What Home Robots Can Reliably Do, and What Still Needs a Human
A task-by-task audit of home robots in 2026, separating reliable products, supervised work, controlled demonstrations, teleoperation and experiments today.
Introduction
A home robot can already vacuum a floor for an hour without an operator. The same standard does not yet apply to clearing a mixed dinner table, sorting laundry or loading an unfamiliar dishwasher. In 2026, the strongest general-manipulation demonstrations include Sunday Memo handling dishes and socks, Figure 03 completing an uncut dishwasher sequence, Physical Intelligence policies folding laundry and 1X NEO performing household chores with a combination of autonomy and Expert Mode. Each result uses a different body, home, speed and intervention policy.
This audit counts only tasks shown in a real home or a credible domestic setup. It separates specialized products that work reliably today from supervised systems, controlled demonstrations, teleoperated sequences and experiments. Company labels are retained, but a video is not accepted as proof of everyday reliability unless the source reports repetition, duration, success rate or field use. Price and availability are also separated: a beta application, deposit, subscription and normal retail shipment are four different commercial states.
Key findings
- Specialized floor cleaning is reliable today; general household manipulation is not.
- Laundry folding has moved beyond one-off clips, but current systems remain slow, task-specific or limited to selected households.
- Dishwasher loading is demonstrated by Memo and Figure, yet neither source provides a standardized household success rate across many kitchens.
- 1X NEO is the clearest purchasable full-body home proposition, but its offer explicitly includes human Expert Mode.
- Elder care remains experimental because physical safety, clinical validation, privacy and emergency reliability are not established.
What home robots can do in 2026
Classification concerns the cited task, not every capability of the robot.
| Task and system | Environment and evidence | Autonomy class | Availability and price | Practical limitation |
|---|---|---|---|---|
| Floor vacuuming and mopping; specialized mobile cleaners | Millions of homes; routine navigation, docking and scheduled cleaning are established product functions. | Autonomous for the bounded task | Widely sold at consumer prices | Corners, stairs, cables, liquids and object pickup still require preparation or human help. |
| Laundry folding; Weave Isaac 0 | Limited California household service; Weave reports more than 2,000 field hours and over 1,000 pounds of laundry per week. | Supervised autonomous | $249 per month with no upfront payment, or $3,999 plus $49 per month; regional program | Stationary appliance dedicated mainly to laundry; service area and garment range are limited. |
| Laundry and mobile chores; Weave Isaac 1 | Company demonstrations and fall-2026 California reservation program. | Supervised autonomous; intervention policy not fully published | $449 per month or $7,999 upfront; $250 refundable reservation | Shipping schedule, generalization and real-home success rates remain unverified. |
| Dishes, espresso and socks; Sunday Memo | Company-published uncut and accelerated clips in homes, plus training-data collection through gloves. | Company labels execution autonomous; independent evidence unavailable | Free late-2026 beta application; no retail sale or final price | Reliability is still being improved; beta households are part of product development. |
| Dishwasher loading; Figure 03 with Helix 02 | Company says a four-minute sequence was autonomous and uncut. | Autonomous, according to Figure | Not sold to consumers | One curated kitchen and company-produced evidence do not establish broad household reliability. |
| Bedroom tidying; Figure 03 with Helix 02 | Company demonstration of object pickup and placement in a furnished room. | Autonomous, according to Figure | Not sold to consumers | Object set, intervention log and repeated success rate are not independently published. |
| General chores; 1X NEO | Home product page and demonstrations; foundational autonomy plus remote Expert Mode. | Human-in-the-loop with autonomous skills | $20,000 Early Access or $499 per month; $200 refundable deposit, US delivery plan | Experts may need to help with unfamiliar chores; privacy and service terms matter. |
| Laundry and kitchen manipulation; Physical Intelligence policies | Research videos and uncut demonstrations on robot platforms, including operation in new homes. | Autonomous research execution after training | Model research, not a consumer robot | Policy performance depends on hardware, scene setup and task distribution. |
| Assistance and object retrieval; Toyota HSR | Research and demonstration programs in human environments. | Teleoperated, supervised autonomous or research policy depending on project | Research platform, not mass-market home product | No single current consumer package or universal autonomy level. |
Task classification by current evidence
The same task can move between categories depending on the robot and environment.
| Band | Examples | Evidence required |
|---|---|---|
| Reliable today | Bounded floor cleaning, scheduled patrol and automatic docking by specialized products | Repeated consumer use, product support and minimal operator intervention |
| Possible under supervision | Laundry handling, object retrieval, simple transport and selected door operations | Field use with monitoring, recovery support or a restricted object set |
| Controlled demonstration | Dishwasher loading, table clearing, sock folding and simple food-machine operation | A successful company or laboratory sequence in a prepared or familiar environment |
| Teleoperated | Remote manipulation, training demonstrations and rescue from an unfamiliar state | Human motion, joystick, VR or remote expert directly influences execution |
| Experimental | General cooking, mixed-clutter cleanup, shower or toileting assistance | Early research without sufficient safety, repeatability or product evidence |
| Not yet reliable | Unsupervised whole-home housekeeping and physical elder care | No public system meets broad task coverage, safety and recovery requirements |
Floor cleaning is the benchmark for household autonomy
Robot vacuums succeed because the task is narrow: move over a mostly horizontal floor, avoid obstacles, collect debris and return to a dock. The environment can be mapped and the tool stays near the ground. Even here, users still remove cords, unblock brushes and handle stairs. The product is autonomous for a bounded workflow, not a general household agent.
General-purpose robots inherit every difficult part that floor cleaners avoid. They must identify objects, select a grasp, estimate fragility, carry items at human worktop height, open mechanisms and recover after slips. A home also changes between runs. Reliability therefore drops as soon as the robot must manipulate unstructured objects rather than navigate around them.
Laundry is becoming a real service, but not a solved skill
Weave’s Isaac 0 is the strongest field-use claim in this sample because the company publishes cumulative hours and laundry volume and offers a limited California service. Its stationary architecture removes walking, doorways and whole-home navigation, narrowing the problem to handling garments around one appliance. That design choice is a practical advantage, not a weakness.
Physical Intelligence and other laboratories show autonomous folding policies across garments, while Sunday Memo demonstrates piles of socks. These clips establish capability. They do not establish that every fabric, inside-out sleeve, wet garment or tangled load will finish without assistance. A useful home product needs recovery and service support as much as a successful grasp policy.
Dishwashers expose the gap between a demo and a product
Figure’s Helix 02 dishwasher sequence is valuable because the company describes it as autonomous and uncut. Sunday also publishes Memo sequences for clearing dishes, dumping scraps and loading a machine. Both show multi-step coordination, fragile-object handling and appliance interaction.
The missing evidence is distribution. Kitchens vary in lighting, counter height, dishwasher racks, plate shape and clutter. Neither company publishes a standardized test across dozens of homes with intervention counts and breakage rates. The correct classification is a demonstrated autonomous task under the company’s conditions, not reliable universal dishwasher loading.
Food preparation remains narrow
Memo can operate an espresso machine in a company demonstration, and laboratory robots can pour or manipulate ingredients. Those are bounded procedures with known tools. General cooking adds heat, sharp objects, contamination, timing and continuous safety monitoring. A robot that can make one espresso has not demonstrated meal preparation.
Food service should be described at the operation level: fetch a sealed drink, press a machine control, pour from a known container or move a plate. Each action needs its own success and safety evidence. No reviewed general home robot publishes broad independent results for cooking complete meals.
Human assistance is part of the product architecture
1X makes the human-in-the-loop layer explicit through Expert Mode. That is more transparent than presenting every successful chore as independent autonomy. A remote expert can recover a task, demonstrate a new one or operate the robot while the system collects useful data. Reflex and other industrial companies use similar supervision logic for uptime.
For a home buyer, the important questions are who can access cameras, when a human takes control, whether consent is per session, where video is processed and what happens if connectivity fails. Autonomy percentage alone does not answer those privacy and service questions.
Elder assistance needs a higher evidence standard
Voice reminders, remote check-ins and fetching light objects are plausible low-contact tasks. Physical support, fall recovery, medication handling, bathing and emergency response carry medical and injury risks. A robot must detect failure, avoid pulling a person off balance and remain available during power or network faults.
None of the general home robots reviewed publishes clinical validation or a broad safety case for unsupervised physical elder care. Such claims should remain experimental until healthcare partners report defined tasks, population, observation period, adverse events and human oversight.
Limitations and missing information
- Company-produced videos may omit failed attempts, setup time and off-camera intervention even when the visible sequence is uncut.
- Task duration and success rates are rarely reported under a common protocol.
- Prices can exclude subscription, remote assistance, delivery, installation, taxes or regional service limits.
- A research policy demonstrated on one robot does not become a consumer product until hardware, support and safety are integrated.
- Specialized appliances and complete mobile humanoids solve different problems and should not be ranked as if they were interchangeable.
Conclusion
A home robot in 2026 can reliably perform narrow, well-engineered jobs. Floor cleaning is established. Weave’s stationary Isaac 0 shows that laundry can become a regional service when mobility is removed from the problem. Memo, Figure 03, NEO and research policies demonstrate pieces of table clearing, dishwasher loading, folding, object retrieval and appliance operation, but their evidence does not yet support unsupervised whole-home housekeeping.
The most honest commercial architecture is supervised. A robot executes known skills, asks for help when confidence falls and records intervention instead of hiding it. NEO’s Expert Mode states that model directly. Buyers should judge each task by repetition, environment diversity, intervention rate, damage risk and service coverage. Under that standard, general cooking, mixed-clutter cleanup and physical elder care remain experimental, while bounded cleaning, laundry and object transport are the first areas moving from video demonstrations toward usable household systems.
Frequently asked questions
Can a home robot clean an entire house in 2026?
Specialized robots can vacuum or mop mapped floors, but no reviewed general-purpose robot reliably handles every room, surface and object without preparation or assistance. Picking up mixed clutter, cleaning bathrooms, moving between floors and handling liquids require manipulation and recovery capabilities that current home platforms have demonstrated only in narrower conditions.
Can a robot fold laundry automatically?
Yes, selected robots and research policies can fold garments autonomously. Weave reports limited real-home operation with Isaac 0, while Physical Intelligence and Sunday publish folding demonstrations. The capability is still constrained by garment type, presentation, speed and recovery. A successful sequence should not be read as universal handling of tangled or highly varied laundry.
Can a humanoid robot load a dishwasher?
Figure 03 and Sunday Memo have company-published dishwasher demonstrations. Figure describes one four-minute Helix 02 sequence as autonomous and uncut. The evidence shows the task is technically possible under those conditions. It does not establish a standardized success rate across different racks, kitchens, dishes and lighting, so everyday reliability remains unproven.
Is 1X NEO autonomous or teleoperated?
NEO combines autonomous skills with a human-in-the-loop service. 1X says the robot arrives with foundational autonomy and can use Expert Mode when help is needed. That may involve remote human operation or demonstration. The correct label is therefore mixed: autonomous for some skills and human-assisted for tasks outside the current capability set.
Which general home robot can I buy in 2026?
1X publishes a NEO Early Access ownership price and monthly option with a refundable deposit. Weave offers Isaac 0 in a limited California program and reservations for Isaac 1. Sunday Memo remains a free beta application rather than a retail sale. Availability depends on region, service coverage and the difference between reservation and shipment.
Are home robots safe for elderly users?
Current robots may support reminders, communication or light object retrieval under supervision. Unsupervised physical assistance is a different risk category. No system reviewed here publishes broad clinical evidence for lifting people, preventing falls or handling medication independently. Families should not treat a consumer robot demonstration as certified medical or emergency-care capability.
Sources and methodology
Only official demonstrations or research pages showing domestic tasks in real homes or credible household setups were included. Each entry was classified by autonomy evidence, commercial status and environmental scope. Specialized floor cleaners are used as a reliability benchmark but not counted as humanoids.
Verification date: July 11, 2026. Company success claims remain company-reported unless a field metric or independent study is identified. Missing intervention rates and failure data are stated rather than estimated.
- NEO product page — 1X Technologies · accessed July 11, 2026
- NEO order program — 1X Technologies · accessed July 11, 2026
- Sunday Memo home robot — Sunday Robotics · accessed July 11, 2026
- Sunday 2026 Beta Program — Sunday Robotics · accessed July 11, 2026
- Weave Robotics home page — Weave Robotics · accessed July 11, 2026
- Isaac 0 order page — Weave Robotics · accessed July 11, 2026
- Isaac 1 order page — Weave Robotics · accessed July 11, 2026
- Helix 02: Full-Body Autonomy — Figure AI · 2026
- Helix 02 Bedroom Tidy — Figure AI · 2026
- π0.5: a VLA model that learns from experience — Physical Intelligence · 2025
- Our robots are learning to fold laundry — Physical Intelligence · 2025
- Advances in robot dexterity — Google DeepMind · 2024
- Toyota Human Support Robot — Toyota Motor Corporation · 2015
Related TechniaHQ guides
Official image recommendations
- A real home scene with NEO, Memo, Isaac and Figure 03 labeled by task and autonomy class.
Home robots performing laundry, dishwasher and object-retrieval tasks with evidence labels — Official manufacturer media - Sunday Memo clearing a dining table in the company’s uncut sequence.
Sunday Memo wheeled home robot collecting dishes from a dining table — Sunday Robotics - Weave Isaac 0 folding laundry at its stationary household workstation.
Weave Isaac 0 robot folding household laundry — Weave Robotics - Figure 03 loading a dishwasher with Helix 02.
Figure 03 humanoid robot loading dishes into a dishwasher — Figure AI - Task matrix from reliable today to not yet reliable, with intervention icons.
Home robot task reliability matrix for cleaning, laundry, dishes, food, retrieval and elder assistance — TechniaHQ original graphic
Fact-check report
Verified: July 11, 2026
Confirmed
- Sunday states Memo is not for purchase until after its 2026 beta.
- 1X publishes NEO ownership, monthly and refundable-deposit terms.
- Weave publishes separate Isaac 0 and Isaac 1 prices and regional schedules.
Not confirmed or incomplete
- No common success-rate benchmark covers the cited household tasks.
- Independent validation of Memo and Figure home demonstrations is unavailable.
- The extent of remote intervention for each commercial home-robot session is not fully public.
Fast-changing information
- Beta schedules, regional availability, subscription terms and learned task libraries can change during 2026.