The Residence Changes From Within

For much of residential history, the home operated through human labour and human memory. Household knowledge sat with the people inside it. They knew what needed to happen, when it needed to happen, how the residence should be prepared and who could be trusted to do the work.

Much of that knowledge was never written down because it did not need to be. One person could explain something to another, and another human being could understand the meaning behind the instruction.

Technology gradually absorbed parts of that work. Electricity changed the physical operation of the home and appliances reduced labour that had once consumed entire days. Computers brought outside information into the residence. The World Wide Web accelerated that exchange. Smartphones then collapsed communication, photography, banking, navigation, shopping, entertainment and administration into one portable interface.

Each shift looked incremental while it was happening. Across time, each changed behaviour, expectations and the structure of daily life.

We are inside another shift now, and it is still too often being treated as an extension of the smart home.

For roughly two decades, residential technology centred on control. Lighting, climate, security, entertainment and access became connected. Household records moved into software. Property platforms and client portals brought maintenance, suppliers, schedules and documents together.

The residence became easier to see and easier to manage from a distance, yet it still depended on people to notice what had happened, understand its significance and decide what should follow.

Artificial intelligence reaches further into that process because it can absorb parts of interpretation, coordination and execution. AI agents can already retrieve records, operate software, organise schedules, monitor information and coordinate administrative work. Physical AI is carrying those abilities into machines that can perceive their surroundings and act within the material world.

The Evidence Has Moved Beyond the Laboratory

On 8 September 2026, XPENG announced that it had commissioned its humanoid robot production lines in China, with more than 80 per cent of core processes automated; the first IRON humanoid walked off the line autonomously. 1X is taking orders for NEO, a humanoid designed for the home, with first deliveries planned for 2026. LG presented CLOiD at CES 2026 as part of its “Zero Labor Home” vision, combining household robotics with connected appliances and Physical AI.

UBTECH is moving in the same direction from the service sector: its Walker C1 is being positioned for hospitality and other public service environments, including reception and visitor guidance.

The important development is not one robot. It is the ecosystem forming around them.

Manufacturing is beginning to scale at speed. Household robotics is becoming commercial. AI agents are gaining interoperability. Robotic hands are beginning to reproduce another capability that has kept physical work firmly human: touch.

Sharpa’s Wave hand mirrors the scale and structure of the human hand, with five fingers, 22 active degrees of freedom and fingertip tactile sensing that detects very small forces and supports slip prevention. Tacta is developing human-sized hands that sense force, contact, texture and temperature. The significance is not that robots now possess a human hand. They do not. It is that dexterity and tactile perception are advancing together, narrowing another barrier between machine reasoning and physical execution.

The infrastructure around those machines is changing as quickly.

Work is also under way to give household robots structured knowledge of the home itself (see When the Residence Stops Waiting to Be Managed).

The technical problem is becoming clear. A machine needs more than the ability to perform a task. It needs enough structured knowledge to understand where it is, who it is acting for, what is permitted and what the intended outcome actually is.

Why Context Is Harder Than Capability

That is where operating autonomously inside a private residence becomes substantially more complex.

A humanoid may be able to feel pressure through its hands, fold clothing, clean a room, carry objects or prepare food. None of that tells the machine what should happen next inside this particular residence.

It does not know from general training that one person requires a room prepared in a specific way while another does not want anyone to enter it. It cannot assume which garments may be handled, which surfaces require specialist care, whether a guest suite should remain untouched, who can authorise a supplier, or when a change in arrival time must alter security, transport, provisioning and staffing.

A robot can learn the physical act. The harder problem is understanding what that act means inside one residence, for one person, under changing conditions.

That knowledge determines which action is appropriate, when it must occur, whose authority permits it, what standard must be met and what happens when the expected conditions change.

Experience Is Becoming Operational Information

Much of it still sits across people, conversations, memory and disconnected systems.

Experienced estate managers, housekeepers, chefs, personal assistants, engineers, security professionals, concierges, lifestyle managers and trusted suppliers carry years of contextual understanding. They recognise patterns, exceptions and preferences that may never have been formally documented because another person could infer what was meant. They know when an instruction is incomplete, when a routine action is sensitive and when a technically correct response would still be wrong.

That knowledge has traditionally been called experience. Increasingly, it also has to be understood as operational information.

A modern private residence already generates and accumulates detailed records of assets, suppliers, maintenance, security, staff, inventories, household preferences and recurring routines. Family offices and residence teams may organise that information through property platforms, client portals and specialist systems.

Those records matter, but they are not yet an operating model.

A record can say what exists. It does not necessarily explain who may act, what takes priority, what depends on what, or when the normal procedure must be suspended.

Autonomous systems need those relationships, not simply the records themselves.

Adoption Will Not Wait for Agreement

The smart home was designed around human control. We are now moving into a different way of living. The intelligent residence has to understand context: who is present, what the intended state is, what has changed, which actions are permitted, which systems may act independently and where human judgement remains mandatory.

That transition will not be accepted simply because it becomes technically possible.

The resistance is already visible. Some people see AI and autonomous systems as a means of removing repetitive work and administrative burden. Others see the same technologies as a threat to employment, privacy, expertise, identity and control.

The technology does not pause while that argument is resolved.

Previous transitions did not require universal agreement either. Automation, computers, the internet, mobile technology and digital finance developed while society debated their consequences. Adoption moved unevenly, but the underlying capability continued to advance.

This transition will be no different.

Preparing the Residence to Be Understood

The home raises the stakes because it contains family, privacy, safety, security, possessions and deeply personal routines. A machine operating in a factory works inside an environment designed for repeatable production. A machine inside a private residence enters a space that changes with the people living in it.

Technology is advancing faster than human acceptance and understanding. For PRC, that does not make the transition less likely. It means the conditions around its use become more important.

Private residences need clear boundaries around what autonomous systems can perceive, retain, decide and do. Authority has to be explicit. Actions have to be recorded. Human override cannot become an afterthought added when something fails. Privacy, security and data ownership become part of residential operations because a machine that can see, remember, learn and act inside the home carries a different order of access from any appliance that came before it.

The deeper change is already happening inside the residence before it is visible from the outside.

Household knowledge that once moved informally between people is increasingly becoming structured, governed and machine-readable. Rooms, objects, routines and preferences will no longer exist only as physical details or isolated records. They will carry relationships, conditions and permissions that determine how the residence functions.

Humanoid robots will not create the intelligent residence merely by entering it. Their arrival will expose whether the residence has been made intelligible enough to direct them.

The machine is learning how to see, reason, move and increasingly feel. Standards are being written so it can understand the home around it. Manufacturing is beginning to scale.

What remains unresolved is the operating intelligence of the individual residence itself, and how humans adapt to living alongside systems that increasingly know how to act.

PRC has to prepare for that direction of capability now, rather than build around the assumption that today’s residential operating model will remain permanent.

That is where the next residential transformation will be decided.