Each March, the Appliance & Electronics World Expo (AWE) in Shanghai draws significant attention, with thousands of companies exhibiting. This year’s event reflected an appliance and consumer electronics sector caught between rapid product iteration and a broader market reset, including the deeper integration of artificial intelligence and the convergence of virtual systems and embodied intelligence.
Inside Dreame’s 10,000-square-meter exhibition hall, crowds remained steady throughout the day. The company estimated daily foot traffic at more than 100,000 visitors. Its robot vacuum zone drew particular attention. In one live demonstration, a machine moved between table and chair legs while its mop plate swung outward at a wide angle to clean around pillar-like obstacles.
The demonstration centered on two technologies in Dreame’s latest robot vacuum. One was the structural evolution of a dual-jointed arm designed to improve coverage in complex spaces. The other was a jump in cleaning capability, using 160 degrees Celsius steam for deeper cleaning.
Together, those developments move beyond the product priorities that have defined the category in recent years, such as suction power, navigation, and hot-water temperature. They point instead to a more basic question: how else can a robot vacuum better serve users?
Dreame has tried to answer that question before, and the market response has been notable. According to 36Kr, Euromonitor International recognized Dreame on March 12 as the world’s leading brand by high-end robot vacuum sales volume.
36Kr also reported that Dreame ranked first by market share in 30 countries in 2025 and held more than 40% market share in 18 countries and regions, including Sweden, Denmark, and Germany.
Those milestones do more than validate past performance. They also offer insight into how Dreame is positioning itself today.
Rather than following product trends, Dreame appears to be trying to set them. Its ability to navigate around obstacles and sanitize surfaces with high-temperature steam is built on mechanical design and intelligent algorithms. More importantly, it is translating those advances into system-level capabilities that can be replicated and refined over time.
Reinventing cleaning
At the most basic level, users want two things from a robot vacuum: broad cleaning coverage and thorough results. The first is about reaching more of a home. The second is about how well the machine cleans once it gets there.
Over the past three years, suction power in the category has risen from 3,000 pascals to 30,000 pascals. With hot-water mop washing, temperatures have climbed from 60 degrees Celsius to 100 degrees Celsius. Robotic arms have gone from nonexistent to single-arm, then dual-arm setups. Even so, one awkward limitation has remained. In the high-frequency dead zones of daily life, such as the gaps around table and chair legs, many devices still cannot reach effectively. In kitchens and other grease-heavy areas, many still struggle to clean thoroughly in one pass.
It is not that manufacturers have ignored the limits of current solutions. Many have continued along existing development paths. Fewer have tried to break with the underlying structure and rebuild the product from scratch.
To deal with blind spots around table and chair legs, flowerpot bases, and the undersides of floating cabinets, the standard industry approach has been to rely on algorithms that repeatedly twist the machine along edges, or to enlarge the side brush to expand the cleaning radius. Both approaches try to widen the working range without changing the basic physical constraint that the machine still cannot fully reach those spaces.
The drawbacks are clear. Repeated twisting reduces efficiency and can sharply increase the time needed to clean the same area. A larger side brush extends reach only in a linear way. When the machine encounters obstacles such as table legs that require wraparound cleaning, it can still leave a ring of untouched floor.
Dreame’s answer is a dual-jointed swing arm that adds a second joint to the original single-jointed design. The first arm handles outward extension, while the second rotates around the first, much like a human forearm. The angle between the two expands from 50 degrees to 140 degrees, and the mop plate’s extension distance reaches 16 centimeters.
The value of those numbers becomes clearer in a real home setting.
“If you only build a single-jointed arm, it just isn’t flexible enough,” Meng Jia, president of Dreame’s robot vacuum division, told 36Kr. In homes crowded with table and chair legs and more complex furniture layouts, extension in only one direction remains limited. The machine may be able to reach the area, but still fail to clean it well. “It’s like the human arm. Without an elbow joint, making it longer only gives you straight-line movement. It still can’t bend,” Meng said.
What the dual-jointed arm adds is not just length, but freedom of movement. When the angle between the two joints can be adjusted dynamically from 50 to 140 degrees, the machine can clean around obstacles such as table legs in a wraparound motion. When it needs to enter spaces with low clearance such as the area beneath a sofa, the arm can swing outward and fold inward, allowing the mop plate to reach corners that were previously inaccessible.
In that sense, the shift from a straight arm to a curved one gives Dreame’s robot vacuum a more active way to enter tight, complex spaces.
But reaching those areas is only part of the challenge. The machine also has to clean them well.
Users are not satisfied with floors that only appear clean. Less visible problems include old grease on kitchen floors, bacterial risks in areas where pets spend time, and hygiene concerns in households with babies and toddlers. These are heavier-soil scenarios that fall outside routine daily cleaning.
Until now, manufacturers pursuing high-temperature cleaning have largely focused on the temperature of the water used to wash the mop, raising it from 60 degrees Celsius to 100 degrees Celsius. But hot water cleans the mop, not the floor. By the time the mop leaves the base station, its temperature has already dropped. When faced with dried grease, the machine may still need to mop the same area five or six times and still fail to remove the grime completely, leaving the user to finish the job manually.
The problem with that approach is that it can drift away from what users are actually asking for: cleaner floors, not just hotter mop washing.
Dreame said its market research found that more than 80% of users wanted better stain removal in kitchens and dining areas, while households with babies or pets placed greater importance on floor sterilization.
Those needs suggest that the key variable is not only the cleaning tool, but also the temperature of the cleaning medium reaching the floor itself.
In its latest product, Dreame has shifted from heated mopping to a steam-based approach. The company said the robot vacuum can generate 160 degrees Celsius steam in eight seconds and use four spray outlets to target stubborn stains directly on the floor. Combined with what it described as a near-flawless sterilization rate, the feature is meant to support what the company calls barefoot freedom.
On the surface, these look like product upgrades. More broadly, they reflect Dreame’s attempt to rethink the form and function of the robot vacuum itself. While much of the category remains locked in a battle over specifications, Dreame is trying to step outside that competition and focus on the most basic cleaning needs users still have.
Capability built around users
From developing a bionic robotic arm for edge and corner cleaning, to introducing a dual-jointed arm with a 16-centimeter extension, to bringing high-temperature steam into a robot vacuum, Dreame has tried to turn each product iteration into a longer-term advantage rooted in user insight and engineering capability.
The visible part of innovation is easy to spot. It appears in the moment when the robotic arm extends. But for that function to work reliably over time, much of the work happens out of sight.
The dual-jointed arm may look like a simple addition of one motor, but the change created a chain reaction. Module space doubled, squeezing the original air duct layout. As a result, the internal wiring, piping, and structural stacking inside the main unit had to be redesigned.
Durability posed a tougher challenge. The second joint needs to rotate about 90 degrees to extend fully, but under those conditions, the core torsion spring could not meet the required lifespan of 300,000–500,000 cycles. Dreame’s R&D team addressed that by adding a set of reduction gears, cutting the torsion spring’s rotation angle to about 30 degrees and effectively tripling its service life. That gear set, though invisible to users, became central to the arm’s stable movement.
A similar pattern appeared in the development of Dreame’s steam technology. High-temperature steam is common in wet and dry vacuum cleaners, but bringing it into a robot vacuum required more than adapting an existing design.
To fit a steam module inside the compact body of a robot vacuum, Dreame found that conventional modules on the market were too large. The module also sat about one millimeter from the mainboard and battery. Sustained heat at that distance could affect motherboard stability and shorten battery life. Dreame said several early prototypes were burned out by heat dissipation problems.
The company said its R&D team repeatedly scrapped and rebuilt its original plan. After seven structural redesigns, it reduced the steam module to one-third the size of mainstream market products, creating enough room inside the machine without compromising the performance of other components.
That still was not enough. For a household cleaning device, safety is a baseline requirement. Dreame said its team then spent months validating thermal insulation, adding insulating materials, ensuring that user-touchable areas would not exceed 50 degrees Celsius, and installing smart detection units to monitor heat in real time. Only after months of functional testing did the company move the steam technology into its robot vacuum lineup.
That focus on detail points to a broader operating method at Dreame: deep user participation.
From user reviews and surveys to prelaunch beta testing, Dreame said it has developed a proprietary process to ensure feedback runs through the full cycle from R&D to commercialization. It said it conducts more than one million user tests each year.
“Users are actually very clear about their pain points. They just can’t always describe what kind of product they want,” Meng said. “Before cars were invented, people would only have said they wanted a faster horse, because they had never seen a car. Once users express the pain point clearly, the rest is our job. We have to build the product and make it real.”
Guided by that idea, Dreame appears to be trying to do more than solve the problems users explicitly describe. It also wants to identify the points of delight users may not be able to articulate, and position itself as a brand that understands them.
That commitment to user needs often requires harder choices: not chasing trends, not piling on specifications, and not rushing products to market. Instead, it means accepting higher costs and longer development cycles in exchange for stronger products.
Much of the most meaningful innovation is invisible to users. But products that work consistently can still win trust and narrow the confidence gap between a brand and the market.
Toward embodied intelligence
As the industry debates how embodied intelligence might enter the home, Dreame used AWE to present its answer: an all-purpose home service robot.
Built on a four-legged wheel-and-limb structure, the robot is designed for movement across multiple household settings. It can travel through living rooms, bedrooms, balconies, and outdoor areas, and move up and down stairs while balancing stability and efficiency. Its humanoid dual-arm design allows it to perform basic household tasks such as wiping tables and folding clothes. Dreame also presented it as a platform that could eventually handle tools and coordinate with other home appliances. Over time, that could extend to cleaning and sterilization, pet care, and elder care.
The jump from robot vacuum to home service robot may appear large. Within Dreame’s broader strategy, however, the move is relatively easy to trace.
That logic rests in part on the company’s accumulation of real-world household data. Through repeated daily use, robot vacuums gradually build an understanding of home layouts, user habits, and interactions with the floor. That accumulation, grounded in ordinary home life, may also provide a foundation for Dreame’s next attempt at embodied intelligence.
“Having entered households at scale earlier than most, the robot vacuum can rely on massive real-world scenarios for rapid iteration, and in turn develop a deeper understanding of the home as a complex, dynamic environment. That is an innate advantage that other intelligent forms find hard to match,” Meng told 36Kr. Users do not need to initiate extra actions. They simply let the robot vacuum run as part of their normal routines, and it completes environmental mapping, obstacle recognition, and behavioral pattern learning.
At the same time, the technologies Dreame has already built into its robot vacuums, including environmental perception, dynamic obstacle avoidance, object recognition, and voice interaction, are not limited to one device category. They form a base that can be applied across different product types. Those functions, already tested in the market, are now being transferred to Dreame’s all-purpose home service robot.
Laboratory simulations and restricted test environments cannot fully reproduce the complexity of real homes. Valuable environmental understanding and behavioral interaction can only be built through long-term use across millions of households and repeated product iterations. The data generated in those homes is becoming a core capability and a competitive threshold for the next generation of household robots.
For Dreame, the key is not to replicate the precision of industrial robots inside the home. It is to rethink home-service scenarios and interaction logic around the user.
That helps explain the company’s confidence in its future direction. Through repeated product launches and commercial validation in global markets, it has built technological depth and stronger user awareness. Dreame’s all-purpose home service robot may be only the first visible sign of that longer strategy.
From solving a single problem to defining a broader set of household problems worth solving, Dreame’s approach has remained consistent: return to the user’s most basic needs, build durable advantages through system-level innovation, and turn that innovation into reliable products through engineering execution.
KrASIA features translated and adapted content that was originally published by 36Kr. This article was written by Xiao Xi for 36Kr.