Figure AI, a perceived leader in the international race to bring smart robotics to the masses, is accelerating its push to get its AI-powered humanoids into households.
The unicorn startup says its self-developed Helix AI model that powers industrial applications can now accomplish the most dreaded domestic tasks, especially on laundry day. If Figure’s robots can crack laundry, it signals that almost any other task could follow.Laundry is the perfect storm of boring for humans , and brutal for robots.
For most people, folding a tee shirt is so routine, they do it without thinking. For robots, every garment is a unique puzzle. Unlike rigid objects like cups and boxes, clothing has no fixed shape, flops around, and changes based on how it’s grabbed.
Real-world chaos like clothes tangling, socks hiding inside sleeves , and fabric bunching unexpectedly makes computer vision and grasp planning far harder than picking up a box in a warehouse. There’s no standardization. Even two identical shirts from the same brand may shrink or stretch differently after washing.
Laundry automation basically plateaued in the 1950s after adoption of fully automated home washers and dryers became widespread. The smart home boom of the 2010s inspired specialized laundry bots, like the Laundroid and Foldi Mate, that ultimately failed to take off.
The international race toward superintelligence has reignited interest in full lifecycle laundry automation. Humanoid robots are seen as the best equipped to take on laundry, a chore that’s built around human environments.
In Silicon Valley, Figure founder Brett Adcock has been talking up his unicorn startup’s humanoids’ laundry day capabilities. He recently shared a video of a second-generation Figure robot loading clothes from a laundry basket and into a front-loading washing machine.
While humanoids like Tesla’s Optimus and 1X’s NEO have shown promising demos, Adcock said Figure’s robots are the only ones that do it fully autonomous. The robot’s brain is an AI system called Helix, a vision language action model that integrates perception and language understanding to control its manipulation of objects.
When laundry biz entrepreneur Tyler Purcell poked fun at the Figure 02’s demo, Adcock fired back, saying: “This guy runs a $100M laundomat business and he’s scared. He should be.”
He said the robot’s hardware is already complete, with the dexterity, reach, and mobility needed to perform general-purpose physical tasks. He said the only limiting factor now is training data for its Helix model.
The founder followed the washing machine video with this demo of a Figure 02 humanoid folding 10 towels for almost three minutes. He called it the first time a humanoid robot folded laundry using a neural network. He said engineers applied the same Helix architecture Figure robots use for logistics tasks, with the only change being a new dataset.
According to Figure, Helix learned to pick towels from a pile, adapt its folding based on how each starts, recover from mistakes, and use fine hand movements to prepare and fold neatly. Figure is also training the AI robot brain to maintain eye contact, direct gaze, and use their hands for more effective communication.
The company is actively fast-tracking its humanoid robots for home deployment, accelerating its original timeline by two years due to internal confidence in the Helix model, which was launched after Figure severed ties with OpenAI. Adcock has said there’s a path toward manufacturing 100,000 humanoids within the next four years.
Figure is one of many well-funded upstarts racing to get its AI-powered machines into homes.
Another Silicon Valley firm, 1X Technologies, has begun testing its NEO androids in select homes and plans to start scaling production in 2026. In a new video shared by 1X founder Bernt Bornich, a NEO Gamma humanoid carries a bag of rice weighing 20 kg (45 lb) which is about 64 percent of its 32 kg (70 lb) frame.
Another California startup, Weave Robotics, plans to start shipping initial orders of its wheeled Isaac robots for homes this year. The robot responds to voice and text commands, and has demonstrated the ability to fold laundry, feed pets, and automatically clean rooms. It’s priced nearly $60K.
In China, Shenzhen-based Astribot is one of many rising robotics firms focused on useful household applications like folding laundry. Its flagship humanoid, the Astribot S1, moves via a wheeled base, with a single bending leg for stability. Pricing hasn’t been publicly disclosed yet.
Others are betting on non-humanoid form factors for mundane domestic applications.
Stanford roboticist Aaron Tan has opened preorders for his Lume concept, a pair of bedside lamps that, when activated, reveal robotic arms to fold laundry placed on a nearby surface. A CGI video showing the system elegantly folding blankets went viral.
Though it’s just a visualization, Tan told The San Francisco Standard that a physical prototype exists in his garage , though it’s not yet as sleek as the concept video. He’s targeting a 2026 shipping date and pricing under $2,000.