Last week, I handed a 50-pound sack of dirty laundry to a courier at my apartment building’s front door. Four hours and $55 later, the clothing and linens came back clean and folded into compact rectangles, stacked edge to edge.
But I didn’t use any old wash-and-fold delivery service and some of that crispness wasn’t human handiwork. Around 15% of my dirties had been folded by a robot named Isaac, who looks a little like Johnny 5 from “Short Circuit” (a human folder completed the rest of my laundry).
Isaac was built by San Francisco startup Weave Robotics (opens in new tab) and recently installed in the North Beach laundry room of the startup Tumble (opens in new tab), which operates app-powered washers and dryers at multiunit buildings in 11 states, plus a four-hour couriered wash-and-fold service in San Francisco.
The partnership launched a few weeks after the companies’ founders, Weave Robotic’s Kaan Dogrusoz and Tumble’s Scott Patterson, connected on X. Dogrusoz posted about his company’s advances with Isaac, catching the eye of Patterson, who’d been tinkering with building his own laundry-folding robot, despite limited experience with robots.
Patterson reached out with an idea: Perhaps, instead of trying to build his own robot from scratch, Isaac might be up for a trial job in a Tumble laundry room? Two weeks later, Isaac was installed.
“Laundry rooms are the perfect place to deploy robots and support autonomy,” said Patterson. “If we can turn them into little autonomous factories for laundry, everyone would send their clothes out.” Plus, traditional wash-and-fold services are slow and pricey, he argued. “With robots, the goal is to make it affordable enough for everyone.”
Dogrusoz, a former Apple engineer, cofounded Weave Robotics in 2024, with a goal of creating in-home robots that handle chores.
“Laundry wasn’t just a question of what we could make a robot do. It’s one of the first things that I would actually want done at home,” he said. “This is a really cool opportunity to prove our entire hypothesis … that a general-purpose robot can be built in a year and then scaled up in the real world, doing real things.”
So how’s it going? We sat down with Patterson and Dogrusoz to learn more.
This interview has been edited for clarity and length.
I’ve seen Isaac at work, and it’s the real deal. But it’s sloooowww, taking a couple of minutes to fold each item. So when will Isaac be folding all my clothes within the four-hour window?
Patterson: We don’t have a timeline yet. Soon we’ll be doing 50%. But this is the slowest it’s ever going to be.
How does the robot work?
Dogrusoz: We have two RGB (red, green, and blue cameras which capture color cues for navigation) for Isaac’s head, so that it has visual input. And then it also has sensors in each of its motorized joints so it can get into what is called “a proprioceptive state”: an internal sense of arm position so that it can basically know what it is doing.
This is one of our stationary units, primarily focused on manipulation tasks like folding. Future models have omnidirectional bases so they can move around apartments. We were able to economize because we don’t have a motorized wheelbase. It has a telescoping torso, so it can reach the floor and then self-compact.
Different fabrics end up really mattering. Some athletic wear is really, really difficult to manipulate, but we’re getting better week over week.
It seems like SF roboticists are obsessed with laundry — there’s the Lume lamp folding robot, Figure’s Helix in Sunnyvale, a humanoid bot (opens in new tab), which can pick up your dirties off the floor and fold them. Physical Intelligence raised $400 million (opens in new tab) for a laundry-folding and -loading robot (opens in new tab), and Dyna Robotics, a Redwood City startup, which has installed its folding robot at a laundry (opens in new tab) in Sacramento. Do you all hate folding laundry that much?
Dogrusoz: I personally don’t think that I need to be folding laundry the rest of my life, although I’m at peace with the fact that we have to live with it for a few more years. I feel, build the world that you want.
We’re focused on the specific tasks that are perhaps the most menial and loathed in the home. That’s the way that robots earn their keep and earn the trust to be in a place that is that intimate. It was one of the first things that I would actually want [a robot to do] around my home.
But why laundry and not, say, loading the dishwasher or cleaning the toilet??
Patterson: Laundry folding is literally a benchmark for robotics. Take folding boxes: There’s six sides to a box and only so many ways you could fold it, but a pair of shorts could have an infinite number of configurations for the robot to pick up and figure out how to fold. That’s why you see laundry folding come up as this litmus test of robotics, because if you can figure out how to do laundry, you can figure out how to do anything.
Is Isaac going to put humans out of work? Or will laundry robots always need human help?
Patterson: It’s really collaborative robotics; it’s meant to be with a person. The washing machine does most of the work, and then the robot does the folding. The person can do the other stuff while the robots take care of folding. We expect this to create more jobs in this area, because we want the robot to work alongside a person.
The doom and gloom about AI has been like, “Oh, it’s meant to take our jobs away.” Instead, it’s just, like, a huge value unlock. If you drive the cost down on this, you’re creating a service that everyone wants and — now — one that most people could afford.