When category creators stop building
Last month, iRobot filed for bankruptcy and agreed to be acquired by Shenzhen Picea Robotics, a Chinese contract manufacturer. The company that invented the robot vacuum, that sold over 50 million Roombas, that once commanded 64% of the global market, will now be owned by the very ecosystem of competitors that displaced it. The sale price? Roughly $140 million in debt forgiveness. Amazon had offered $1.7 billion just two years earlier.
What happened to iRobot isn't a mystery. It's a pattern. And it reveals something fundamental about why Product First matters, especially in markets where incumbent innovation slows, and new technology curves make rapid iteration possible.
The stagnation trap
iRobot didn't fail because it lacked talent or resources. They invented the category. They pioneered the circular robot design, Virtual Wall boundaries, acoustic dirt detection, and self-emptying docks. For 14 years, they defined what a robot vacuum could be.
But somewhere along the way, they stopped building with urgency.
They persisted with camera-based navigation while competitors adopted LiDAR. Their product cycles stretched to two years while Chinese rivals shipped major updates every six months. They maintained premium pricing without the differentiation to justify it. By the time iRobot finally introduced LiDAR in 2025, Roborock had been shipping it for nearly a decade.
This is the stagnation trap.
Category creators build early advantages through genuine innovation, then gradually shift their energy toward protecting market position rather than extending it.
R&D slows. The organization prioritizes efficiency over innovation. And when new technologies emerge that enable faster, cheaper, and better products, the incumbent can't respond quickly enough.
Builders who brought something new
What struck me about the Chinese competitors wasn't that they copied iRobot. Each one brought genuinely differentiated insights from adjacent domains and translated them into product innovations that users could immediately feel.
Roborock's founder, Richard Chang, came from Baidu Maps. Applying that expertise to robot vacuums, he achieved something iRobot hadn't: affordable, accurate LiDAR navigation that made methodical room-by-room cleaning possible for the mass market.
Dreame's founder, Yu Hao, was an aerospace engineer who'd studied computational fluid dynamics at Tsinghua. He applied aerospace turbine principles to achieve 150,000 RPM, shattering an industry barrier that had stood for years.
Narwal's founder, Junbin Zhang, heard his father complain about back pain from mopping floors. Zhang combined his technical background with a specific user problem and invented the first robot that could mop and clean itself automatically.
The builder's advantage
What iRobot's collapse teaches us isn't that incumbents always lose. It's that innovation velocity determines outcomes in technology markets.
When new curves emerge, whether in motors or sensors, or AI, the companies that move fastest to incorporate them into genuinely better products will win.
This requires a specific kind of builder: someone with deep domain insight who sees an opportunity to reimagine what's possible. Chang understood mapping. Yu Hao understood turbines. Zhang understood the problem of keeping floors clean.
These builders didn't wait for permission or consensus.
They shipped. They learned from users. They iterated with urgency. They combined technical depth with the conviction that they could make something fundamentally better.
This is what we mean when we say we invest in A+ builders. Not credentials or pedigrees, though those can signal potential. But the product itself. The choices embedded in it. The craftsmanship that reveals how someone thinks, prioritizes, and executes under constraints.
A great early product is living evidence of courage to act, capacity to learn from feedback, and tenacity to persist when the path isn't clear. It carries the fingerprints of its maker in every decision.
Conclusion
At N47, when we talk about investing when the product speaks loudest, this is exactly what we mean.
The robot vacuum market is a $9 billion category growing at 17% annually. The company that created it is now worth less than a single year's R&D budget for its leading competitor.
That gap didn't emerge overnight. It accumulated through thousands of product decisions over a decade.
We believe there are builders right now, in categories we haven't yet imagined, who are about to create this kind of shift. They're working on products that will make everything before them feel obsolete.
They may not have impressive metrics yet. But they have insight, urgency, and a product that's starting to capture genuine user love. This is what helped Chang, Hao, and Zhang stand out and overtake the incumbent iRobot.
Roborock's robots cleaned in efficient patterns instead of bouncing randomly. Dreame's vacuums picked up debris that competitors left behind. Narwal's mops actually cleaned themselves instead of spreading dirty water.
Users noticed. They told friends. They posted videos. The products generated the kind of organic advocacy that can't be bought.
This is what we look for at the earliest stages: the unmistakable moment when a product creates that response. Not vanity metrics or manufactured buzz, but the real signals that emerge when users encounter something that solves their problem in a way that feels both obvious and brilliant. Social posts from a delighted customer. Feature requests that reveal unexpected use cases. The early organic pull that proves genuine problem-solution fit.
We invest before the consensus arrives, when most VCs would say it's too early.
Because by the time the charts and cohorts look impressive, the opportunity to partner with builders at the decisive moment has passed.
That's the moment we want to find you. Not when everyone agrees you're building something important, but when your product speaks for itself, and the conviction to keep building is still an act of faith.
If you're that kind of builder, we'd like to see what you've made.
