A Conversation with Dreame's Yu Hao: I Was Born a Madman

By Lili Yu

Edited by Jing Liu

The robot vacuum war is a blood-soaked arena.

Crouching here are global giants like iRobot, Ecovacs — once worth over 100 billion RMB — and Roborock, whose stock briefly surged past 1,000 RMB per share. Lurking too are aggressive newcomers like Narwal, which burst onto the scene with a single hit product, not to mention traditional appliance titans like Midea and Haier.

But Dreame remains a force that cannot be ignored. Founded in 2017 and landing a record 3.6 billion RMB financing round in 2021 — the largest ever in smart cleaning — the company originally set its sights on Dyson, the vacuum giant. After cracking the core technology of the "high-speed motor," Dreame quickly built up brand recognition, distribution channels, and supply chain resources by joining the Xiaomi ecosystem.

In many ways, Dreame plays like a company that "doesn't follow the rules." Within the Xiaomi ecosystem, most companies try to dominate a single niche. But Dreame does everything — from robot vacuums to cordless vacuums, smart wet-dry vacuums, high-speed hair dryers — and pushes them all forward simultaneously, as if operating with the expansion logic of an internet company in its golden age.

In founder and CEO Yu Hao's roadmap, Dreame's future looks more like a generalized robotics company: taking core technologies — the "brain" (intelligent algorithms) and the "heart" (motors) — and applying them across agriculture, healthcare, transportation, and more, transforming from a single company into a robotics ecosystem.

This relentless expansion, in Yu's view, is simply "habit replication." The man who gained admission to Tsinghua University through his physics Olympiad performance, and who was among China's earliest quadrotor developers and the inventor of a tri-rotor aircraft, has lived a recurring cycle since childhood: from a rural elementary school in Jiangsu, to a town middle school, to a county high school — each time starting from the bottom, but within six months, always finding the trick to becoming number one. "Once someone gets used to being first, it becomes an addiction." Years later, the slogan "Either don't do it, or be number one" now covers every wall of Dreame's offices and factories.

"I've had setbacks, but not failures." Yu makes no effort to hide his swagger. In 2022, while more tech consumer companies chose to lie low, this company — which had just begun pushing into the China market and was still figuring out its own-brand strategy — told Waves that it was putting many established players on its competitor list. "Only by benchmarking against the strongest opponents will you become stronger." The CEO believes this absolutely.

This is a man who still believes "everyone is a genius," who believes "logic can deconstruct anything."

Can he achieve his goals? We cannot say. But what we can confirm is this: in an era when experience no longer works, the order of arrival no longer matters.

Only by benchmarking against the strongest opponents will you become stronger

Waves: Dreame recently released its 618 shopping festival results — 900% year-over-year growth. Are you satisfied with this number?

Yu Hao: We set this year's target at ten times last year's, and we grew tenfold. We actually could have grown twenty times, but we didn't stock enough inventory.

Waves: Why not?

Yu Hao: Supply chain was impacted by COVID in the first half of the year; we were already at maximum capacity. For wet-dry vacuums, we kept opening duplicate molds and ran nine rounds of pre-sales in between. Opening duplicate molds takes at least three months, and there's a ramp-up period too. We never had inventory sitting in warehouses — products went straight out the door once made.

Waves: Has weak consumer spending not affected the smart cleaning sector?

Yu Hao: No, the industry isn't growing — some categories have slight growth, many are declining. Dreame is growing against the trend.

Waves: But in some industry rankings, Dreame isn't that prominent. Meanwhile your office space is covered with slogans like "Either don't do it, or be number one."

Yu Hao: That's our culture. We have clear internal targets: who to catch up to this year, what position to reach next year.

Waves: Sounds like employee motivational slogans.

Yu Hao: You need to benchmark against the most powerful opponents, because a person will increasingly resemble their competitors until they surpass them. Strong enough opponents make you stronger.

Waves: I heard when Dreame first entered the Xiaomi ecosystem, you were dead last, then started climbing back. What was the key?

Yu Hao: First, going deep into core technology. Second, deep supply chain mastery. We started building both capabilities from that point. Some Xiaomi ecosystem companies say Dreame is like an alien, that it doesn't follow the rules, because most Xiaomi ecosystem companies focus on one domain, but Dreame does everything. But this creates massive compounding effects. We built two layers above and below our entire R&D center. The lower layer targets things that won't change in ten years — motors, vision algorithms. The upper layer combines core technologies with external suppliers, channel partners, various technologies. These two layers give R&D teams abundant inputs. When developing new products, you don't start from scratch — you get huge compounding effects.

Waves: After launching two or three products, Roborock split from Xiaomi. Many people think Dreame is following the same path.

Yu Hao: For Xiaomi ecosystem companies to build their own brands is logical. From another angle, pushing the industry forward together is also the mission and vision of the Xiaomi ecosystem. Dreame and Roborock definitely have similarities, but the differences will grow larger over time. In a couple years, people might say we're like Ecovacs or iRobot or even Tesla. Two companies may intersect, but they ultimately head in different directions.

Waves: What about the split itself?

Yu Hao: I've thought about this, but not in terms of splitting — I thought about whether we'd face the same problems with Xiaomi at that time, even when our own-brand sales were still zero. Many companies think at the tactical level, but I believe we need to think from the root, from vision, mission, and values. So on day one of founding, I clarified the company's mission: to achieve technological progress and product accessibility in tech-enabled living.

Waves: But vision can't replace one fact: Lei Jun once set Xiaomi's hardware profit margin cap at 5%.

Yu Hao: Manufacturing for Xiaomi earns one yuan; your own brand earns three. Many people fixate on that two-yuan gap. I don't care that much. If a company wants to do a few hundred million in revenue, there won't be problems with Xiaomi — it's even quite comfortable. If you want to do a few billion and become a qualified public company, that's the biggest challenge. But if you want to do tens of billions, that problem dissolves again. Through May this year, our own-brand domestic revenue grew nearly 300% year-over-year.

Waves: What did you learn most from Xiaomi?

Yu Hao: Honestly, I didn't directly learn any specific methodology from Xiaomi. But Xiaomi is an excellent business school. The best business schools use case studies — you can directly observe what problems a company faces at 20,000 employees. Also, Lei Jun still reviews every product, down to the details. This respect for product is an attitude hardware companies need.

Waves: Back to the 618 numbers — how is Dreame catching stronger opponents?

Yu Hao: If you want the bold claims you've made to look effortless in hindsight, you still need execution. Dreame's path is to strategically benchmark against the strongest opponents, tactically do the long-term things step by step, then find where you have advantages to overtake. Right now Dreame is in a ten-billion-RMB track. Once this battle is won, I'll pick a trillion-RMB competitor.

Waves: A trillion-RMB competitor — like who?

Yu Hao: Hard to say. Companies come in many types. A third-rate company basically does what others have done, then lowers costs. A second-rate company tries many things, has some innovation, but low success rate. A first-rate company basically doesn't make mistakes, can execute on a chosen direction, but hasn't truly innovated in any domain. Super-first-rate companies are Tesla and Apple. What I can do is gradually move Dreame from fourth-rate to super-first-rate.

Waves: For consumer electronics companies, breaking through in a single market is also critical. Is Dreame's position in its main battlefield solid enough?

Yu Hao: Our main battlefield goal this year is to get the own-brand system running smoothly. Our own-brand volume domestically is still small; overseas we haven't fully reached the three-year mark either. But the new battlefield is basically set: building around the generalized robotics ecosystem. Last year we launched the biomimetic quadruped robot Eame One. Based on that quadruped team, we established Dreame Innovation Research Institute, with five cross-domain research directions, gradually building out the Dreame robotics ecosystem.

Waves: Why expand to the larger robotics ecosystem?

Yu Hao: I've been exploring two things: what's happening in the world, and what I can bring to it. Our ultimate goal is to do things that can boost productivity tenfold, a hundredfold. If human productivity doesn't advance, global involution will only intensify. In prehistoric civilizations, people fought over water sources. New agricultural tools resolved that. But if productivity had stalled at that level, it would just be me taking from you, you taking from me — that's all there is. So I think robotics is fantastic; it represents a direction of human light, it can bring social progress.

Waves: That sounds like a very grand strategy. When you decide to explore new businesses, how do you determine whether such a large leap should continue?

Yu Hao: Most of my choices are based on logic. As long as the general direction isn't wrong, many mid-course adjustments are fine. Take our first five research institutes — maybe two won't succeed that well, then we'll do a sixth, a seventh. The core is that these things are worth doing.

In red-ocean competition, whoever makes mistakes first gets eliminated

Waves: We did a rough count — there are at least ten robot vacuum companies with name recognition in the market. Competition in this industry is already very homogenized.

Yu Hao: Homogenization precisely shows we're on the eve of mass adoption. Market demand combined with media amplification can create a bigger tipping point.

This is a process of group evolution. In automotive history, Ford's Model T went through something similar from T1 to T5. This evolution can even happen simultaneously across many companies.

Waves: What is Dreame's differentiation?

Yu Hao: We long-term define ourselves as a generalized robotics company, continuously building advantages in two areas: high-speed motors and AI algorithms. Everything else comes down to comprehensive operational capability, including cost reduction.

Waves: There's been an ongoing debate about vacuuming versus mopping. One common view is that Chinese homes mostly have floors and tiles, so the ceiling for vacuum cleaners is limited.

Yu Hao: Vacuum cleaners are also evolving; you could even say washing machines are an evolution of vacuum systems. After vacuum cleaners came to Asia, brush head forms already changed. Also, Japan and Korea, which have similar living habits to China, have 106% vacuum penetration; Western countries are at 126%. So it probably depends more on whether purchasing power is sufficient.

In the 1960s-70s, when Japan's per capita GDP exceeded $10,000, it experienced explosive vacuum growth. This means crossing a threshold where so-called non-essentials become necessities.

Waves: Last year Dreame raised 3.6 billion RMB. What did you do with it?

Yu Hao: Mainly top-tier R&D talent recruitment and technical team building. We didn't celebrate, didn't do anything superficial. It was mainly strategy-driven, pushing into the China market.

Waves: The Steve Jobs story shows that a product-focused CEO can take a consumer company further, but you emphasize strategy more?

Yu Hao: These aren't contradictory. When a product-focused company establishes its first product advantage, if it doesn't promptly build out supply chain, quality systems, and continuously convert that serial entrepreneurial spirit, the company's vitality is hard to sustain.

Waves: What kind of opponent do you respect?

Yu Hao: Someone who, no matter how high they stand, is willing to get their hands dirty at the front lines. Only this way can you acutely grasp market demand.

Waves: In an increasingly homogenized battlefield, who is more likely to break out?

Yu Hao: In red-ocean competition, it's about who makes mistakes more easily. Phase one, everyone will defeat imported brands in China. Phase two, globally defeat foreign brands. Phase three, it's these domestic brands fighting each other.

Life is short — why become someone who blocks progress?

Waves: Early in its founding, Dreame identified Dyson's motor as the hardest part and decided to spend two years on R&D. Where did this confidence come from?

Yu Hao: First, high-speed motors are quite different from regular motors. Traditional motors stack coils and spin; engineers long cultivated are used to this. But high-speed motors have the spinning coil on the outside, with a high-speed rotor inside. Magnetic circuits, electromagnetic design need algorithmic drive — completely different from before. When experience alone can't solve it, going back to first principles creates breakthrough opportunities.

Also, my hometown is in Nantong, Jiangsu. There's a huge electric tool industry there; brush production is roughly one-third of China's total. I used to wind coils as a kid.

Waves: Have you ever thought: why you, and not someone with years in this field?

Yu Hao: Many technological breakthroughs need the combination of insider and outsider perspectives.

Waves: What was the breakthrough difficulty in high-speed motor R&D? Why is cross-industry perspective so important?

Yu Hao: The difficulty is in optimizing startup code, then finding position, optimizing violent ramp-up. To push everything to the extreme, you can't rely on experience.

Experience has its benefits — based on experience, you can always achieve something. But after 10, 20 years, experience becomes obsolete. Outsider-insider collaboration mainly changes this variable.

Take space computers — radiation particles in space interfere with electron flow in chips. You say, improve it more. Traditional industry people habitually make the walls thicker. But SpaceX moved clusters onto satellites and rockets.

Waves: Do breakthroughs in this era require more generalists?

Yu Hao: Every major breakthrough needs some new type of generalist. Many great entrepreneurs are super product managers, even social inventors. Take electric cars — you need to design charging networks, understand consumer psychology. Look at Elon Musk — he's even more suited to making sports cars. He even knows what kind of car celebrities would like.

We're used to cultivating specialists; many PhDs become narrowists. Actually, reading more leads to generalization, because your methods, tools, and boundaries multiply.

Waves: I heard when you discovered your motor efficiency exceeded Dyson's, the first reaction was self-doubt.

Yu Hao: The team's first reaction was to test again. When we confirmed the surpass, everyone was like: holy shit, we actually did it!?

Waves: Were you surprised too?

Yu Hao: My reaction was: raise the target another 100%.

Now Dyson has reached 125,000 RPM, but we've stockpiled 200,000 RPM and are about to mass-produce 180,000 RPM.

Waves: Because of China's knockoff phenomenon, and frequent patent litigation by overseas companies, I heard Dreame has also had battles with them.

Yu Hao: I've never thought overseas giants were unchallengeable. We chose a different approach ourselves, but even so, they like to sue over everything. I followed the early patent cases closely, but after winning a key patent lawsuit, I let the legal department handle it as normal operations — it became routine.

Waves: How do you view giants protecting their rights this way?

Yu Hao: Contempt. I look down on it.

In 2013-14, Elon Musk said he would open Tesla's patents. There was lots of conspiracy theorizing around it at the time. I thought: what's the big deal? In 2012, when I was still doing Skyworks at Tsinghua, I had already thought about intellectual property and human technological progress. I thought: one day I'll file many patents, the next day I'll make them public.

Intellectual property does protect and drive progress, but today some of it is blocking technological progress. I suspect half the engineering in this world is building patents.

Waves: When you become an industry "giant" yourself someday, will you be cooler about it?

Yu Hao: I don't know — will I become rigid then? History does show many people who, in their last decade of life, did many foolish things. But life is made of cells; you bring nothing when born, take nothing when you die — so why keep things that don't drive progress in your control?

What survives is right, what doesn't is wrong

Waves: Five years into entrepreneurship, what's the most worth sharing insight?

Yu Hao: Choose the hardest goal, start with the simplest thing. Too many people choose small goals but make their first step too difficult, thus neglecting the accumulation that long-termism brings.

Waves: You started with drones in school. Why later choose to begin with high-speed motors?

Yu Hao: It was indeed drone mania at the time — almost every investor's office had one to show they were in tech investing. But I thought drones were a niche market.

We set two principles: first, it must be in millions of households; second, it must have certain technical barriers, and that technology could over time be applied to other fields. Seeing that all Dyson's products in China revolved around high-speed motors, we said: let's directly tackle the hardest thing.

Waves: While R&D-ing high-speed motors, Dreame also developed robot vacuums. "Wanting both" is often a mistake students-turned-entrepreneurs make right after graduation.

Yu Hao: We internally debated whether to do both simultaneously, because it meant double the expense. But this relates to your resource management capability — there's no absolute right or wrong. In hindsight, what survives is right, what doesn't is wrong.

Waves: If you hadn't broken through on high-speed motor core technology then, what would Dreame look like now?

Yu Hao: Honestly, I never thought about it. I have many regrets, but more often I think: should I have taken even bigger steps, rather than wasting life in transition phases.

Waves: Dyson took twenty years from lab to product. Dreame took roughly two to three. As a later entrepreneur, how did you solve your resource scarcity problem?

Yu Hao: Do massive logical deduction, avoid many pits. Then you discover: because your long-term logical choices are correct, resources follow.

Waves: Why choose to join the Xiaomi ecosystem early on?

Yu Hao: We didn't know Lei Jun or anyone around him, so this was a logic-derived choice. The real either/or was between Xiaomi and another company at the time — but the other preferred strong-strong partnerships, while Xiaomi's ecosystem model provided upstream and downstream resources that could help small companies grow fast.

Waves: Though younger brands understand consumers better, giants can quickly incubate new brands. How does Dreame compete?

Yu Hao: Young brands' advantage is no historical baggage, teams with more fighting spirit. But many easily neglect long-term accumulation of core competitiveness. Look at many cosmetics brands — they rise fast by catching one trend, but without their own R&D and factories, without continuous accumulation, they have no pricing advantage. But if you have your own core competitiveness and run faster than others, you can be the better one.

Waves: Strategically, Dreame chose overseas first, domestic second — was this planning or expediency?

Yu Hao: Mainly because we had no money then. Domestic promotion requires lots of money; overseas was much better by comparison. Tactically, overseas happened to catch product upgrade and channel transformation dividends. Also very important: thanks to Chinese companies like Huawei and Xiaomi doing market education overseas in the previous decade, we could bypass ODM and go directly to brand.

Waves: What differentiated feedback have you gotten from domestic versus overseas markets?

Yu Hao: Overseas users care more about hard specs; they treat it more as a pure tool. For a hair dryer, flashy surface treatment, flowers painted on the finish — they don't need that. Chinese consumers are more discerning; years of product progress in China have raised everyone's expectations.

Waves: When will young Chinese consumer electronics brands achieve premium pricing?

Yu Hao: From white-label to brand, this will happen within five years. But brand premium is hard. Because behind brand premium is cultural confidence, is people's aspiration toward a lifestyle. Only when people no longer think European aristocratic life is the best, no longer think American Hollywood life is the best, can brand premium be established.

Waves: What's Dreame's core methodology in building its own brand?

Yu Hao: We didn't deliberately seek methods, but there are several methodologies worth sharing: attack from outside, hit high from low, beat old with new.

Attack from outside means overseas, we first do online well, then offline channels.

Hit high from low means establishing flagship products. Do the flagship well, and everything below flows smoothly. So our average selling price keeps rising, but mainly through new products.

Beat old with new means channel momentum gradually amplifies. When resources are scarce, first use them on small channels — then big channels will gain confidence in you.

Choose to believe everyone is trustworthy

Waves: Dreame now has nearly 1,800 employees. Is management your main contradiction at this stage?

Yu Hao: Many investors ask me this too — how does someone who's never had a job manage so many people now.

My solution was to study why enterprises exist. The conclusion: enterprises must exist to minimize internal transaction costs.

In human social organization, this has upper limits. Basically, a tech company's management ceiling is 100,000 people; a manufacturing company, 1 million; companies with dispatched labor, maybe 10 million. As long as you don't exceed this boundary, there are many tools available.

For example, early on you need flatness; gradually as things get tangled, you need matrix structures; later you need platforms for things to be reusable — that's why many internet companies build middle offices. But middle offices have collaboration efficiency problems, so they get dismantled, become divisionalized.

Waves: Specific methods?

Yu Hao: Find different tools for different stages. In management, I don't pursue 100% goodness — that's nearly impossible. Just stay above average, and you can keep moving forward.

Waves: What's Dreame's most thorny management problem right now?

Yu Hao: Talent density. Dreame has been going for five or six years now; we'll inevitably hit some employee fatigue. Why do many companies go public in year seven or eight? Because that's when fatigue sets in.

My method: give enough opportunities, like having veterans do new things. Also, whether you want to rest or be a coach, you need to reach consensus on new goals.

Waves: Are you fast or slow at decision-making?

Yu Hao: Slow. Regarding decisions, we have a commonly used internal methodology: systematic comprehensiveness, grasp the key, extreme execution — three steps.

Most people think about ten problems, do ten problems. My search range expands tenfold — I think about 100. I'll jump to extremely many dimensions, from global environment and human future down to whether a screw might break. After thinking, I'll pick one or two most critical things, then abandon the rest. Then extreme execution — break one thing into 100 key points and do them thoroughly.

Waves: How do you implant your thinking into the team?

Yu Hao: I basically turn every management meeting into a training session. Besides directly saying what to do and not do, I tell people why and how.

Waves: Why?

Yu Hao: Many people work their whole lives accumulating more experience without understanding why. Once they switch industries they're completely lost. You need to tell them the thinking process behind it — on one hand they can better understand your decisions, on the other they can think this way too.

Waves: When you make personnel adjustments or layoffs, do you morally interrogate yourself?

Yu Hao: Definitely. I believe everyone is good. When you make adjustments, if you've truly considered their interests, given enough opportunities, and what remains is a very small portion of cognitive mismatch — then there's no way around it.

Waves: Talking to you now, your competitiveness is very strong, and completely unhidden.

Yu Hao: Being number one is a habit; once you have the habit, it becomes an addiction. I attended elementary school in the village, middle school in town, high school in the county. At first, you're definitely not number one — there are always classmates getting better education. But basically after half a year, you find the trick and become number one. It's the same at a company — once people get used to this temperament and atmosphere, it can be sustained, then infect the whole team.

From university to now, I've encountered many setbacks, but no major failures. I also went to the US for exchanges, which gave me a kind of confidence when researching high-speed motors: we're all human, why couldn't we?

Waves: But many people easily feel defeated, even paralyzed by it.

Yu Hao: Everyone is a little genius at birth. But when you repeatedly tell them to follow the rules, to do what age-appropriate things, to accept this and that, they gradually become ordinary. But even now, I still believe in myself, believe that 90% of life choices don't matter — I only do the 10% that matter most.

Waves: These years of entrepreneurship, any new understanding of people?

Yu Hao: When doing Skyworks at school, a friend told me: every organization you establish needs a basic assumption about the people in it — are they trustworthy or not? I chose to believe everyone is trustworthy. Back then school reimbursement was troublesome; we directly hung bags at the door for people to put invoices in, then automatically paid them.

After founding a company, I've encountered betrayals, even harm. But after filtering all information, I find I can still return to this assumption: Everyone is trustworthy.