO, who has worked at companies like Baidu and Horizon Robotics. The backers include star investment institutions such as Hillhouse Capital and Capital Today.
He Jiabin, the founder of Ropet, an AI robotic pet company that completed a tens of millions of RMB Series A1 investment in early September, has a career spanning Microsoft, Baidu, and ByteDance.
In addition, for Luobo Intelligence, an AI emotional companion hardware company that received angel - round financing in June, its founder Sun Zhaozhi was previously the person in charge of product design at XPeng Robotics. A widely - circulated investment story about Luobo Intelligence is that Zhu Xiaohu, an investor who once "poured cold water" on the robotics sector, decided to invest after just ten minutes of conversation with the company.
Top - tier investment institutions are flocking to the AI toy sector. With the support of capital, startups are accelerating their expansion.
The sales volume testifies to the market's popularity. BubblePal, the world's first AI toy launched by Haivivi in July last year, sold 200,000 units in one year, and the cumulative sales volume has exceeded 250,000 units so far. Calculated at its price of 389 yuan, the sales revenue of this single AI toy product has reached 100 million yuan.
Regarding the second - generation product, the CocoMate series, launched in August, a staff member of Haivivi told Tech Planet that the new product was sold out as soon as it was released. It is still in the pre - sale stage and is exclusive to online platforms. CocoMate and BubblePal have slightly different functions. CocoMate is the world's first end - to - end AI interactive toy that can more accurately capture people's tones and emotions. The first - generation product, BubblePal, is relatively weaker in dialogue response, but it can be hung on the neck of any doll, bringing the favorite doll to life.
Fuzai, an AI companion product under Luobo Intelligence, is priced at 399 yuan, similar to BubblePal. Data shows that during this year's "618" shopping festival, its sales volume ranked third in the trendy toy category on JD.com, only behind products from Pop Mart and Miniso. Its founder revealed that 5,000 units produced in July were sold out, and the production capacity of 10,000 - 20,000 units in August was almost sold out.
Fuzai comes in five colors, corresponding to the five elements in Chinese culture (metal, wood, water, fire, and earth). On the product page of JD.com, Fuzai is labeled as an "AI robot that can talk back to the boss." The salesperson said that the wood and fire - colored products are selling better.
Listed companies such as Tom Cat and Aofei Entertainment have also achieved growth through AI toys. Data shows that Tom Cat officially launched its AI emotional companion robot at the end of last year. On the first day of the "6.18" e - commerce promotion this year, its single - live - stream sales exceeded 1.8 million yuan, topping the list of AI toys on Douyin. Aofei Entertainment started mass - producing its first batch of AI - enabled Pleasant Goat toys at the beginning of the year, and it has already iterated over three generations of products, showing an obvious growth trend.
High - end Products Priced at Tens of Thousands, with Gross Margins over 90%
The concept of AI toys is broad. It is not a simple combination of AI and toys, nor is it limited to traditional children's toys. AI plush toy pendants, AI plush toys, children's companion robots, AI pets, AI robots, and AI trendy toys all fall into the category of AI toys.
The market positioning of AI toys has expanded from children to people of all ages, including young people and the middle - aged and elderly. The main business models are hardware sales and subscription services.
Currently, most AI toys on the market are priced between 300 - 400 yuan, mid - end products are around 1,000 - 3,000 yuan, and some high - end products are priced at tens of thousands of yuan. Take LOVOT, an AI companion robot launched by Japanese company GROOVE X, for example. Its basic model costs about 30,000 yuan, and the high - end version is as high as over 60,000 yuan.
The price is positively correlated with the product's functions. Dimensions such as language ability, the ability to perceive human emotional fluctuations, reaction speed, interaction ability, and whether it is an IP determine the price of a product.
The gross margin of AI toys is much higher than that of traditional toys and is comparable to trendy toys. After toys are AI - enabled, their prices are 5 - 10 times higher than traditional plush toys. According to data from Qianzhan Industry Research Institute, the gross margin of general AI toy products such as the basic model of BubblePal is about 50% - 65%. For products like the Tom Cat robot and RoPet priced between 1,500 - 3,000 yuan, the gross margin is maintained at 70% - 85%. For some high - end products such as LOVOT and Tesla Bot, the gross margin can even exceed 90%, higher than the 70.3% gross margin of trendy toy leader Pop Mart.
The imaginative profit margin is attracting various players to enter the market. The reduction in the cost of large - model technology has lowered the threshold for AI toy startups. Large companies, traditional toy manufacturers, IP companies, and AI toy startups are all targeting the AI toy sector.
However, there are probably far more companies in this industry good at raising funds than those that can finally mass - produce products.
Qin Shuguang, the co - founder of the technology of Youni Robot, an AI toy product, was a programmer for more than a decade. As the industry's dividend gradually disappeared, he was forced to transform. After a market survey, he finally chose to start a business in the less - competitive AI toy sector. He told Tech Planet that they cooperated with a technology product company in Shanghai. The company has invested about 5 million yuan since last year. A team of more than 30 people spent a year on independent research and development before finally launching the first desktop robot product, Youni, priced at 399 yuan, competing with the Eilik robot.
The company's main profit model is B2B, providing contract manufacturing for brands. Shuguang revealed that they only make a few yuan from a product sold for a few hundred yuan.
The pitfalls he encountered in the past year made him realize that many domestic products are not successful because many people only stay at the stage of hyping concepts and raising funds with demos. Many technology companies follow a similar pattern: they raise millions or tens of millions of yuan, make a demo, and then nothing happens. The past era of PPT - based financing has evolved into the era of demo - based financing, which has also led to many domestic processing factories being established for making demos rather than mass production. Many novice entrepreneurs usually have a hard time avoiding this test, and he and his company paid hundreds of thousands of yuan in tuition fees.
Low Success Rate in the Hundred - Billion - Dollar Market, Some Lose Tens of Millions
The hundred - billion - dollar market is waiting for the next Pop Mart to emerge.
According to a research report released by Market Research Future (MRFR) in September 2025, the market size of AI toys is expected to grow from $42.15 billion in 2025 to $224.75 billion in 2034, with a compound annual growth rate (CAGR) of 20.48% during the forecast period (2025 - 2034).
However, capital's attitude towards AI toys is not unified. Some investors have bluntly stated that they are not optimistic about the AI toy sector.
The progress of startup players in the sector is also not optimistic. An AI analyst told Tech Planet that after talking with AI toy entrepreneurs recently, he felt that the development of the sector is not going well. AI toys have a relatively high cost and are not selling as well as ordinary plush toys. The large number of financings may be related to investors' wide - spread investment strategy, covering every sector. The financing amount can also reflect the degree of capital's attention.
The success rate of AI toys is extremely low, and some people have withdrawn from AI toy projects. Lao Shi, an AI hardware entrepreneur, told Tech Planet that there are many AI toy entrepreneurs in a community he is in. He has talked with 40 - 50 of them, and only 2 - 3 are still in business.
He himself gave up the AI toy project at the beginning of this year and switched to developing learning companion hardware. Lao Shi said that they were among the early players in the field, starting in April. Their product was positioned as a children's learning companion. In his perception, the sector started to heat up in August last year, and many toy manufacturers and brand owners came to ask if they could cooperate. At that time, the industry solution was still in a blank state. So, Lao Shi's team provided the technical solution and cooperated with IP companies and toy manufacturers to develop AI companion products.
But soon, Lao Shi realized that the direction was wrong. Lao Shi said that as he came from a product - technology background, he initially often started from his own or his surrounding people's needs without a clear direction. By December last year, he felt something was off. On the one hand, the product retention rate was extremely low. On the other hand, there was a serious inventory backlog, and the market price had dropped significantly. The price that was originally over 100 yuan dropped to 70 - 80 yuan at the beginning of the year and has now dropped to 19.9 yuan, and even 9.9 yuan. There were also technical problems, such as low sensitivity and inconvenient operation.
Lao Shi said that after trial and error, he found that the industry hopes to empower toys with AI large models. However, AI large models are rational and have no emotions. All large models learn human language and text, which is a logical thing and cannot provide emotional companionship. What it needs to do is to provide companionship in specific scenarios and solve people's rigid needs.
Secondly, even if AI toys can be successful in the future, it won't be the technology solution companies. The final successful ones must be IP owners like Disney. He has cooperated with IP owners and knows how large their traffic is. In his view, the core competitiveness of AI toys lies not in the technology itself but in content creation and continuous influence on user mentality. "If these IP companies enter the market themselves, technology - based companies may be the first to be eliminated." Technology - based companies cooperating with IPs may be doing all the trial - and - error and exploration work for the IP owners in the early stage.
Just because a company has received financing doesn't mean it has succeeded, and good sales don't mean it can rest easy. In Lao Shi's view, AI toys are still in the exploration stage, with a high return rate and a low retention rate. Moreover, the category positioning is wrong. The current excessive investment in technology is like "decorating a turd." Once the price goes up after positioning as a toy, no one will buy it. If the price goes down, the company will suffer losses. Among the people he knows, some have lost tens of millions of yuan.
This is the case for any emerging trend. A few unicorns will emerge, and some will lose money and withdraw. With the boost of capital, there is a possibility of success, but there is also a risk of failure.
This article is from the WeChat official account "Tech Planet" (ID: tech618), author: Zhai Yuanyuan. It is published by 36Kr with authorization.