FOLD is a physical-AI appliance that automates the last manual step of laundry.
Dump. Close. Done.
Every other room in the house has had its machine moment. The kitchen got the dishwasher. The washer and dryer became background infrastructure. Folding is the one chore we still do with our hands — forty hours a year, per household, by hand.
A physical-AI appliance. The size of a washer. One button.
Each cell of FOLD is a small answer to a hard robotics problem. Together they close a loop that has resisted automation for decades: perceive the pile, isolate one garment, fold it, remember who it belongs to.
A multi-camera and depth-sensor array builds a live 3D reconstruction of the laundry. It identifies garment type, dimensions, fabric behavior, seams, orientation, likely owner, and matching pairs — before anything moves.
Controlled airflow, vibration, rotating surfaces, vacuum pickup, and computer vision cooperate to make a tangled pile behave, briefly, like discrete objects. The hardest unsolved problem in textile robotics — isolating one garment from many — handled with physics instead of fingers.
Twelve processing cells operate simultaneously. Thirty mixed garments resolve in roughly thirty seconds because the work happens in parallel, not in sequence. Throughput becomes an engineering number, not a user experience.
FOLD builds a private, per-household wardrobe model. Michael's black T-shirt. Sarah's leggings. Daniel's football shirt. Over time it knows your closet better than you do — on-device, nothing leaves the home.
Removable per-person containers sit at the base of the machine. Folded garments drop directly into the right owner's stack. Open the lid; lift Michael's dock; walk it to his room. The last twenty feet of the laundry journey — gone.
Robotics has solved rigid objects. It has not solved textiles — deformable, high-dimensional, infinite in configuration. FOLD is the wedge: a mass-market consumer appliance that generates the training data to make general textile manipulation possible. Everything else is downstream.
FOLD One is a one-time purchase. Later generations will drop into mainstream price bands. Folding your own clothes should never require a subscription.
The math.
135 million US households. 40 hours of folding a year, per household. A chore that has resisted automation precisely because it is hard.
Price it like a washer. Build it like Apple. Sell it as the thing you never knew you'd always wanted.
Every cycle logs a physical garment interaction — perception, grasp, singulation, fold outcome. The flywheel is simple and brutal: more machines, more interactions, better models, better folds, more machines. In three to five years, FOLD holds one of the largest datasets on physical manipulation of textiles in the world. There is no shortcut to buying that.
Each entry on this line erased a weekly chore. We're building the next one.
A seed round for a category-defining appliance. We're talking to a small group of institutional partners who understand hardware, AI, and consumer category creation.
A short note goes to Mike directly. Expect a reply within a day.