Sunday Robotics exits stealth with Memo, a home robot trained on 500 real households.

Sunday Robotics has emerged from stealth with Memo, a home robot pitched at everyday chores — washing dishes, doing laundry, tidying up — and a pointedly pragmatic design. Rather than legs, Memo rolls on a wheeled base, a choice its founders argue trades away bipedal spectacle for the stability and cost profile a domestic product actually needs. The company is backed by $35 million from Benchmark and Conviction and led by Stanford roboticists Tony Zhao and Cheng Chi.

Memo's differentiator is its data strategy. Instead of learning primarily in simulation, Sunday trained the robot on authentic routines captured in more than 500 real homes using a patented 'Skill Capture Glove,' a roughly $200 wearable that records how people actually move, clean, and organize. That bet — that cheap, high-volume human demonstration data beats synthetic data for messy domestic tasks — is a direct counterpoint to the simulation-heavy approaches favored by better-funded rivals, and it is the thesis investors are underwriting.

The go-to-market is deliberately narrow: a 'Founding Family Beta' opened November 19, with about fifty households slated to receive individually numbered units when the program launches in late 2026. That cohort scale is honest about where home humanoids really are, and it reflects a strategy of tightly controlled real-world learning loops over a splashy mass launch.

For the category, Memo is the latest evidence that the home is the contested frontier — arriving in the same quarter as Figure 03 and 1X's NEO pre-orders — and that not everyone believes the winning form factor is a full bipedal humanoid. A well-credentialed team arguing that wheels plus a wearable-data flywheel can beat legs plus simulation is exactly the kind of contrarian bet worth watching as the field sorts out what a home robot should actually be.