Product Watch | ZhuiMi enters the smart mattress market, company valuation reaches 1 billion in 3 months

robot
Abstract generation in progress

Ask AI · Why does the smart mattress segment quickly attract capital?

Author | Qiao Yujie

Editor | Yuan Silai

China exports mattresses worth 10 billion dollars to the world every year, but for a long time, the star companies in the smart mattress industry did not emerge from here.

The industry’s star entrepreneur is an amateur athlete from a small town in Italy, Ferrara. In 2014, he and his wife founded a smart mattress company in New York called Eight Sleep. Today, it sells $500 million per year, with a valuation of $1.5 billion.

Needless to say, Eight Sleep’s contract manufacturers and R&D partners are all in the Pearl River Delta. After seeing Eight Sleep’s surprisingly impressive sales, Chinese companies finally entered this market.

Earlier this year, the smart mattress brand Xirui (stareep) under the Pursnatch ecosystem announced the completion of its second round of financing—nearly 100 million RMB. The post-investment valuation reached 1 billion RMB, which is only about three months after it completed its first angel round.

The rapid follow-up by capital, to some extent, also reflects the industry’s renewed assessment of this form factor of “smart beds.”

Sleep monitoring has long been a big arena. In the March 2025 release of the “2025 China Sleep Health Survey Report” by the China Sleep Research Society, the sleep difficulty rate among people aged 18 and above in China is about 48.5%, meaning more than 300 million people have sleep disorders.

Atuo sells 3 million pillows per year, and retail revenue already accounts for half of the total. As a high-priced, high-ticket product, mattresses look even more tempting.

“When we surveyed consumer acceptance of smart mattresses two years ago, most users were still in an observation phase. In retail settings, the average time for sales to go from product introduction to closing is about two hours. But now, it only takes half an hour,” Cai Yanming, president of Xirui, said in an interview with Hard-Tech.

However, for a very long time, smart beds were almost synonymous with electric beds. The so-called “smartness” mostly stayed at the level of posture adjustment and was often joked about by users as “pseudo-smart.”

Today, a new generation of smart beds is beginning to comprehensively introduce AI technology. By using AI algorithms and continuously sensing the body’s condition, it enables adaptive adjustment, helping improve sleep quality. A series of changes is allowing the “bed” to move from a standardized industrial product toward an intelligent product that actively adapts to human needs.

AI Agents in the Bedroom

The 2025 China Sleep Health Research White Paper shows that, on average, adults in our country sleep only 6.85 hours per night. Insomnia caused by stress, “revenge late nights” triggered by scrolling on phones, and passive late sleep driven by the pace of highly urbanized life all compress people’s sleep duration for various reasons.

In Cai Yanming’s view, sleep time shortening is the big backdrop. Under these real conditions, to help users achieve a better sleep experience, it is necessary to start by improving sleep quality.

There is actually very limited room for traditional mattresses to change in structure and support. In common sleeping positions such as side-sleeping, shoulders and legs are prone to excessive localized pressure. If the sleeping posture is even slightly off, waking up can lead to a sore back and aching waist, numb legs—this is exactly the user pain point that has long been ignored.

Xirui’s solution is to treat the mattress as a continuously operating sensing system. Its adaptive mattress includes an AI chip that can sense the pressure distribution on the body, body contours, and changes in sleeping posture in real time, and transmit the data to multiple independent adjustment units inside the mattress. The system dynamically adjusts key pressure-bearing areas based on the load conditions under different sleeping postures, so that the support state always matches the user’s current body data.

(Image source/Company)

To make sure these adjustments truly “happen during sleep,” rather than interrupt the sleep experience, Xirui has carried out engineering optimizations at the underlying hardware level.

The R&D team drew on the technical experience of Pursnatch in motors, electric drive, and electronic control, and jointly developed an intelligent bed frame adjustment system with Pursnatch engineers. With patented motor and structural design, the bed frame experiences almost no jolting during adjustment, and operating noise is controlled to about 20 decibels—close to the lower limit of sounds that humans can perceive.

Under this system, the mattress and the bed frame no longer act independently. Instead, they collaborate, intervening throughout the entire cycle of the user’s sleep: before sleep, posture adjustment helps the user relax and fall asleep; during sleep, it performs seamless adjustments in response to snoring or changes in load; after sleep, it completes waking up in a more natural way.

And it is precisely on this foundation of being quiet enough and natural enough that intelligent adjustments can truly take on the role of “improving sleep,” rather than evolving into yet another intelligent device that users must actively adapt to.

From One Bed to a 24-Hour “Sleep Business”

Compared with pillows, a large share of mattress consumption still takes place in offline stores.

To help users complete the experience more quickly, Xirui introduced the Matchfit matching system. Users lie down and test for 5 to 6 minutes. The system then matches a more suitable mattress softness-and-firmness plan based on body-type data, pressure distribution, and a basic sleep model.

Supporting this process is a set of AI agent algorithms trained on a million-level sample set. It can serve smart mattresses with adaptive capabilities, and it can also provide traditional mattress users with more deterministic selection recommendations.

Once in the usage stage, the mattress continuously records the user’s sleep performance each night, generates sleep reports and scores, and keeps learning from the user’s long-term data—so that the adaptive adjustment plan becomes different for every person as time goes on.

At present, Xirui is extending its monitoring function to hardware such as smart rings, smart ambient lights, smart pillows, and smart temperature-controlled bed toppers. This means daytime body data will be linked with nighttime mattress data. The daytime movement data monitored by the smart ring will also be learned by the mattress, which then responds at night.

Xirui’s rapid valuation growth is not a one-off case. Earlier this year, Eight Sleep just secured a round of financing of $50 million. Jintian Yixiu, founded by Wang Teng, a former Xiaomi executive, completed financing of tens of millions of RMB within just a few days of its establishment.

Traditional home furnishing brands are also smart about choosing to collaborate with software companies. Over the past year, Sleepy Dormitory partnered with BrainCo Technology to launch a brain-computer interface AI mattress; Qisheng Technology released a Shuofude smart bed product to further deepen synergy with the HarmonyOS ecosystem; and Qiusuo Technology, which works closely with Xiaomi’s ecosystem chain, jointly released a professional sleep analysis service with DeepSeek.

If you trace it back, smart mattresses have never been a new concept. When the IoT concept emerged in 2013, this category had already ridden a wave once. But it was soon pushed into the corner due to poor product experience and high prices.

This time, although major brands have rolled out “brain-computer interfaces” and “adaptive algorithms,” the question of how to prove the intervention effect on sleep disorders from medical and clinical perspectives—not merely expensive psychological hints—is the gap the industry must cross to move from “geek toys” to “mainstream, universally needed demand.”

View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments