Unready but Unstoppable

Humanoid robots are coming to our homes. 1X has thousands of pre-orders to lease its Neo robot for $499/month and at the China International Import Expo (CIIE) last month, which hosted 900,000 visitors, humanoid robotics had prominent placement across the technology halls. AgiBot, Astribot, Unitree all demonstrated systems aimed at household tasks and each is working to bring component costs into consumer range. Unitree’s G1, a 3/4-size humanoid can be purchased by consumers now for just $13,500.

Humanoid robots in the home are not speculation about a distant future. They are happening, but not because of any real technological breakthroughs, because of simple, yet brutal economics: for most people, we have priced human care out of reach.

The mathematics of why robots are coming

For the elderly, the gap between coping and not coping is not a gradual slope. It is a trap door. Most do not need continuous clinical attention. They need help in the moments where daily life turns dangerous: the high shelf, the low chair, the stairs in the dark. These moments might add up to a few minutes a day. The rest of the time, they are fine.

But the system does not offer a few minutes of help. It sees independence or dependence, nothing in between. Once you cannot manage every task alone, the only option it recognises is residential care. You don't need a nurse; you just need a steady hand. But to get that hand, you are given a bed in a home and the loss of everything else.

Struggle with the stairs and you could lose your house.

And that shift carries a price. A typical care-home place runs £5,000 or more per month. Not because the care is medically intensive. Because the system has few other options and more are undesirable. This is where imperfect robots find their opening. Not because they are good enough, but because the gap has been left empty. A teleoperated humanoid cannot match human compassion or judgment. But it can fetch medication at 3am, steady someone rising from a chair, call for help when needed. It is not a solution. It is not a solution. It is the absence of one, made into a product.

The mathematics are stark: three weeks in a care home costs the same as an entire year of robot support at home. If those are the only options, patchy support from a humanoid is going to beat forced dependency in residential care.

Why humanoid robots aren’t ready for the home

Stage demonstrations make humanoid robots look nearly ready. They can fold laundry, put items in a fridge, and whisk an egg, really! But a stage is not a kitchen, and the gap between a one off performance and what works in a home has a name. Software engineers call it the ‘ninety-ninety rule’. It’s a humorous adage attributed to Tom Cargill of Bell Labs who noted “the first 90 percent of code takes 90 percent of development time, and the remaining 10 percent takes the other 90 percent.” In robotics, this is even less forgiving. The first ninety percent of capability is visible, fundable, demonstrable. The remaining ten percent contains the improvisation, the edge cases, the irregular conditions that define domestic life. That final fraction costs as much as everything before it, and with infinite edge cases, there is no finish line.

LLMs got as far as they did because they had something to train on: the entire searchable internet, compressed into statistical relationships. Data was the foundation. Robots have no equivalent. The human hand alone contains around 17,000 touch receptors and free nerve endings in the palm. The fingertips have one of the highest densities of nerve endings in the body, with some 2,500 receptors per square centimetre. And there is no dataset to help a robot navigate this complexity. No corpus of touch experiences that can be compressed and learned at scale. Every capability must be earned through trial, failure, and supervision. Over and over again.

This is why impressive demonstrations fail to map cleanly to real homes. A humanoid can pour a drink on stage, under controlled lighting, on a stable surface, with a glass it has seen before. It cannot generalise that action across the variability of real households: the shifting light, the wet counter, the unfamiliar mug, the cat. The gap between public perception and operational reality hits here, in the absence of data that would make physical intelligence learnable at scale.

The ghost in the machine has always been human

So what fills that gap? Well, the same thing that has filled it for the last 250 years. Us.

The technology industry has a recurring pattern. What is advertised as artificial intelligence is often human labour in disguise. The story begins with the original Mechanical Turk, Wolfgang von Kempelen's 1770 chess-playing automaton that toured Europe defeating opponents for decades. The secret was simple: a chess master hidden inside the cabinet, operating the machinery through an elaborate system of levers and magnets.

This deception has repeated itself with remarkable consistency. Amazon's "Just Walk Out" technology, touted as being powered by computer vision and artificial intelligence, actually relied on roughly 1,000 workers in India reviewing what customers picked up and walked out with from stores. In 2022, about 700 out of every 1,000 transactions required manual review by these workers. The magic of picking up groceries and walking out was magic. But it was human magic, not AI magic.

Now Neo repeats the trick. 1X CEO Bernt Børnich explained that anybody purchasing Neo for delivery next year must accept that human teleoperators will be controlling the robot remotely and seeing inside their homes through its cameras. The company admits that much of the work will be done by teleoperators in the beginning, with owners scheduling when operators can take over Neo through an app.

If you were hiring a nanny, would you not interview them, check their background and references? Well, what about the person operating Neo? Don’t you need to interview them too? This isn’t Alexa, this isn’t a camera. This is a human embodying a humanoid that can move freely around your home. Let’s say you did interview the operator and decide for yourself that you can trust them in your home with your possessions, pets and family. Then are you happy for the robot to power down when they finish their shift. Or will you interview everyone on the roster, every new hire?

1X is asking for a level of trust that we are not yet set up to deal with.

The company says it will blur faces and allow no-go zones, but the fundamental reality remains: inviting Neo into your home means inviting humans you have never met to see everything, touch anything. Full autonomy is not only a long way off, it might be impossible unless you want to redesign and rebuild a home from scratch to suit humanoid robots.

Why it doesn't matter that humanoid robots aren’t ready for the home

The limitations are real, but so is the arithmetic.Which means none of this will stop adoption. The generation now entering old age can barely afford care at current prices. The generation behind them has less wealth than their parents, and will inherit less still if those assets get spent down on residential fees before they can be passed on. In many countries human care is a luxury, not a baseline.

This is often framed as a demographic problem: ageing populations, shrinking workforces, too few hands to meet the need. But that framing obscures our choice. There is no global shortage of people willing to do care work. There is a political unwillingness to let them in and a market unwillingness to pay them properly. Humanoid robots let us sidestep both. They turn a distributional problem into a technology purchase and call it innovation.

The robot won't be chosen because it works well. It will be chosen because it lets us avoid the harder conversation about what care is worth and who should provide it.

What we are inviting into our homes

We need the foresight now to understand what we are inviting into our homes. Not just machines, but the humans operating them, the companies controlling them, and the structural dependencies we are creating.

Humanoid robots in the home will not mark the arrival of perfected autonomy. They will mark the point at which imperfect autonomy, operated by unseen humans, becomes economically unavoidable. And perhaps that point has already passed.

Understanding what we are accepting matters more than being ready for it. The economics have already made that decision for us.

References:

1X Neo Leasing & Pre-orders Yeo, A. (2025) 1X has launched NEO, a humanoid household robot. Here's how to preorder, Mashable. Available at: https://mashable.com/article/1x-neo-humanoid-robot-preorder

Confirms the $499/month lease model, the $20,000 purchase price, and the 2026 shipping date.

Unitree G1 Price & Availability Unitree Robotics (2024) Unitree G1 Humanoid Robot, Unitree Shop. Available at: https://shop.unitree.com/products/unitree-g1

Verifies the $13,500 consumer price point for the G1 robot.

China International Import Expo (CIIE) Visitor Numbers Xinhua (2025) Breaking records reveals 8th CIIE success, China International Import Expo Official Site. Available at: https://www.ciie.org/zbh/en/news/exhibition/news/20251114/54606.html

Confirms the "900,000 visitors" figure (specifically 922,000 visits) and the prominent placement of robotics.

UK Care Home Costs Carehome.co.uk (2024) Care home costs 2024: How much do you pay?, Carehome.co.uk. Available at: https://www.carehome.co.uk/advice/care-home-fees-and-costs-how-much-do-you-pay

Supports the figure of £5,000+ per month (citing average nursing care costs of over £1,500 per week / £6,000 per month).

Amazon "Just Walk Out" Human Reviewers Singh, R. (2024) Amazon's 'just walk out' checkout tech was powered by 1,000 Indian workers, Business Standard. Available at: https://www.business-standard.com/companies/news/amazon-s-just-walk-out-checkout-tech-was-powered-by-1-000-indian-workers-124040400463_1.html

1X CEO on Teleoperation & Privacy Humanoids Daily (2024) 1X CEO Details NEO's 'Two Modes' and Defends Teleoperation, Humanoids Daily. Available at: https://www.humanoidsdaily.com/feed/1x-ceo-details-neo-s-two-modes-and-defends-teleoperation-as-more-secure-than-a-cleaner Supports the quotes regarding human teleoperators seeing into homes and the "nanny" comparison.

The Ninety-Ninety Rule Bentley, J. (1985) 'Bumper-Sticker Computer Science', Communications of the ACM, 28(9), pp. 896–901. Available at: https://dl.acm.org/doi/10.1145/4284.315122 The original academic source attributing the adage to Tom Cargill at Bell Labs.