In November 2025, Rabbit, once regarded as the “most dazzling new species” in the AI era, finally collapsed due to a broken capital chain.
The Rabbit R1 was once regarded as the “iPhone” of the AI era, setting a miracle of selling 100,000 units in just four days. However, this popular product was prematurely packaged as an “entry-level product,” raising user expectations. Later, as the product failed to meet the expectations in terms of performance, the return rate soared, and its reputation declined significantly. Eventually, this team of only 26 people fell into a dilemma of wage arrears and exhausted cash flow.
“There were too many negative feedbacks on this product,” said Silvia, who has years of experience in hardware investment. “The more units were sold, the more negative feedbacks were received from users, and the high return rate ultimately dragged down the company’s cash flow.”
Among all the people we interviewed, Silvia was the one most curious about the overall development of the hardware industry. In the past few years, she has been actively involved in the investment and startup circles and has since been directly involved in the R & D process of specific hardware products. This was driven by a frustrating reality: current AI hardware products simply cannot meet users’ expectations.
Rabbit’s experience is not an isolated case. We once visited a store selling AI hardware on a digital culture street in Hangzhou and tried out several products priced between 399 and 1499 yuan, which were mainly focused on interaction and English learning. They felt more like toys and were basically unable to achieve real – time voice interaction. Even those with visual functions were simply mechanical and far inferior to software assistants like Doubao, Qianwen, Yuanbao, and KiMi on mobile phones.
In the mainstream narrative where single products become popular overnight and capital is flocking to AI hardware, Rabbit is undoubtedly the most extreme example. Its implication is clear: It’s not that the path of AI hardware lacks imagination; rather, the AI hardware at this moment is far from reaching the “iPhone moment.” Most AI hardware products are just transitional and are destined to be phased out by the times.
So, in this track with the most certain opportunities, where are the opportunities for entrepreneurs to enter the game?
01
AI for the Sake of AI
As AI companion toys gradually move from concepts to the public’s lives, the real – world experience often comes with a sense of disappointment.
Last week, Jianzi spent 1,000 yuan on an AI picture – book robot online. After the experience, he was disappointed but also relieved. He is a veteran in the hardware industry. After leaving Hema, he started his own business and is currently developing an English enlightenment product for children called Baobao Long. In the past few months, he has bought no less than seven AI educational robots, but none of them satisfied him.
Some robots have a mechanical voice when reading picture books and lack the ability to interact randomly. Others have inaccurate voice recognition, with a response time of up to three to four seconds. “People, especially children, can only tolerate a delay of one second. Currently, large – model image processing or cloud – based large – model voice response time can basically be kept within 2.5 seconds,” said Jianzi. A response time of three to four seconds is already unacceptable.
Based on his market research, most products only use the concept of AI without actually solving real – world problems. The practical reason is that the computing power cost is too high, and edge – side chips cannot be widely used in consumer – level hardware products. As a result, cloud – based large models have to be used, and there are many inaccuracies in the audio – text conversion of cloud – based model solutions.
In the specific scenario of reading picture books, the AI picture – book robot he bought had relatively timely voice interaction, but the AI voice lacked warmth, making it difficult to create a pleasant learning atmosphere for children, and it could not be interrupted midway. Otherwise, the token computing power consumption would be huge, and the business model would not be viable.
“This kind of AI experience is very strange. It’s more like doing AI for the sake of AI,” said Jianzi. He wants to focus on in – depth product development to truly solve users’ problems.
Silvia shares the same view. Due to limitations in software and hardware technology, most AI hardware products fail to meet users’ expectations. Even the popular AI products on the market mostly rely on other elements.
The most representative case is fuzozo, which was launched in June 2025. It pioneered the paradigm of hanging – style AI hardware and was called the “AI version of Labubu.” It sold thousands of units within a few days of its launch and a cumulative total of 200,000 units in half a year.
Silvia believes that people are more attracted to the trendy toy elements. Some users buy AI hardware on a whim, and these products will ultimately end up gathering dust. Even Meta’s AI glasses, which sold 2 million units, are more like high – end sunglasses with AI photography and voice – interaction capabilities. The core user experience is closer to the concept of “sunglasses +.”
However, these current issues do not deter entrepreneurs and investors from their long – term goals. The prospect of betting on the next “iPhone” in the midst of bubbles and chaos is still very appealing.
02
Betting on the Next New Species
In October 2024, Silvia attended an exchange meeting with less than 100 participants in Silicon Valley. The meeting was filled with engineers from tech giants such as Apple, Meta, Google, and Microsoft, as well as front – line investors and entrepreneurs like Kai – fu Lee.
The meeting’s theme was quite forward – looking. At that time, the concept of intelligent agents was not yet popular, but a group of engineers had already predicted what the new hardware entry point would be and that there would be a “battle of a hundred glasses” in China in 2025.
This closed – door meeting strengthened her belief in the AI hardware track: AI technology has created a new interaction paradigm that urgently needs a new “body” to carry it. Traditional mobile phones and computers are not suitable carriers, and this interaction paradigm will surely bring about a major transformation in the hardware industry.
In this once – in – a – century transformation, the opportunities for hardware are destined to be in China. There are practical reasons for this. China has a mature supply chain and a large number of experienced engineers.
One statistic that left a deep impression on Silvia was that the engineer population in the United States is aging seriously. In 2024, the average age of engineers was 43 years old, and engineers under 30 years old accounted for only one – tenth of the total. Most of these engineers are concentrated in large companies, and there are few in startups.
“This is the best opportunity for Chinese entrepreneurs,” said Silvia. Since the exploration of novel AI hardware is still in its early stages, large companies usually do not allocate sufficient energy and resources to set up dedicated departments for it. They focus more on large – scale tracks.
The investment boom in AI hardware soon erupted in China, and the valuations of some companies changed dramatically. At the end of 2024, an AI glasses company held a product launch event, and its products sold well at the event. Some investors who did not invest in the company before the launch would probably regret it because the company’s valuation increased by 240% from 500 million yuan within two hours.
At the beginning of 2024, the valuation of an AR glasses company that only had a prototype and no specific functions launched was only 200 million yuan. One year later, its valuation increased tenfold.
2025 also became a significant year for AI glasses. In addition to startups in the AR glasses field such as Xreal, Rokid, and Leiniao launching their AI glasses, tech giants like Alibaba, Baidu, Huawei, Xiaomi, and ByteDance also followed suit. Similar stories have spread from wearable devices to more niche tracks, such as companion robots, AI toys, AI recording cards, AI headphones, AI rings, and AI pet monitors.
According to data released by 36Kr, in May 2025 alone, the funds flowing into AI hardware accounted for more than half of all investment and financing activities.
Since the second half of 2025, the enthusiasm of capital for AI hardware investment has remained undiminished, but there is a more obvious preference for star entrepreneurs at the top, such as high – level executives from large companies. In June 2025, after the cumulative sales of Robopoet’s AI companion hardware exceeded 120,000 units, it quickly completed an angel – round financing of tens of millions of yuan.
In January, Looki, an AI multi – modal wearable device startup with a luxurious team, received over 100 million yuan in Series A financing in less than a month after selling more than 10,000 units. After Wang Teng, a former senior executive of Xiaomi, announced his startup, he quickly received tens of millions of yuan in seed – round financing from Hillhouse Capital, Zhipu Robotics, Yunjiu Capital, etc., even without a prototype.
Not only entrepreneurs with backgrounds from large companies but also those with high – level academic backgrounds are very popular. Entrepreneur Feichai currently has three AI hardware projects. In December last year, the new project he joined had only four people. The founder is a computer doctor from Tsinghua University, and another is a supply – chain management director from a large company. Even when the team only had a vague direction, there were more than a hundred investors contacting them.
Entrepreneurs without a glamorous background also have ways to raise funds quickly. A post – 2000 entrepreneur from Shenzhen told us that after they received positive feedback at the CES, dozens of investment institutions showed interest. The initiative has completely shifted to the entrepreneurs.
03
AI Entrepreneurship Is Losing the Standard Answer
Although the valuations of companies in the AI hardware track are much higher than those in other tracks, there is currently no standard answer as to what form future AI hardware should take.
“Traditional hardware can mostly solve specific problems in specific scenarios, such as lawn – mowing robots, coffee machines, and bean grinders,” said Silvia. “They can be compared in terms of parameters to see who can solve problems most effectively.”
However, the problem – solving ability of current AI hardware may be overshadowed by the evolution of large – model capabilities. Investors simply cannot make deductions through pure logic.
In the past decade, a quantifiable standard has been abstracted for the commercialization path of traditional hardware. After having a product prototype, it is verified for PMF (Product – Market Fit, where the market quickly absorbs the product without active promotion) on global crowdfunding platforms like Kickstarter and Indiegogo, then raises funds, and finally goes into mass production.
This logic soon became invalid. The AI recording pen Plaud, which was a global sensation last year, actually raised a relatively small amount of funds on the crowdfunding website in its first – generation product. However, this did not prevent it from achieving global sales of over 1 million units in the following two years.
A partner of a VC firm with years of experience in the hardware industry said that Plaud’s success story of becoming popular without external financing almost “shocked the investment circle.”
Plaud is not the standard answer, and the story of becoming popular overnight does not mean the end. In addition to Rabbit mentioned above, AI Pin also performed well on Kickstarter initially but has now stopped selling due to high false – touch rates in gesture recognition, high learning costs, and a voice – interaction delay of 5 – 10 seconds.
Many investors are in a dilemma. On the one hand, they believe that the AI era has broken traditional investment logic, and there is no basis to follow. On the other hand, when company valuations are soaring, they unconsciously use traditional experience, such as sales data, for evaluation.
Facts have proven that short – term sales do not necessarily mean that a product has the potential to survive in the long term.
Looking back at Plaud’s success, it simply found a differentiated scenario in the recording niche market. Different from the recording pens of large companies, it mainly serves users who have more conversations rather than those driven by documents. This positioning determines the product – definition logic, which is to help users who mainly engage in conversations, such as doctors, lawyers, salespeople, and investors, explore the boundaries of intelligence in unstructured daily conversations.
In addition to finding the right scenario, products need to be iterated more quickly. “The difference in the final results of the application layer may only be 20%,” said Silvia. “For users, such a difference is difficult to translate into continuous incremental value. What really makes a difference is whether the founding team can respond quickly and continuously iterate in a highly uncertain technological cycle.”
Li Yunzhou, the CEO of Xuanyuan Technology, is a veteran in the robotics industry. He started with fighting robots and then switched to companion robots. Different from most entrepreneurs, Li Yunzhou decided from the very beginning that Xuanyuan Technology would not rely on external suppliers and would maintain self – control in software development, hardware development, and even front – end development, emphasizing engineering capabilities.
This ensures that Xuanyuan Technology can iterate its products faster than other companies. Li Yunzhou once said in an interview, “Since our technical architecture is complete and the content library is becoming richer, content can be updated in as fast as one day, and hardware R & D can be completed in 1 – 3 months.”
For most AI hardware entrepreneurs, the risks are still direct and harsh. Due to the uncertainty of technology, product R & D is ahead of its time. Entrepreneurs’ anxiety lies in how to develop a product that suits the market in the next year based on the anticipation of future technology. Feichai said that you also have to hope that no other entrepreneurs will launch a similar product before yours.
An entrepreneur described the process from the product demo to the crowdfunding prototype and then to the mass – production prototype as “a nerve – wracking journey.”
Pan Zhidong, the head of the Global AI Hardware Business Group at Dreame, told us that even if entrepreneurs find the right scenario, they still need to constantly balance the adaptation between hardware and software. ‘Taking Dreame’s AI smart ring as an example, integrating the vibration motor into a tiny volume and making the vibration frequency match users’ physical sensations requires countless rounds of debugging and optimization.’
There are still unicorns emerging in niche markets. In addition to Plaud, Oura, a Finnish company representing AI rings, has sold 3 million smart rings in the past year, and its valuation has soared to $11 billion.
The significance of these unicorns is that although no one can be completely sure that the direction they are betting on is the correct answer for the next – generation products, entrepreneurs still have great opportunities. Before becoming a world – class company, entrepreneurs need to take advantage of China’s supply – chain advantage to get their products off the ground and then try their best not to fall behind.
This article is from the WeChat official account “Baijing Lab,” author: Liu Jia, editor: Baichi. Republished by 36Kr with permission.
link