I'm pretty certain that if these were actually ready there'd be commercial uses of them first, where they see a lot more use and thus generate a lot more value than any household has laundry.

Robot operated laundry on a cruise ship or something.

The goal at this stage is largely training data collection so it can reach widescale use. Just like self driving variations in multiple different cities, the data needed for AI robotics is broad with a million niche usecases, so it makes sense it's not strictly commercial.

They need visual recording of tele-operated robots (or humans with headset cameras) doing normal household stuff like folding laundry in real environments so it can be fully automated. Which is what funds a lot of this stuff since that training data is a goldmine right now if a company can collect enough of it.

Robots are used extensively for commercial purposes, though.