Executive summary
For decades, robots have been confined to factories and warehouses – purpose-built machines performing tightly scoped tasks in controlled environments. Now, advances in AI, cheaper hardware, and billions in venture capital are converging to make a once-distant goal feel suddenly achievable: a humanoid robot that can navigate the unpredictability of everyday domestic life.
- Tesla’s Optimus learns to perform household tasks by watching human videos and can be instructed via natural language commands.
- Figure AI’s Figure 03 features tactile sensors sensitive enough to detect as little as three grams of pressure.
- 1X’s NEO is the first domestic humanoid available for pre-order at US$20,000, with US deliveries expected in 2026.
- Morgan Stanley forecasts that 10% of US households could own a humanoid robot by 2035, equating to roughly 15 million units.
- Safety, regulation, insurance, and consumer acceptance remain major hurdles that could slow deployment significantly.
Despite the hype, domestic robots remain far from ready for unsupervised daily use. But the trajectory is hard to ignore. With global populations ageing and workforces set to shrink, humanoid robots could help sustain economic growth, drive down the cost of goods and services, and give people a way to age independently at home. The question is no longer whether this technology will arrive but when – and on whose terms.
Over the years, we’ve gotten pretty good at building robots for specific jobs. Autonomous shelving units now glide through Amazon’s vast warehouses, ferrying packages with mechanical precision. Robotic arms weld, paint, and bolt together cars across the auto industry. And in parts of China, some factories even operate with the lights out, running around the clock without a single human on the floor. Robots, in other words, are already deeply woven into the fabric of the modern economy. Yet for all their capability, most of these machines look nothing like us. Each one was designed to do a narrowly defined job in a highly predictable environment, and most of them would be completely lost the moment something unexpected happened. A box in the wrong place, a door left ajar, a staircase – and everything falls apart.
Getting a machine to handle the unpredictabilities of everyday life – to pick up something unfamiliar, navigate a cluttered room, or interpret an ambiguous instruction – requires solving a cluster of engineering and AI challenges that have stumped researchers for decades. The hardware was too expensive, the AI too limited, and the real world too unpredictable. But things are finally picking up. A confluence of advancements in AI, sensor technology, and a flood of venture capital has converged to make the previously impractical suddenly feel within reach. Dozens of companies are now competing to build a humanoid robot capable of functioning in a domestic setting. In this article, we’re going to look at where the domestic robotics industry actually stands, what’s driving this latest wave of innovation, and when a humanoid robot might realistically find its way into your living room.
In the next 10 years – maybe under 10 years – the biggest company in the world will be a humanoid robot company.
Brett Adcock, Figure AI CEO
Say goodbye to chores
A robot that cleans your home, takes out the trash, and cooks lunch. Where do we sign up?
As we noted above, numerous companies are now actively working on developing humanoid robots for household use. Among those that have garnered the most attention so far is Tesla’s Optimus, a project that has steadily evolved from a tongue-in-cheek stage reveal – where a dancer in a bodysuit stood in for the actual robot – into something that increasingly demands to be taken seriously. The latest iteration of Optimus runs on the same neural network architecture that powers Tesla’s self-driving vehicles, with xAI’s Grok large language model layered on top to handle conversational interaction. In practical terms, that means the robot can process natural language commands – spoken or typed – rather than relying on pre-programmed routines. It walks at speeds of up to eight kilometres per hour and can carry around 11 kilograms, enough to haul a bag of groceries or a small box of tools.
A video released by Tesla in mid-2025 offered the most detailed look yet at what Optimus can do in a domestic setting. The footage shows the robot working through a range of household tasks with smooth, deliberate precision: taking out the trash, wiping down a surface, vacuuming, stirring food on a stove, tearing paper towels from a roll, and closing curtains. Other clips have shown Optimus taking a person’s blood pressure, carrying boxes down a flight of stairs, and assembling pizza. All of it, Tesla claims, was handled by a single neural network trained not through manual programming or teleoperation but by watching videos of humans performing the same tasks.
According to the company, the next goal is for Optimus to learn directly from third-person footage found across the internet, a method that could dramatically accelerate the rate at which the robot picks up new skills. When Musk first presented Optimus back in 2021, he made some rather lofty promises, claiming that the robot would babysit children, run errands, and serve as a personal companion. While we’re not there yet, recent demonstrations have made that kind of future seem closer than it once did. However, with the price reportedly set between US$20,000 and US$30,000, Optimus may be out of reach for the average household, though Musk has floated the idea of rental options further down the line.
Safety first
Figure 03 learns to perform new tasks by watching humans do them first.
Another strong contender is Figure 03, the third-generation humanoid from California-based Figure AI. Built as a ground-up redesign of its predecessor, Figure 03 runs on the company’s proprietary Helix system, a vision-language-action AI model that allows the robot to perceive its surroundings, interpret spoken or written commands, and carry out tasks in real time without following pre-written scripts. In effect, Helix lets the robot learn by watching humans do things, then figure out how to replicate those actions on its own.
Because a robot working alongside people in their homes raises obvious safety concerns, Figure AI has rethought the physical design accordingly. Hard machined surfaces have been replaced with washable soft textiles and strategically placed multi-density foam to cushion any accidental contact. The overall frame is 9% lighter and more compact than Figure 02, making it easier to manoeuvre in the tight spaces that define most real homes, such as narrow hallways or cluttered kitchens.
Beyond the AI, Figure 03 packs a dense array of hardware upgrades. A redesigned sensor suite gives each camera a 60% wider field of view at double the frame rate, feeding the onboard processor a richer, faster stream of visual data for navigating complex real-world environments. The hands are equipped with custom tactile sensors capable of detecting forces as light as three grams – roughly the weight of a paperclip resting on your fingertip – along with embedded palm cameras that provide close-range visual feedback during grasping tasks. An upgraded auditory system, with a larger speaker and repositioned microphone, gives the robot sharper sound recognition and clearer voice interaction.
And for those who care about aesthetics, Figure 03’s soft exterior panels are removable and interchangeable, meaning the robot can be dressed to suit different settings. When it’s time to recharge, each unit steps onto a wireless charging pad and simultaneously offloads data at high speed, keeping the fleet continuously learning from every task performed. The robot moves with a slow, deliberate grace through kitchens and living rooms, folding towels, loading dishwashers, clearing tables, and operating household appliances. The way it interacts with people is equally striking – measured, unhurried, and smooth enough that it doesn’t trigger the instinctive unease most of us feel around large autonomous machines.
Your friendly robotic helper
NEO is the first commercially available domestic humanoid robot – and it can be yours for US$20,000.
Both Optimus and Figure 03 are undoubtedly impressive, but neither is commercially available yet. One domestic humanoid that is – or will be very soon – is NEO, built by 1X, a Norwegian-founded robotics firm now headquartered in Silicon Valley. Made available for pre-order in October 2025, NEO is priced at US$20,000 with priority delivery slated for 2026, or available through a US$499-per-month subscription. It comes in three colours (tan, grey, dark brown) and ships first to US customers, with other markets following in 2027. NEO’s Scandinavian roots show in its minimalist aesthetic: a soft, head-to-toe body sheathed in custom 3D lattice polymer, camera-equipped eyes behind a dark visor on a human-shaped head, and multi-jointed hands driven by elastic motors that mimic the way human tendons work. The goal is a machine that feels like it belongs in a living room rather than a lab.
NEO weighs just 30 kilograms but can lift over 68 kilograms and carry 25 kilograms. The hands now have 22 degrees of freedom, close to the human hand’s 27. A three-stage speaker system and four-microphone array handle voice interaction, while built-in Wi-Fi, Bluetooth, and 5G connectivity keep the robot linked to 1X’s network for continuous learning and over-the-air software updates. Operating noise sits at 22 decibels – quieter, in fact, than a refrigerator. On the AI side, NEO uses what 1X calls its World Model as the robot’s cognitive core, enabling it to generalise to tasks it hasn’t been specifically trained on. It can recognise when it’s being spoken to, use visual context to enhance its responses – spotting ingredients on a kitchen counter and suggesting what to cook, for instance – and remember past conversations to adapt over time.
From day one, 1X says NEO will be able to handle tasks like opening doors for guests, fetching items, and turning off the lights at night, with capabilities growing through regular software updates. Critically, though, the robot isn’t fully autonomous yet. When NEO encounters something it can’t handle on its own, it falls back on an “Expert Mode” service where human teleoperators step in remotely to assist. It’s a hybrid model that doubles as a training mechanism, feeding real-world data back into the system so that future NEOs get smarter with every interaction.
“For humanoid robots to truly integrate into everyday life, they must be developed alongside humans, not in isolation,” says 1X CEO Bernt Børnich. “The home provides real-world context and the diversity of data needed for humanoids to grow in intelligence and autonomy. It also teaches them the nuances of human life – how to open the door for the elderly, move carefully around pets, or adapt to the unpredictability of the surrounding world. Robots confined to industrial space or lab development miss out on this critical understanding.”
I believe we’re coming close to the day when every home that can afford it will have a humanoid robot.
Adrian Cheok, a Professor at the Nanjing University of Information Science and Technology
Coming to a home near you (or not?)
While some industry experts believe that humanoid robots are the future, others are concerned about safety and the technology’s impact on human labour.
So, when can we actually expect humanoid robots to become a common feature in our homes? Ask around, and you’ll get wildly different answers. Unsurprisingly, the people building the technology are the most bullish about its prospects. “In the next 10 years – maybe under 10 years – the biggest company in the world will be a humanoid robot company,” predicts Figure AI CEO Brett Adcock. “Every home will have a humanoid, which will do domestic chores from emptying the dishwasher to making the bed.” Adcock goes even further, suggesting that billions of humanoid robots will eventually fill roles across the workforce, healthcare, and even space, where they will be helping us build off-world colonies.
Claims like that obviously need to be weighed against the fact that people like Adcock have staked their careers and considerable fortunes on the technology succeeding. Still, they’re not entirely without outside support. Morgan Stanley’s most recent forecast projects that around 10% of US households could own a humanoid robot by 2035 – roughly 15 million units – with adoption accelerating sharply through the late 2030s and 2040s. Adrian Cheok, a Professor at the School of Automation at the Nanjing University of Information Science and Technology, is similarly optimistic: “I believe we’re coming close to the day when every home that can afford it will have a humanoid robot. They may be a bit expensive at the start, but they’ll rapidly come down in price to create a mass market through the sheer utility of doing the chores.”
Not everyone shares their enthusiasm, though. Serious questions remain about the practicality and affordability of the technology, as well as who will actually benefit from it – and who might get left behind. “This technology has a tremendous potential to provide value, and provide good, but if it just makes large corporations richer, that’s not going to be a good outcome,” says Hans Peter Brondmo, a former vice president at Google’s Everyday Robot project. “I believe this is less of a technological challenge, and more of a policy challenge. We need to fundamentally rethink the social contract.”
And then there’s the question that tends to surface whenever the conversation moves from factory floors to family kitchens: safety. “If they’re cooking, at what point are you comfortable with robots carrying a knife or pan of boiling water?” asks applied futurist Tom Cheesewright. “These robots are also physically strong and are capable of doing you harm, so there’ll undoubtedly be tabloid stories about people coming a cropper.” Cheesewright draws a parallel to the self-driving car debate. “People underestimate the scale of these challenges. Many of the technical issues have been solved, but the challenges are now economic and social, which includes the need for regulatory frameworks, insurance, and sheer consumer acceptance.”
Closing thoughts
Domestic robots clearly have a long way to go. The demos are impressive, but the gap between a controlled showcase and a machine that can reliably navigate your home day after day without human supervision remains wide. Yet if the progress in robotics follows anything close to the trajectory of the broader AI industry over the past few years, that distance could close faster than most people expect. And if even a fraction of these companies deliver on what they’re promising, the downstream effects would fundamentally reshape how the world works.
The global population is expected to peak this century before tipping into decline, and with it, the supply of human labour. Robots capable of picking up the slack could allow economies to keep growing even as workforces shrink. The cost of goods and services could fall dramatically, potentially lifting the quality of life across income levels. And as populations skew older, domestic robots might offer something harder to quantify but no less important: the ability for people to age at home, with dignity and independence, long after they might otherwise have needed full-time human care. None of that is guaranteed, of course. But the money, the talent, and the momentum all point in one direction. The robots are coming. The only real questions left are how fast and whether we’ll be ready for what happens when they do.