The laundry-folding robot is this decade's self-driving car — overfunded, under-delivered, and a perfect test of whether AI can handle the real world's wrinkled chaos.
The Summary
- Figure AI, Sunday Robotics, and Weave Robotics — collectively worth billions — are racing to build laundry-folding robots, betting that home automation is finally ready for mass adoption
- Roboticists say folding clothes is an ideal "safe chaos" testbed: hard enough to prove dexterity with deformable objects, low-stakes enough that failure just means wrinkled shirts
- The industry has hyped laundry bots before and failed to deliver — the difference now is foundation models and real-time learning from human demonstrations
The Signal
Laundry folding is what roboticists call a "deformable object manipulation" problem. Clothes bunch, wrinkle, shift unpredictably. They don't have rigid edges or fixed geometries. Every T-shirt folds slightly differently depending on fabric weight, how damp it is, how crumpled it was in the dryer. For a robot trained on pristine simulations, this is nightmare fuel.
Sunday Robotics CEO Tony Zhao says the company "looks for hard things to do." Translation: if you can nail laundry, you can probably handle dishes, groceries, and eventually eldercare. Figure AI called it "one of the most challenging dexterous manipulation tasks for a humanoid robot." But here's the key insight roboticist Ayanna Howard gives: when the robot drops your shirt, nothing breaks. No shattered plates. No spilled hot coffee on a toddler. Just a wrinkled shirt you were going to wear anyway.
"If you drop a piece of clothing, it's not going to break."
This is the unlock for home robotics. Not that the task is easy, but that the failure mode is acceptable. You can't beta test a surgical robot or a self-driving car in someone's living room. But a laundry bot? Worst case, you refold the jeans yourself. Best case, you never touch laundry again and the robot learns from watching you fix its mistakes.
The three companies racing here — Figure, Sunday, Weave — have raised billions collectively. Figure alone is valued north of $2 billion. That capital is betting on two things:
- Foundation models for robotics (think GPT, but for physical actions) are finally good enough to generalize across tasks
- Consumers will pay premium prices for robots that do multiple chores, not just vacuum in circles
The laundry focus is strategic. It's visible, it's visceral, it's a chore people actually hate. Nobody demos a robot that quietly optimizes your home energy usage. You demo the one that folds your kids' tiny socks while you watch Netflix. It's proof the agent economy isn't just software eating the world — it's hardware folding your world.
The Implication
If laundry bots work, the home robotics market goes from "maybe someday" to "investor feeding frenzy." These companies aren't building laundry appliances. They're building general-purpose humanoid platforms that learn new tasks over time. Laundry is task one. Dishes, cooking, eldercare, home security — those come next, uploaded as new skills.
Watch what happens when one of these companies ships a $3,000 robot that actually works. Not perfectly. Just well enough that early adopters tolerate the occasional misfolded shirt because the robot also vacuums, loads the dishwasher, and puts away groceries. That's when humanoid robots stop being a CES curiosity and start being a category.