Dreame: the vacuum company that is now selling cars
the Xiaomi supplier turned global cleaning-robot brand to now selling AI cars
Hey everyone, I spent the afternoon at the annual journalism conference at the Foreign Correspondents’ Club in Hong Kong. During the newsletters panel, the speakers shared how they got started and what hacks they did to jump start their initial following base.
Listening to them, I realized how incredibly lucky I’ve been. AI Proem has grown almost entirely organically, simply by harassing 2-3 friends in the beginning and things took off on its own on Substack. Now, with over 10k followers online and having met so many of you offline, it feels a little surreal. As you know, nothing is behind a paywall but I do provide paid private sharings sessions. Over the course, I’ve simply written about the things I genuinely care about, that excites me, and only when I felt I had something meaningful to say. [For more, see here]
Now, I rarely get writer’s block because there’s always far more I want to cover than time allows. So when I do let my anxiety get the better of me and publish something that feels a bit of a ‘filler piece,’ I apologize.
So today, I wanted to humbly ask for a small favor: If you’ve ever enjoyed any of my writing or interviews, would you please share AI Proem with a friend today? It would mean a lot. Your support is what keeps me going. Thank you.
Everyone’s talking about humanoid robot marathons, and Silicon Valley’s hottest livestream is of a warehouse robot. But there is actually a more interesting physical AI story unfolding before us that is going unnoticed in the West: a Chinese vacuum company that now makes sports cars.
You know which Chinese company is making a lot of noise right now? Dreame. The vacuum company. The home appliance company. The robot vacuum company. The “AI-powered whole-home ecosystem” company. The one that debuted its huge demo event in San Francisco in late April, talking about cars, phones, smart rings, robotic arms, AI laundry, robotic lawn mowers, and a future where every consumer electronic will be reinvented.
How do you even categorize a company like this?
At Dreame NEXT 2026, held at the Palace of Fine Arts in San Francisco, Dreame staged its largest international showcase to date. Officially, it spanned smart mobility, smart home appliances, personal devices, and personal care. Unofficially, it looked like a hardware founder trying to convince America that a Chinese vacuum company could become a full-stack consumer technology platform. Dreame’s own messaging was that “in the age of AI, every product deserves a reinvention.”
The Verge’s coverage was maybe a little less generous: hundreds of influencers, some celebrities, and many products that either had no pricing or launch date, or felt more like concept theater than shipping hardware. There were phones, smart rings, a rocket-powered EV concept, and enough robotic arms to make the whole thing feel like a showroom for a future, but everything was still somewhat in the R&D process and hadn’t passed procurement yet.
Both versions are true to some extent, but I think the company is worth learning about because its pursuits and vision- not in humanoids - seem like a much more realistic immediate next step in the next stage of consumer robotics, if one were to put robots in their homes.
Yu Hao doesn’t do quiet, or lose
In today’s Chinese tech environment, as in after the whole Jack Ma saga, most Chinese tech (all rich people) founders have become careful, sanitized, or gone out of their way to be invisible. Yu Hao (余浩) has been going the opposite direction. He talks a lot, he makes public statements a lot, and he’s very active on social media. All of which is probably a headache for his PR team but somewhat entertaining and refreshing in an environment where everyone’s either hush or reading off corporate scripts. He talks about building one of the greatest companies in human history. He talks about a future Dreame ecosystem worth 百万亿市值. Now what is that? One hundred trillion dollar market cap company, and for context, that is like ~20 Nvidias. Like I said, not a humble guy. His ambition is to grow the Dreame ecosystem to the largest company in the world, essentially.
While most Chinese tech leaders of this generation have retreated from public life, they have let their PR teams handle the messaging if there is even a messaging KPI, and have stayed off social media. Yu Hao is doing the opposite, actively commenting on other companies’ high-profile social media moments, making himself the product before every product is ready. Some say this was a deliberate strategic decision due to the fact that the company used to serve a 2B audience, to pivot to 2C, it needed to recreate itself or bolster its street creds. And what gets more attention than an unhinged, unfiltered boss who embodies everything he has built. Because in a consumer hardware market where features get copied fast, and product cycles compress, attention itself becomes a moat, or at least a financing tool, a recruiting tool, and a channel tool.
Now there is also ample authenticity to this. His personality, boisterous, courageous, and arrogant, seems to have come with his success from an early age. In many ways, he even admits, he’s never not won.
Yu Hao is one of those founders whose life story reads almost too cleanly. He was 保送, which means guaranteed through every stage, meaning he was recommended to the next level without taking the grueling nation-wide/provincial-led middle school-to-high school exam zhongkao, nor the high school to university exam gaokao, or graduate entrance exams. He entered Tsinghua through the physics competition results. He studied computational fluid mechanics and launched 天空工厂 (Sky Workshop), a student initiative on campus that hand-selected its members. The student aerospace project even attracted attention and investment from Boeing. And guess what? Upon graduating from Tsinghua, he founded Dreame, and his core team was mostly from that group as well. He famously has said that once becoming number one becomes a habit, you get addicted to it. That line later became the spirit of Dreame’s internal slogan: either don’t do it, or do it to be number one. And in most categories, they have achieved that.
His ambitions and arrogance know no ceilings. Talking to the media, he once said, paraphrased: “ Every bluff I’ve made, I’ve also made it come into reality. Jensen Huang and Elon Musk created 8-10 trillion-dollar businesses, but they’re a generation older than me. I can create a new ceiling.
The Xiaomi launchpad and escape
As we know, he studied aerospace engineering, so when he thought about starting a company, there were two paths. Something about spaceships or something that might be easier to commercialize.
Looking at how it all started, Dreame did not begin with the idea of making a better vacuum. It began with flight, airflow, and motors. Yu Hao has said that when drones were hot, he decided that drones were too niche. He wanted a product that could enter millions of homes, had a real technical barrier, and had technology that could transfer into other fields. Dyson had already proven that high-speed motors could define premium home appliances. So Dreame went directly to the hard part, spending two focused years on motor technology to match Dyson’s capabilities.
And with the team’s commitment, it found an opportunity to join the Xiaomi ecosystem at the end of 2017. At that time, 2B manufacturers were largely content with making white-label products for distributors and brands. And Xiaomi’s home appliances, ranging from kettles to vacuums, largely depended on these firms.
Little did Lei Jun expect that Xiaomi would give Dreame its platform and a launch pad for its later sprawling business categories. Of course, protective measures were put in place, but Xiaomi thought that it was hard for any new players to come in and challenge its channel access, low price, premium quality, and, especially, for a young company to replicate.
Within the Xiaomi ecosystem, most companies remained in a single category. Dreame didn’t, of course, Yu Hao didn’t let that happen. It moved from cordless vacuums to robot vacuums, wet-dry floor cleaners, hair dryers, and more, often in parallel. Yu Hao wanted to expand across the supply chain. He pushed into robotics vacuums and hair dryers, which I’ve heard actually led to some friction within the ecosystem. Peers were shocked that it was breaking the unspoken gentlemen’s agreement, which is to stay in your lane and be fed comfortably by Xiaomi.
But now we know, it was all a conscious push on Yu Hao’s end; he was slowly pushing the envelope.
Yu Hao’s answer was that Dreame had built two layers: a bottom layer of long-term technologies (motors, vision algorithms) and an upper layer that combined those with external suppliers and channels. New products didn’t need to start from zero. The company could reuse the stack.
The escape strategy was clever. By 2020, Dreame’s own-brand products already accounted for about half of the company’s sales, while overseas sales were roughly 70% according to 36Kr’s reporting. Yu Hao explained the overseas-first move simply: they didn’t have enough money. Domestic promotion was expensive. Overseas markets were more open, and Chinese companies like Huawei and Xiaomi had already spent years educating the market abroad. His shorthand was: use overseas to hit domestic, use high-end to hit low-end, use new channels to hit old channels.
And here’s a detail from their early brand-building that I think is underappreciated. When Douyin (TikTok’s Chinese version) opened up to e-commerce at the end of the 2010s (2018ish), there were basically no legitimate premium brands selling on the platform. Dreame went all in on Douyin commerce early, making it a core channel for 2C marketing when everyone else was still skeptical. For a resource-constrained company competing with Xiaomi’s own distribution, that was a genuinely differentiated bet on a channel no one else was taking seriously yet.
Whether intentional or not, the success came about when a few ingredients converged. Dreame used Xiaomi to learn how to make and figure out what sells, used overseas markets to prove it could stand alone, used Douyin to build a brand when no one else was premium on the platform, and then came back to the domestic market with higher-end positioning.
Core tech vs Core customer
There are two types of companies that can expand beyond their original category, usually.
The first type expands through distribution and user ownership. Tencent can push a new service because it owns attention. Meituan adds categories because it owns local commerce behavior. Apple adds services because it owns the device relationship.
The second type expands through the reuse of core technology. Honda and Mercedes make motorcycles and cars, because the underlying technology and advantage is the capability of the engines and the ability to assemble them together. DJI moves from drones into cameras, gimbals, and robotics-adjacent products because stabilization, sensor, imaging, motors, and control systems transfer.
I’d call this Core Tech vs Core Customer. Dreame is betting on Core Tech.
Motor. Airflow. Battery. Sensor. Algorithm. Motion control. Manufacturing. Channel. Yu Hao’s argument is that these are transferable hardware primitives, and if you build them right, entering each new category becomes cheaper because everybody’s favorite word - sYnerGy. This is why he’s described Dreame as ‘没有边界的公司’, a company with no boundaries.
If that sounds grandiose, it is. The team spent the first two years of their initiation just figuring out how to optimize the motors. At the time of its inception, as I mentioned above, Dyson was the holy grail of luxury home appliances, and their goal was simply to match Dyson’s motors. Motors go into vacuums, hair dryers, fans, air purifiers, lawn robots, and maybe cars. AI goes into navigation, cleaning, laundry, and phones. Robotic arms go into vacuums, mowers, laundry machines, and refrigerators.
Now, Yu Hao, the young entrepreneur, but legend has been called a Elon Musk + Lei Jun, funny enough, the two ‘seniors’ seem to have had their encounter recently. He also famously spent 2.282 billion yuan (~340 million) to acquire control of Jiami Packaging, a listed company on the A-share market in 2025, which even fueled speculations that the company would venture into food packaging (lol I guess no borders/ boundaries right). But that claim was quickly denied as “false information” by the company, and since then, IPO rumors have not really resurfaced. Part of the reason it hasn’t been smooth for the company to list publicly is its ambitious (insane) goal.
So how do you put a valuation on a company that does everything? I mean, beyond the vacuums we talked about, it claimed to build an AI native EV to challenge Bugatti Veyron.’
Wheels before legs
So the reason Dreame caught my eye is that amid all the humanoid mania, a robotics company like itself seems to be rejecting it altogether.
Everyone wants to talk about humanoids because, well, we’ve written before 1/ ego 2/ sense of self-importance 3/ buzz and what people say “our physical world is designed for humans’ but you guys know how I feel about it all. It’s simply just not there yet. And I, for one dont understand why industrial six-axis arms don’t get more love.
At CES 2026 in January this year, 21 of 38 humanoid robotic displays were Chinese. The marathons and demos definitely fuel the excitement. The discourse follows Elon Musk’s vision of a general-purpose humanoid, and the rave followed. But Yu Hao understands that hardware does not scale as easily as software. There’s no Moore’s Law for legs, actuators, batteries, reliability, and service networks. In fact, it is said that more than half the cost of a humanoid robot is in creating the legs and the form factor. Using wheels is much cheaper. And that’s why we see Sunday Robotics or Dyna Robotics optimizing with the arms on wheels, adding a ‘head’ like sphere, I guess, for funsies, because a headless bot could turn some people off.
Someone shared a stat with me that I keep thinking about. For every step in a robotics process, assume a 99% success rate. Sounds great. But if you’re running a 30-step process, which is considered long for consumer robotics, your total success rate drops to about 74%. That’s the math that makes hardware hard. The chip isn’t cheap, the software models for robotics aren’t strong yet, and the cost structure is overwhelmingly in manufacturing.
The thing is, if you push through the noise and hype, think about what an actual practical use case of robots in a home environment is. A vacuum that can know not to suck up a sock, and elevate itself over a step, or a human-looking bot that can run fast.
And why Dreame’s consumer specialized robots and not humanoids, have actually built a business, a real, viable business. Some may say Dreame’s bots dont look ‘revolutionary,’ but a functioning bot that can clean the pool or mow the lawn really adds more tangible value than most of what we’re seeing on the market.
We are the results of our times
There’s a saying in Chinese, 时代对个人的塑造, which essentially means “How our times shape who we are.” It is often used to describe how time and place determine a person’s outcome more than their efforts. Some use it to describe their own success as they rode the wave of privatization and the opening up of the economy in China in the 1990s. Others use this phrase to showcase that, at times, even the smartest and most hardworking won’t get their break because of a dire environment.
Dreame’s founding was closely tied to the rise of OEM suppliers to big electronic brands. Dreame’s consumer- branding success may be due to the genius of Yu Hao’s strategic positioning. But now the technology itself can be attributed at least partially to the bigger movement. As the country’s decades-long efforts in building, enhancing, and grooming a generation of companies, technology, and talent in the spaces of manufacturing, lidar sensors, battery storage, and so on have definitely helped. It’s the convergence of technologies that is embodying the boom of China’s AI robotics industry now. So the story of Dreame is not really just about Dreame, it’s the whole ecosystem around the robot.
Just as Rui Ma wrote in TechBuzzChina after two trips across China visiting factories and showrooms, she “came away less convinced that humanoids are close to becoming household helpers, but more convinced that China’s advantage in robotics is not any single robot. It is the closed loop around the robot: capital, factories, data, supply chains, customers, and local governments all pushing in the same direction.”
The market data reflects this. IDC’s 2025 data shows the global cleaning robot market shipped 32.72 million units, up 20.1% YoY. Lawn-mower robotics grew 63.8%. Window-cleaning robots grew 70.4%. Chinese manufacturers are setting pricing benchmarks, commercializing AI navigation, and scaling globally through e-commerce. IDC’s conclusion was blunt: supply-chain depth and rapid iteration have become decisive competitive advantages.
And you can see the category reshuffling in real time. iRobot filed for Chapter 11 in late 2025 and was acquired by Shenzhen PICEA Robotics in January 2026. Reuters reported that iRobot entered bankruptcy with roughly $190 million in debt. The company that popularized consumer robot vacuums ended up being owned by a Chinese manufacturer. Meanwhile, Chinese brands like Roborock, Ecovacs, and Dreame compete in the premium segment.
Two decades ago, German and Nordic brands dominated the premium consumer electronics space, and Japanese and Korean companies were the suppliers of many. A decade ago, consumer electronics used to be made in Japan, then Korea, then China as a low-cost factory. (Btw, even Boston Dynamics is now owned by Koreans)
The old story was the price of export but the story now is iteration speed and truly quality design. Dreame’s own numbers claim overseas revenue at nearly 80% of total sales in 2025, North America revenue up 189% YoY, products in 42 million households across 120 countries, over 10,000 patents filed with 3,000+ granted, and R&D personnel at more than 70% of the company. A Dreame-distributed release citing IDC’s tracker claimed 27.6% unit share of European robot vacuum shipments in 2025.
Even though I’ve written about how some Chinese companies are still positioning themselves as robot OEMs for global brands. That is not Dreame’s endgame. No, for Yu Hao, born in the late 80s, like many other entrepreneurs who grew up in an economically vibrant China (DeepSeek’s Liang Wenfeng), their confidence and ambitions since day one have been in the global market. They’re the 主角s main characters, not NPCs.
*Company exec just pulled out of intv ….
Dreame’s real innovation: leveraging national capital(国资委）while telling bigger market-size stories (cross-category market-entry, global gtm), still lack of innovation in real ideas/products.