A San Francisco woman has given a glimpse of what home cleaning could look like in the age of humanoid robots after hiring one for $30 an hour to work inside her apartment. Gianna Capezio booked a one-hour cleaning visit from San Francisco startup Tau Robotics, with the humanoid robot vacuuming, making her bed and cleaning her bathroom during the session. The robot was not operating autonomously. A trained human operator controlled it remotely through its cameras, with artificial intelligence assisting the system. Capezio said the robot was slower than a human cleaner but still did a solid job and she did not have to redo its work afterwards. Her video of the unusual cleaning session quickly spread online, attracting millions of views and sparking debate over the technology, its usefulness and the future of household work.
Get breaking news anytime, anywhere. Download the TOI app now!
San Francisco woman hires a humanoid robot for a $30 cleaning visit
Capezio's experience came after Tau Robotics began offering its humanoid cleaning service to selected customers in San Francisco. The company currently operates the service on an invite-only basis, with customers paying a flat $30 for a one-hour visit. Tau advertises its robots for tasks including vacuuming, wiping surfaces, taking out rubbish and clearing light clutter, while its current service also offers options for kitchen, bathroom and deeper cleaning. Customers cannot simply leave the robot alone inside an empty property. Tau's terms require the customer or another authorised adult to remain at home throughout the visit, and the company does not accept keys, lockboxes or door codes. The robot also depends on a working internet connection and electricity because it is operated remotely. The service is therefore still an early-stage experiment rather than a conventional autonomous cleaning business, but it represents a notable step because a humanoid machine is now being sent into real homes to perform paid household work. Founded in 2024, Tau has raised $2.6 million in pre-seed funding and is still operating with a small engineering team, underscoring how early the company’s commercial rollout remains.
