Nobody pays attention to consumer hardware anymore.

"Many fund managers are now cautious about hardware categories, and they don't have the same enthusiasm as they did last year."

Last year, consumer hardware was still one of the hottest tracks in the primary market. Investors flocked to Nanshan, Shenzhen, and we once wrote an article titled *The FA Moved Its Office Next to DJI* for this. As the saying goes, fortune shifts, this kind of scenario is rarely seen this year.

"Many senior executives from DJI are starting their own businesses now, and they can't even get financing for their first round. From my personal experience, this track has turned extremely cold, and almost no one is looking at it," Investor J told me. He previously engaged with a smart sleep hardware project whose founder used to be a vice president at a large factory. Due to a gap in his career, the founder missed the last financing window, and it has become much more difficult for him to raise funds now.

"To be honest, I personally quite like that project. It has excellent product aesthetics and brand tonality, and its valuation is not expensive either, as a round of financing can be completed with less than 200 million RMB. The problem is that many fund managers are now cautious about hardware categories, and they don't have the same enthusiasm as they did last year," he said.

It is worth noting that along with the boom of consumer hardware, investments in this track have become increasingly abstract. Since last year, we have successively received information on some seemingly novel projects, such as sewing machines, tofu machines, ice makers, and so on. When we first saw these projects, we also found them quite novel and shared them with several investors. However, even post-2000s investors, who are considered "good at accepting new things," felt somewhat confused when receiving such project information, and most of their feedback was "can't understand it," "won't invest in it," and "how to exit?"

Investor Z, who has long focused on hardware investment, told me that these seemingly niche and gimmicky hardware projects actually raised funds at the peak of the hardware boom last year. Even many FAs would first introduce such projects to institutions like HSG, GL Ventures, and IDG. For these top-tier US dollar funds, most of the time they will not make overly detailed judgments. They bet on hit rates, focusing on not missing good opportunities rather than fearing making wrong investments.

However, the cooling of the boom does not mean that this track has fallen into silence.

Siming has long been deeply involved in hardware investment, and not long ago he started his own business and founded an AI Native-focused FA institution — Puhua Capital, which focuses on AI hardware and AI application globalization. His understanding is that AI hardware is not "ignored" as the outside world imagines, and investment institutions are still paying continuous attention to it. Compared with last year, the market sentiment has obviously calmed down a lot. "Now everyone prefers to wait and see first whether this batch of companies that have received financing can finally succeed."

This is understandable, as no track can remain in a state of frantic financing forever. After the phased completion of fundraising and investment, it is time to wait for these companies to grow up and deliver their performance reports.

"The Next Bambu Lab" May No Longer Be a Viable Pitch

First, let's take a look at the recent cases in this track that have raised financing of over 100 million yuan.

Not long ago, a company named XBOT announced that it had successively completed two rounds of financing totaling several hundred million yuan. Among them, the 200 million yuan Series A round was invested by Hong Kong Jiankun Capital GPTX, and the Series B round was about 300 million to 500 million yuan in scale, jointly participated by multiple government funds, US dollar funds and industrial investors. Founded in 2022, the company is positioned as a general-purpose catering embodied robot, with its core product being the AI coffee robot, which integrates multi-modal interaction systems, visual perception and large model technologies, and is priced at about 219,000 yuan per unit.

Also in the coffee machine track, another commercial fully automatic coffee machine brand, Ka Ye Technology, has also completed nearly 400 million yuan Series B financing, setting the single-transaction financing record in this track, which was led by Meituan Longzhu and followed by institutions including GL Ventures.

At the Hicool event last year, I saw XBOT's product, and I chatted with a friend from Brazil who runs a coffee trading business at that time. Judging purely from the flavor, it may be hard to tell the difference between an Americano or a latte made by the robot and those made by human baristas, but if you order a latte and upload a photo, you can get a latte with the exact same pattern on the foam, and the emotional value at that moment is indeed fully maximized.

Interestingly, the founder of this company, Tang Mu, once served as the general manager of Tencent User Research and Experience Design Center and the vice president of Xiaomi's Ecosystem Chain. Many of the classic QQ avatars that have been popular for generations were created by him and his team; he also participated in the development of consumer hardware hits such as Xiaomi routers and Xiao Ai speakers. At present, its coffee robots have been deployed in more than 100 cities around the world with over 1,000 units in total, having made over 4 million cups of coffee cumulatively; its revenue in 2025 exceeded 100 million yuan, and the order amount in hand in 2026 is already close to 300 million to 500 million yuan.

Whether for capital or consumers, the emotional value brought by DIY does constitute a reason for people to pay for related products. The most typical example of similar product logic is consumer-grade 3D printing.

"Essentially, what people buy is still a consumer product, but they are paying for emotional value," Siming mentioned that he is currently preparing to serve a company that makes smart sewing machines, which is a typical project with emotional value. In his view, when combined with AI, such products can achieve high customization, attracting more and more young people who love fashion and DIY to participate, just like Bambu Lab, which opened a new era of DIY.

Beneath the surface, Clawlab, a consumer-grade smart textile company, recently announced that it has successively completed several rounds of financing totaling over 100 million yuan, with investors including HSG, Shunwei Capital, Yuanjing Capital and miHoYo, etc. Founded in December 2022, Clawlab's founder Hu Wenxin has an engineering background and once worked at top-tier large companies such as DJI and Meituan. He said that the textile industry, as a long-standing sector, naturally ranks first among the basic needs of "clothing, food, housing and transportation", which is a sufficiently large and real demand. Customized DIY is the next step of people's inward exploration and outward expression after basic needs are met. Over the past three years, he has mainly focused on three things: targeting the direction with the highest ceiling for consumer-grade textile DIY to overcome difficulties along the way; accumulating sufficient resources to prepare for the subsequent big explosion of related categories; and building a sufficiently open and extreme atmosphere to provide a matching environment for more talents to join in the future.

It sounds quite impressive. Of course, not everyone approves of such logic. Investor Z said that hardware products divided by hobbies, such as fretless guitars, have relatively low technical barriers, so latecomers can easily enter the market, and price wars will be extremely fierce. Unless the company has strong capabilities in ecosystem construction and rapid product expansion from the very beginning, it will easily fall into vicious competition, leading to a rapid decline in growth rate and gross profit margin.

For another example, 3D printing is indeed a good direction, but at present, Creality's market value is only 9 billion Hong Kong dollars, with its revenue reaching 3 billion Hong Kong dollars, so its PS ratio is already very low. "Think about it, with 3 billion yuan in revenue, the corresponding market value in the secondary market is only 9 billion Hong Kong dollars. Most smart hardware companies cannot even achieve 3 billion yuan in revenue, which is not an easy task."

From Z's perspective, in the first round of financing, since the team itself is not particularly outstanding, the first-round valuation is usually not set too high. They usually solve the core pain points of the market through crowdfunding and overseas sales, and their products sell well, so US dollar funds or RMB funds will invest from the PE perspective, for example, 1 yuan of revenue corresponds to 1 yuan of valuation.

The impact of the secondary market is the most direct. When we wrote about the boom of the AI hardware track last year, the market even said that US dollar funds had gone all in on consumer hardware. For one thing, backed by China's strong supply chain, hardware-related sectors are easily the preferred direction for entrepreneurs. For another, the listing of Insta360 undoubtedly further boosted investors' enthusiasm for focusing on hardware.

But everyone knows the subsequent performance: Insta360's market value plummeted from 150 billion yuan to about 50 billion yuan now. After Creality, the "first share of consumer-grade 3D printers", went public, its market value has fallen by almost half, approaching the issue price all the way. "Last year, 1 yuan of revenue could correspond to 10 yuan of valuation, but this year the situation has completely changed. The valuation has dropped a lot, and many companies may only have a PS ratio of 2 to 3," Z told me.

J believes that in the consumer hardware sector, if a company cannot take the leading position in its track, first, its upside potential is very limited; second, even if you become the most powerful player in the track, you may not choose to go public. Companies with excellent cash flow such as Bambu Lab and DJI do not have clear listing plans.

"Everyone has started to return to rationality, because facts have proved that the old logic does not work. If you push the valuation too high, and the final performance fails to meet expectations, you have to pay the price yourself."

Insufficient AI Capabilities, Make It Up with Embodied Intelligence?

Another trend is that more and more consumer hardware companies are actively embracing the concept of "embodied intelligence".

Take the AI coffee robot mentioned earlier as an example, its story is far more than just a coffee machine that can draw latte art, but is further positioned as a general-purpose household humanoid robot. According to the company's description, this intelligent system does not only execute preset programs, but has a "generalizable task brain" that can understand, transfer and make independent decisions. Through continuous deployment in thousands of scenarios across more than 100 cities around the world, the system continuously acquires data and iterates models, finally forming a data-driven evolution closed loop, enabling the product to have the ability of continuous growth.

For example, the coffee machine we mentioned earlier, the team is not satisfied with just making latte art, but positions the product as a general-purpose household humanoid robot. In the founder's vision, the intelligent system not only executes preset programs, but also has a "generalizable" task brain that can understand, transfer and make independent decisions. Through large-scale deployment in thousands of scenarios across hundreds of cities around the world, a unique data-driven evolution closed loop is built, giving the product the vitality of continuous growth.

Siming's view is that many entrepreneurial teams are telling such stories, but he thinks there is a bit of a false proposition here. Because startups do not have sufficient shipment volume in the early stage, it is very difficult to accumulate data, which is a typical "chicken or egg first" problem. Many people say that data can form a closed loop, but investors will ask in return: where does your data come from?

However, he also believes that the value of data itself is not problematic, and the key lies in whether the data is sufficiently scarce. For example, a team that makes pet hardware may not only sell hardware in its future business model, but also provide long-term accumulated pet behavior data to robot companies or embodied intelligence enterprises. "I think this logic is valid, because no one else can get such data."

For household scenarios, the real problem is whether the data has sufficient generalization capabilities. Z mentioned that household embodied intelligence is essentially consumer electronics, except that its service objects are families. Automated and intelligent equipment in the home such as sweeping robots and floor scrubbers, and there may be robotic arms that do housework in the future, can of course complete tasks in fixed scenarios and fixed processes, but the data collected is difficult to generalize to the entire household environment, let alone support the real general intelligence. At present, there is still a considerable gap between the value it creates and the price of the product.

Compared with the concept of "embodied intelligence" itself, Siming pays more attention to who can truly define a new product category. Siming believes that AI hardware itself is a very good concept. If a company has the opportunity to invent a new product category and take the lead in defining it, its valuation space will be several times higher than that of copycats or followers.

Another type of investor has the idea of first making an extremely niche small category the No.1 in the world, and then gradually expanding to adjacent markets. At that stage, consumers are not just buying the product, but the brand itself. After the brand mindshare is established, it will be much easier to extend to new categories. For example, Insta360 took exactly this path.

Specifically, Z told me that he is currently focusing on two directions: one is large hardware plus large software, such as complex products like DIY hardware, CNC, 3D printing, because these are large categories with technical barriers, which are his key focus areas. The other direction is small hardware plus large software, or small hardware combined with more underlying applications. He believes that such products involve not only upper-layer algorithms, but also underlying materials, devices, and then upper-layer algorithms, forming a complete technical barrier. With the development of AI, such as Codex and various Coding Agents, the threshold for software development has been greatly reduced. Software development, which was once considered to have high barriers, is changing. Therefore, for small hardware or other equipment, he believes that it is more necessary to have underlying technical barriers, such as establishing barriers from the material level, so as to have the opportunity to achieve good development in the consumer electronics sector.

J's view is more straightforward. He believes that using hardware to build an ecosystem may not be the most essential approach, and starting from software may be more essential. Because software can not only expand upward to make hardware, but also expand downward to build underlying infrastructure (infra). "If you only have hardware without model capabilities, or many capabilities cannot be continuously scaled, the value will also be greatly limited."

In the final analysis, investors ultimately care about exit paths. For consumer hardware, IPO may still be a hard-to-reach end point, and a more ideal outcome is to become a company with excellent cash flow.

Siming mentioned that at present, neither the ChiNext Market nor the Sci-Tech Innovation Board in China has truly opened up for consumer hardware companies. The Hong Kong Stock Exchange has opportunities, but the cumulative profit of over 100 million yuan in three years is a threshold. However, he also emphasized that those companies that can truly define new product categories, build brand mindshare, and have continuous horizontal expansion capabilities will still be sought after by capital. "Such companies will not find it difficult to raise financing, and most of them will not easily go bankrupt," he said.

This article is from WeChat Official Account "China Venture", Author: Wei Xianghui, Editor: Liu Yanqiu, Published with authorization from 36Kr.