Humanoid robots at home: for when?
Every funding round, every investor, and who is really ahead in humanoid robots, in one document. Have a look ↓
Summary
The first humanoid robots are arriving in homes in limited numbers in 2026–2027, but reliable household assistants will probably become more common toward 2030. Affordable robots capable of independently handling most chores remain a prospect for the 2030s or later.
Consumers can already preorder domestic robots. 1X offers NEO for $20,000 or $499 per month, while Weave Robotics prices Isaac 1 at $7,999 or $449 per month. Neither has established a proven, large-scale consumer business, and both may still rely on remote human assistance.
The first serious household robots may look quite different from the humanoids people imagine. Weave and Sunday Robotics have chosen wheeled designs, avoiding some of the complexity of balancing on two legs. Getting laundry done reliably may prove more valuable than making a robot walk like a person.
Commercial adoption is still tiny compared with established domestic robotics. Approximately 7,000 full-size humanoids were sold worldwide in 2025 for professional and industrial applications, against 34.2 million consumer service robots. The existing consumer market is roughly 4,900 times larger, and overwhelmingly built around specialized machines.
Figure provides one of the clearest measurements of actual household progress. Its Helix 2.5 system completed 237 of 420 full-task attempts across 30 unfamiliar homes, achieving 56% success compared with a 9% experimental baseline. That is a substantial technical advance, but a household assistant that fails nearly half its assignments is still difficult to depend on.
Household tasks also vary more than demonstrations suggest. Figure achieved approximately 67% success at making beds and 62% at folding towels, but only 40% when tidying toys. General-purpose intelligence in the home is proving much harder than mastering one repeatable chore.
Tesla's strongest advantage may ultimately be manufacturing rather than household AI. Optimus is entering early production, and Tesla is planning infrastructure capable of supporting very large volumes. But its targeted external commercial deliveries around late 2027 remain far removed from evidence of a dependable mass-market housekeeper.
Early domestic robotics will probably depend on a hybrid operating model: AI handles familiar activities while remote humans intervene when necessary. This could accelerate useful deployments, but it introduces ongoing support costs and a potentially uncomfortable privacy question for households with cameras inside their homes.
Safety is another major obstacle. Softer materials, protective components, and wheeled movement can reduce specific risks, but manufacturers have not established extensive independent evidence that general-purpose home robots can operate safely without supervision around young children and pets.
The most plausible path to adoption is gradual specialization before genuine versatility. Robots that reliably handle a few repetitive chores could find paying customers considerably sooner than machines capable of cooking, cleaning, organizing, and adapting independently to an entire household.
The decisive milestone will be the amount of useful work a robot completes without intervention, not how human it looks or how impressive its best demonstration appears. Prices, task-completion rates, teleoperation frequency, safety performance, and actual deliveries will tell us when home humanoids have truly arrived.
Written by the New Market Pitch teamWe track new markets so founders and investors can move faster. About us
We keep an up-to-date report on humanoid robots. Have a look ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Can you actually buy a humanoid robot for your home today?
You can now preorder a humanoid robot for your home, but there is still no established consumer market where ordinary families can buy a proven, fully autonomous robot housekeeper.
Two companies have made the clearest move toward real customers. 1X offers its NEO home robot for $20,000 or $499 per month, with initial U.S. deliveries targeted for 2026. Weave Robotics offers Isaac 1 for $7,999 or $449 per month, with California customers first in line. Both companies accept refundable deposits.
Meanwhile, Tesla is targeting external commercial deliveries of Optimus around the end of 2027, and Figure is testing increasingly advanced household skills. Neither has announced a comparable consumer ordering program.
These products are at very different stages. Unitree already sells programmable humanoids, but buyers largely get robotics platforms rather than ready-to-use housekeepers. NEO and Isaac 1 promise practical home assistance, although both can still depend on remote human operators.
The realistic timeline starts with limited deliveries in 2026–2027. Useful household assistants could become more common toward 2030, while affordable robots capable of independently handling most chores are more likely to arrive in the 2030s or later. The dates beyond announced deliveries are our projections, based on current technical results, prices, and manufacturing progress.
If you want more recent data on this point, see the latest figures in our Humanoid Robots report ↓.
Why are humanoid robots suddenly getting closer to our homes?
Humanoid robots are getting closer to home use because AI systems are finally learning to handle physical tasks across different environments, although their reliability is still far behind what consumers expect.
For years, industrial robots could repeat specific movements with exceptional precision but struggled when objects or surroundings changed. Homes are full of those changes: clothes move and fold unpredictably, dishes come in different shapes, and furniture varies from one room to another.
Recent advances in vision-language-action models are helping robots connect visual information, instructions, and movement. Figure's Helix systems provide one example, while 1X's Redwood AI is being developed to learn and repeat household tasks.
The progress is becoming measurable. Figure's Helix 2.5 demonstrated three household behaviors across 30 unfamiliar homes in September 2026. The company reported that its broader pretraining approach raised complete-task success from 9% to 56% under its testing conditions.
That is a major improvement. The catch is that the system still failed 44% of evaluated attempts.
Hardware has been improving alongside the software. Figure 03 introduced changes aimed at domestic use, while 1X developed NEO around a softer, lighter design. Weave took a different route by using wheels instead of legs to simplify movement and stability.
For the first time, several companies are trying to solve the full household problem rather than concentrating exclusively on walking, manipulation, or AI in isolation.
These pages are excerpts from our report. See the rest ↓
Sample pages from an earlier edition. The report has been updated with new data since.
How many humanoid robots are actually being sold today?
Humanoid robot sales are growing from a very small base, and almost all confirmed commercial activity still takes place outside ordinary homes.
The International Federation of Robotics reported that approximately 7,000 full-size humanoids were sold worldwide in 2025 for industrial and professional applications.
That sounds substantial until we compare it with established consumer robots. According to the IFR's World Robotics 2026 report, consumer service robot sales reached approximately 34.2 million units in 2025, an increase of 37% from the previous year.
The comparison works out to roughly 4,900 consumer service robots for every full-size humanoid sold.
These categories are different, so the ratio should be read as a measure of commercial maturity rather than direct competition. Most consumer robots are specialized machines such as robotic vacuum cleaners. Humanoids are primarily entering factories, logistics sites, and professional environments.
The IFR also reported nearly 250,000 professional service robots sold in 2025, up 24%. About 117,500 were used in transportation and logistics, a category where moving goods between known locations offers a clearer commercial opportunity than general household work.
The striking thing is that domestic robots already sell in the tens of millions without looking remotely human. Consumers have clearly accepted robots when they perform a useful job reliably.
Humanoid manufacturers now have to prove that a more versatile machine is worth the additional cost.
If you want more recent data on this point, see the latest figures in our Humanoid Robots report ↓.
Which humanoid robot companies are closest to delivering a home assistant?
1X and Weave Robotics currently have the most concrete consumer offers, while Figure leads some of the most ambitious public household manipulation experiments and Tesla is concentrating on manufacturing Optimus.
The companies are approaching the market differently.
1X wants NEO to become a general household assistant capable of learning new chores. Weave is starting with a narrower promise: laundry handling and daily tidying. Sunday Robotics is preparing residential testing of Memo, another wheeled domestic assistant.
Tesla and Figure have wider ambitions. Both want robots that can eventually perform many different jobs, including industrial work and domestic tasks, but neither has established a broad consumer delivery business.
Weave is trying to commercialize a smaller set of useful chores sooner, while Figure is tackling the harder problem of generalizing physical skills across different homes. Those are quite different bets on how household robotics will develop.
| Company and robot | Main approach | Price or cost | Current commercial status |
|---|---|---|---|
| 1X: NEO | Full-size domestic humanoid | $20,000 or $499/month | U.S. delivery target in 2026 |
| Weave: Isaac 1 | Wheeled laundry and tidying assistant | $7,999 or $449/month | California shipments targeted for fall 2026 |
| Sunday Robotics: Memo | Wheeled household assistant | Approximately $20,000 reported handmade build cost | Residential beta testing planned |
| Tesla: Optimus | General-purpose bipedal humanoid | No confirmed consumer price | Early production; external commercial deliveries targeted around late 2027 |
| Figure: Figure 03 | General-purpose humanoid | No announced consumer price | Household research and industrial development |
| Unitree: G1 | Programmable humanoid platform | Starting around $16,000 | Commercially offered, but not a turnkey housekeeper |
| Agility Robotics: Digit | Industrial logistics humanoid | Enterprise pricing | Workplace deployments, no consumer home offer |
Want the whole picture? Check our report ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Is Tesla Optimus really coming to people's homes soon?
Tesla Optimus is finally entering early production, but ordinary consumers should not expect a dependable robot housekeeper from Tesla in the immediate future.
Tesla's latest quarterly update confirms that early production-quality Optimus units are being built at its Fremont facility. The company is also establishing an Optimus Academy to generate training data and develop robot skills.
Elon Musk has reaffirmed a target for external commercial deliveries around the end of 2027. The company has discussed a Fremont production line designed for eventual annual capacity of approximately one million robots.
That capacity figure is particularly easy to misunderstand. It describes what the production infrastructure is intended to support, rather than robots currently leaving the factory.
Tesla still needs to improve the robot's hands, reliability, durability, and safety. Those are major engineering challenges for a machine expected to handle fragile objects and move around people.
The company's record also gives us a reason to be cautious about dates. Musk previously suggested that Tesla could build thousands of Optimus robots in 2025, yet the company did not demonstrate commercial deployment at that scale. Its third-generation robot has also faced delays.
The near-term opportunity looks stronger in workplaces, where tasks can be controlled and repeated.
Tesla could eventually manufacture humanoids on an enormous scale, especially given its experience with motors, batteries, electronics, and automotive production. For now, the evidence supports a serious development program with early production underway. A mass-market consumer robot business is still some distance away.
If you want more recent data on this point, see the latest figures in our Humanoid Robots report ↓.
Can Figure's humanoid robots actually do housework?
Figure's humanoid robots can already perform selected household chores autonomously, but the latest published test results show that they are still too unreliable to replace human housework.
Figure introduced Helix 2.5 in September 2026 and tested it in 30 homes around the San Francisco Bay Area that had not been used to collect task-training data.
The robot performed three types of work: tidying toys, folding towels, and making beds. Figure reported 237 successful complete-task attempts out of 420, approximately 56%.
The individual results reveal much more than the headline figure.
Making beds achieved approximately 67% success, towel folding reached 62%, and tidying toys managed 40%. These figures were reported under Figure's own evaluation method, with complete-task failures receiving no partial credit.
That means the robot performed substantially better on some chores than others, even after learning from a large dataset of human behavior.
The progress becomes clearer when compared with the earlier experimental baseline. Figure reported only 9% success without its Index pretraining approach. The improvement to 56% represents a gain of 47 percentage points.
Figure also says the system needed half as much task-specific data as a representative earlier Helix 02 behavior while transferring its skills across 30 unfamiliar homes.
The catch is that the evaluation remains company-reported, and the published results do not establish how the robot would perform over months of ordinary household use.
At 56% complete-task success, a robot would still need substantial improvement before consumers could depend on it without checking its work. Nevertheless, the experiment gives us unusually concrete evidence that household manipulation is improving.
Prefer verified numbers to headlines? Our report is here ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Will home humanoid robots really work without a human controlling them?
The first home humanoid robots will be partly autonomous, with remote human operators stepping in when the machines cannot complete a task themselves.
1X openly describes this model for NEO. The robot arrives with basic autonomy, while customers can schedule Expert Mode sessions for more complex tasks. A remote operator can supervise or control the robot and help it learn new abilities.
Weave uses a similar arrangement for Isaac 1. Its website states that the robot is autonomous by default for Laundry Flow and Daily Reset, with teleoperation assistance available when necessary.
This changes how we should interpret early home-robot demonstrations. A successful household task might involve independent AI control, remote assistance, or a mixture of both.
There is nothing inherently wrong with supervised operation. Human intervention can make an immature technology more useful while its autonomous abilities improve.
But we need to know how much intervention customers actually require.
If a robot needs help once every few weeks, remote support could be perfectly acceptable. If someone must intervene several times during each household chore, the economics and convenience become much less convincing.
Neither company has published a comprehensive, independently audited breakdown showing the percentage of household tasks completed entirely autonomously across its consumer fleet.
Remote operation also creates a privacy trade-off. A robot equipped with cameras may need to share visual information from inside bedrooms, living rooms, or other private spaces with a human operator. Manufacturers can introduce consent controls and access restrictions, but customers still need to decide whether that arrangement is acceptable.
For now, teleoperation is an important part of how early domestic robotics can work commercially. The real progress to watch will be how quickly manufacturers reduce their dependence on it.
If you want more recent data on this point, see the latest figures in our Humanoid Robots report ↓.
Are humanoid robots safe enough to live with children and pets?
Today's home humanoids are still too new to be considered proven, unsupervised companions around children and pets, even when manufacturers design them specifically for household safety.
A robot operating inside a house encounters risks that are uncommon in controlled factories. Children may approach it unexpectedly, pets can move beneath its legs or wheels, and people can leave fragile objects in its path.
1X has emphasized NEO's soft exterior and safety-focused mechanical design. Figure 03 introduced protective materials and battery-related improvements. Weave's wheeled approach also reduces the balancing challenges associated with walking on two legs.
These choices may reduce particular risks, but the robot must still manage unexpected arm movements, collisions, pinching hazards, battery faults, and mistakes while handling objects.
ISO 13482, the international standard covering personal-care robot safety, provides an established framework for evaluating relevant risks. Compliance with applicable standards and independent testing become especially important when manufacturers begin putting robots into family homes.
1X's own safety information warns customers to remain attentive around children, pets, and vulnerable people.
That warning matters more than promotional videos of friendly-looking robots. No manufacturer has yet published extensive independent evidence showing that a general-purpose humanoid can safely handle unrestricted daily household activity around young children.
Parents should therefore treat early home robots as machines requiring supervision, rather than substitutes for childcare.
If you want more recent data on this point, see the latest figures in our Humanoid Robots report ↓.
Want to know who is really ahead? Our report is here ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Our methodology
To assess when humanoid robots will realistically arrive in people's homes, we broke the question into six dimensions: consumer availability, household task performance, autonomous operation, manufacturing readiness, commercial adoption, and safety. Each addresses a different obstacle between an impressive robotics demonstration and a machine that families can genuinely use.
For each dimension, we examined the freshest available evidence, combining published prices and ordering conditions, delivery announcements, technical evaluations, production updates, industry sales statistics, and relevant safety standards. We gave greater weight to measurable results and documented commercial activity than to promotional videos, proposed production capacity, or ambitious company timelines. This allowed us to assess where the industry is advancing rapidly and where progress remains difficult to demonstrate.
We paid particular attention to how the evidence fits together. A robot completing a household task successfully tells us something about its intelligence, but considerably less about everyday usefulness until we know its failure rate, dependence on human intervention, and ability to repeat the same work in unfamiliar homes. We treated company-reported experiments as valuable technical evidence while distinguishing them from independent verification and long-term consumer performance.
We also separated three milestones that are often confused: the first robots becoming available to customers, robots becoming consistently useful for a limited range of chores, and affordable machines handling most household tasks independently. Announced delivery dates were kept distinct from our own adoption projections. We then assessed the combined evidence across these dimensions to form a timeline that reflects both the industry's progress and its remaining constraints.
Key sources include 1X's official NEO ordering and pricing information, 1X's NEO product announcement, Weave Robotics' Isaac 1 product details, Weave's preorder and delivery information, Sunday Robotics' residential testing plans, and Unitree's commercial G1 offering.
For technical performance and manufacturing readiness, we used Figure's Helix 2.5 evaluation across 30 homes, Figure's hardware development disclosures, Tesla's July 2026 quarterly update, Tesla's January 2026 quarterly update, and Agility Robotics' commercial deployment announcement.
Broader market adoption and safety assessments draw on the International Federation of Robotics' World Robotics 2026 figures, its previous service robot market benchmark, ISO 13482:2014 for personal-care robot safety, ISO/TR 23482-1:2020 on safety-related test methods, and the updated ISO 13482 standards project. Taken together, these sources provide a basis for comparing technological capability, commercial maturity, and practical household readiness without relying on any single manufacturer's roadmap.
Written by the New Market Pitch teamWe track new markets so founders and investors can move faster. About us
