Robotics research and company milestones · Updated 24 October 2026
Home robots are getting better at unfamiliar tasks
October has brought a care-specific robot prototype and a new domestic companion pre-order programme. Figure’s September demonstration points to something just as important: machines that can transfer what they have learned between homes.
Home robotics is moving beyond machines trained for one particular setting. The most promising programmes are starting to combine visual understanding, language and increasingly capable movement. There is a substantial gap between a company demonstration and a reliable home carer, but that gap is becoming easier to describe and measure.
Kawasaki Home LEO: a robot designed around later life
On 6 October, Kawasaki announced a completed prototype of Home LEO. Unlike a general-purpose factory robot, this machine is being developed specifically for older people living at home. The company describes goals including personalised conversation, noticing unusual situations and finding or bringing small objects. Some of these functions are planned for later development, not established capabilities of the existing prototype.
Kawasaki plans collaborative trials with Japanese public bodies and care organisations, aiming for a launch in Japanese fiscal year 2028. The programme is worth following because it connects a major industrial robot manufacturer directly with the needs of people ageing at home. UK availability and price have not been announced.
Figure Helix 2.5: familiar tasks in unfamiliar homes
Figure reported on 17 September that its robot carried out three long-running household tasks across 30 previously unseen Bay Area homes: tidying living rooms, folding towels and making beds. According to Figure, the model did not need data collection or retraining in those homes.
Figure reports complete success in 56% of trials for the model trained on its broad human-behaviour dataset, compared with 9% for an otherwise comparable model trained from scratch. That is a substantial gain in transfer to new surroundings, and a useful reminder that the robot still failed to finish many attempts. A future care assistant has to cope with changing rooms, objects and routines. These were company-reported household tests, not independent trials of personal care or unattended emergency response.
Figure’s demonstration and technical account · Figure in the humanoid tracker
AiMOGA Mobo: carmaker investment in domestic companions
Chery announced global pre-orders for Mobo on 21 October. AiMOGA describes a humanoid for household companionship, with family recognition, memory and physical navigation as parts of its intended system. The company has named an initial group of 35 early users.
It is a revealing move from a large vehicle manufacturer into domestic robotics. The announcement does not establish UK stock, an orderable UK price, delivered home systems or independent results from older-adult care.
What would count as the next big advance?
Regular success across many ordinary homes would be more persuasive than another impressive video in a test room. We are watching for measured task completion, safe behaviour around people, service and repair arrangements, and independent testing. Cost and support will matter too.
There is reason to be optimistic. The underlying AI is improving quickly, manufacturers are investing in hardware at scale, and specialised care robots are beginning to receive serious development programmes. The first useful household assistants will probably do fewer jobs than the robots in concept art, then improve as the systems learn to handle more of daily life.