For a home buyer, this matters more than the headline number. A robot that can briefly lift a heavy object may still be bad at carrying groceries through a hallway, handing you a glass, or putting a loaded laundry basket on a shelf. The useful question is not "which robot is strongest?" It is: what kind of load, in what posture, for how long, and around whom?
Here is the practical way to read humanoid robot payload specs, robot lift capacity, and carry claims before you trust them in a real home.
The Three Payload Numbers Mean Different Things
Lift capacity is the biggest, least home-specific number. It usually means a robot can raise or support a load briefly, often close to its body, sometimes in a lab-friendly posture. A lift number is useful for understanding peak strength, but it does not prove the robot can walk, turn, stop, or manipulate the object.
Carry capacity is closer to a real-world chore. Carrying means the robot must stay balanced while moving with the object. The load affects foot placement, torso posture, braking distance, battery use, and safety around people. For a biped, a 10 kg bag is not just 10 kg. It is a moving mass that can swing, shift, or block sensors.
Arm payload is the most important spec for household manipulation. It asks what one arm can hold while reaching, grasping, or placing an object. This is the number that matters for mugs, laundry, packages, plates, cabinet doors, and most chores above floor level.
1X makes the distinction unusually clear on its official NEO order page. NEO is listed with a 154 lb lift, 55 lb carry, and 18 lb arm payload. Converted roughly, that is about 70 kg lift, 25 kg carry, and 8.2 kg arm payload. The arm number is far lower than the lift number, and that is exactly the point. For a kitchen or laundry room, the lower number is usually the honest one.
The Same Kilogram Can Be Easy or Hard
A 3 kg dumbbell, a 3 kg grocery bag, and a 3 kg casserole dish are not the same robotics problem. The dumbbell has a predictable handle. The grocery bag swings. The casserole may be hot, fragile, slippery, and wider than the gripper.
That is why posture language matters. Unitree's official G1 page lists about 2 kg arm maximum load for the base model and about 3 kg for the EDU configuration, while also warning that maximum arm load varies greatly under different arm extension postures. AGIBOT is even more explicit for the X2: the database records 3 kg max in specific postures and ≤1 kg across the full range. Those two figures describe very different buyer expectations.
The distinction is not a weakness. It is good disclosure. A robot arm is a lever. The farther the object is from the shoulder, elbow, or wrist, the harder the job becomes. A robot that can hold 3 kg close to the torso may only handle a much lighter object when reaching into a cabinet or extending toward a counter edge.
Payload Clues for Robots Tracked in ui44
The table below compares several home-relevant and manipulation-relevant robots tracked in ui44. It deliberately avoids ranking every robot by one number, because the numbers describe different jobs. It also separates the source basis: some payload clues are already in ui44's robot records, while the 1X NEO and Unitree G1 values are official product-page specs that are not yet represented as payload fields in the ui44 database.
| Robot | Status | Price / access | Published payload clue | Source basis | What the number is good for |
|---|---|---|---|---|---|
| 1X NEO | Pre-order | $20,000 early access | 18 lb arm payload; 55 lb carry; 154 lb lift | Official 1X order page, not ui44 payload data | Separating household manipulation from peak strength |
| Boston Dynamics Atlas | Active | No public price | 50 kg instant, 30 kg sustained capacity | ui44 description/capability record | Industrial material handling, not a home promise |
| Agility Digit | Active | Enterprise / RaaS | Box carrying up to 16 kg | ui44 capability record | Tote and warehouse-style carrying |
| Unitree G1 | Available | $13,500 base | About 2 kg arm load; EDU about 3 kg | Official Unitree page, not ui44 payload data | Research and light manipulation |
| AGIBOT X2 | Available | $24,240 list | 3 kg max posture; ≤1 kg full range | ui44 payload field | Honest reminder that reach changes capacity |
| ROBOTIS AI Sapiens K0 | Development | Not announced | 3 kg max arm payload | ui44 payload field | Open-source research baseline |
| NEURA 4NE-1 Mini | Pre-order | €19,999 standard | 3 kg payload | ui44 payload field | Compact education/research manipulation |
| Hello Robot Stretch 3 | Active | $24,950 | 2 kg payload | ui44 capability record | Real-home mobile manipulation research |
| PAL TALOS | Active | Contact for quote | 6 kg per arm fully extended | ui44 payload field | Heavy research humanoid manipulation |
| Reachy 2 | Active | About $70,000 | 3 kg per arm | ui44 payload field | Teleoperated and open-source manipulation |
The buyer takeaway is blunt: a robot with a smaller, well-described arm payload may be more useful than a stronger robot with vague demo claims. A ui44 robot page is still the right starting point for price, status, battery life, links, and comparable specs, but payload numbers should be read with their source and posture caveat attached. Hello Robot Stretch 3 only lists a 2 kg payload, but it is built around a compact mobile manipulator that can reach real shelves, cabinets, and tables in homes. For assistive and research users, that may be more actionable than a humanoid that looks stronger on video but has no clear manipulation envelope.
What Counts as a Home Payload?
A home robot payload should be judged against real objects, not gym numbers. Most everyday chores fall into four bands:
- Small objects under 1 kg: remotes, mugs, medicine bottles, and snack bowls. The hard part is not mass. It is grip, perception, and not knocking things over. A robot needs compliant fingers, wrist control, and object recognition more than brute force.
- Light chores around 2-6 kg: a few groceries, a folded towel stack, or a small trash bag. This is where many current arms begin to look plausible, but only if the object is easy to grasp and the route is simple.
- Heavier household carrying around 6-15 kg: laundry baskets, heavier grocery bags, pet food, and water packs. Now robot carry capacity, two-arm handling, balance, and safety limits matter more than a single-arm spec.
- Human lifting: a separate category entirely. A robot that can lift a large number on paper should not be assumed safe for lifting a child, pet, older adult, or person who has fallen. Human handling requires medical-grade safety, compliance, perception, liability coverage, and certification. A payload number alone is nowhere near enough.
Peak Strength Is Not the Home-Safety Limit
The most impressive number in this space may be Boston Dynamics' Atlas. Boston Dynamics' current Atlas page lists 50 kg instant and 30 kg sustained weight capacity, with a 90 kg robot, IP67 rating, tactile sensing, 360-degree camera view, self-swappable batteries, and a 4-hour battery life. That is serious industrial strength.
It is also not a consumer home buying signal. Atlas is positioned for enterprise material handling, workflow integration, and selected customer deployments. It shows what high-end industrial humanoids can do, not what you should expect from a 2026 home assistant.
The same caution applies to Digit. ui44 lists Digit at 175 cm, 65 kg, roughly 4 hours of battery life, and a 16 kg box-carrying capability. That is meaningful because Digit's job is logistics: moving totes in warehouses and factories. A tote is standardized. A home is not. A half-open grocery bag, stairs, a dog underfoot, and a child walking nearby are different constraints.
For home buyers, the safe operating limit is usually below the lab or industrial limit. The robot may need to slow down, use two hands, avoid stairs, refuse fragile objects, or ask for human help.
The Posture Clause Is the Most Important Fine Print
Whenever a spec says "arm payload," look for the posture clause. Ask whether the load is measured at full extension, close to the body, with the wrist straight, with the elbow bent, at a specific speed, during walking, or standing still.
Those details decide whether the number describes a useful chore or only a best case. A 3 kg arm payload at full reach is not the same promise as a 3 kg payload held tight against the torso.
PAL Robotics is clear about this for TALOS: the robot can lift 6 kg with one arm fully extended. That is a useful spec because it tells you the condition. ROBOTIS is similarly direct: AI Sapiens K0 is 1.3 m tall, 34 kg, has 23 degrees of freedom, and supports a 3 kg maximum arm payload. That is not a consumer claim, but it is a clean research-platform number.
Unitree's G1 page includes one of the best caveats in the market: maximum arm load varies greatly under different arm extension postures. Every manufacturer should say something like that. Without posture, a payload spec is a headline, not a buying guide.
How to Compare Two Robots Without Fooling Yourself
Use this sequence before comparing robots in /compare or a spreadsheet:
- Match the payload type. Do not compare lift against arm payload or sustained carry against a one-time peak number.
- Convert units. Pounds and kilograms hide differences. Convert everything before ranking.
- Check posture and reach. A full-extension arm payload is more useful than an unspecified max.
- Check the hand. A 5 kg arm with a weak gripper may be worse than a 2 kg arm with compliant fingers for kitchen tasks.
- Check status and access. A pre-order, enterprise robot, and available research platform are not equal buying options.
- Discount for home safety. Pets, children, stairs, glass, water, and narrow hallways all reduce the practical limit.
What Payload Actually Tells You About a Home Robot
Payload is a useful spec, but it is not a proxy for intelligence. It tells you what the hardware may be able to support under certain conditions. It does not tell you whether the robot can identify the right object, choose a grip, avoid spilling, recover from a slip, or decide that a task is unsafe.
That is why the best home-robot payload spec is not a single number. It is a bundle: arm payload, carry payload, object examples, posture limits, hand design, speed limits, safety behavior, and real deployment status.
In 2026, the clearest home-focused payload disclosure comes from 1X because NEO separates lift, carry, and arm payload on the same page. The most powerful public spec comes from Boston Dynamics Atlas, but that is an industrial robot. The most useful buyer fine print comes from AGIBOT and Unitree, because they make posture limits explicit. And the most quietly practical home-manipulation example may still be Hello Robot Stretch 3, because it trades humanoid theater for a modest, specific 2 kg payload in real indoor spaces.
If you are choosing a robot for chores, do not ask "how much can it lift?" Ask: what can it safely hold at arm's length, carry through my home, and put down without making the problem worse? That is the payload spec that matters.
Database context
Use this article as a privacy verification workflow
Turn the article into a real verification pass
Humanoid Robot Payload Specs Explained already points you toward 10 linked robots, 10 manufacturers, and 6 countries inside the ui44 database. That matters because strong buyer guidance is easier to apply when you can move immediately from a claim or warning into concrete product pages, manufacturer directories, component explainers, and country-level context instead of treating the article as an isolated opinion piece. The fastest next step is to turn the article into a shortlist workflow: open the linked robot pages, verify which specs are actually published for those models, then compare the surrounding manufacturer and component context before you decide whether the underlying claim changes your buying plan.
For this topic, the useful discipline is to separate the editorial lesson from the catalog evidence. The article gives you the framing, but the robot pages tell you what each product actually ships with today: sensor stack, connectivity methods, listed price, release timing, category, and support-relevant compatibility notes. The manufacturer pages then show whether you are looking at a one-off launch, a broader lineup pattern, or a company that spans multiple categories. That layered workflow reduces the risk of buying on a single marketing phrase or a single support FAQ.
Use the robot pages to confirm which products actually expose cameras, microphones, Wi-Fi, or voice systems, then use the manufacturer pages to decide how much of the privacy question seems product-specific versus brand-wide. On this route cluster, X2, NEO, and Atlas (Electric) form the fastest reality check. If you want a quick working shortlist, open Compare X2, NEO, and Atlas (Electric) next, then keep this article open as the reasoning layer while you compare structured data side by side.
Practical Takeaway
Every robot, manufacturer, category, component, and country reference below resolves to a real ui44 page, keeping the follow-up path grounded in database records rather than generic advice.
Suggested next steps in ui44
- Open X2 and note the listed sensors, connectivity methods, and voice stack before you interpret any policy claim.
- Cross-check the wider brand context on AGIBOT so you can see whether the privacy question touches one model or a broader lineup.
- Use the linked component pages to confirm how common the relevant sensors and connectivity layers are across the database.
- Keep a short note of which policy layers you checked, which device features are actually present on the robot page, and which items still depend on region- or app-level confirmation.
- Finish with Compare X2, NEO, and Atlas (Electric) so the policy reading sits next to structured product data.
Database context
Robot profiles worth opening next
Use the linked product pages as the evidence layer
The linked robot pages are where this article becomes operational. Instead of asking whether the headline is interesting, use the robot entries to inspect the actual mix of sensors, connectivity options, batteries, pricing, release timing, and stated capabilities attached to the products mentioned in the article. That is the easiest way to see whether the warning or opportunity described here affects one product family, a specific design pattern, or an entire buying lane.
X2 is tracked on ui44 as a available humanoid robot from AGIBOT. The database currently records a listed price of $24,240, a release date of 2025, ~2 hours at 0.5 m/s walking battery life, ~1.5 hours charging time, and a published stack that includes 3D LiDAR (Ultra), RGB-D Camera (Ultra), and RGB Cameras plus Wi-Fi and Bluetooth.
For privacy-focused reading, this page matters because it shows the concrete device surface behind the policy discussion. Use it to verify whether X2 combines sensors and connectivity in a way that could change the in-home data footprint, and compare the listed capabilities such as Bipedal Walking, 25-30 DOF Articulation, and Object Manipulation (with OmniHand accessory) with any cloud, app, or voice layers.
NEO
1X Technologies · Humanoid · Pre-order
NEO is tracked on ui44 as a pre-order humanoid robot from 1X Technologies. The database currently records a listed price of $20,000, a release date of 2025-10-28, ~4 hours battery life, Not disclosed charging time, and a published stack that includes RGB Cameras, Depth Sensors, and Tactile Skin plus Wi-Fi and Bluetooth.
For privacy-focused reading, this page matters because it shows the concrete device surface behind the policy discussion. Use it to verify whether NEO combines sensors and connectivity in a way that could change the in-home data footprint, and compare the listed capabilities such as Household Chores, Tidying Up, and Safe Human Interaction with any cloud, app, or voice layers.
Atlas (Electric)
Boston Dynamics · Humanoid · Active
Atlas (Electric) is tracked on ui44 as a active humanoid robot from Boston Dynamics. The database currently records a listed price of Price TBA, a release date of 2026, ~4 hours battery life, Not disclosed charging time, and a published stack that includes 360° camera view and Tactile plus Wi-Fi and Ethernet.
For privacy-focused reading, this page matters because it shows the concrete device surface behind the policy discussion. Use it to verify whether Atlas (Electric) combines sensors and connectivity in a way that could change the in-home data footprint, and compare the listed capabilities such as Heavy Lifting (50kg Instant, 30kg Sustained), Precise Manipulation, and Dynamic Recovery with any cloud, app, or voice layers.
Digit is tracked on ui44 as a active humanoid robot from Agility. The database currently records a listed price of Price TBA, a release date of 2023, ~4 hours battery life, ~2 hours charging time, and a published stack that includes LiDAR, RGB-D Cameras, and IMU plus Wi-Fi and 5G.
For privacy-focused reading, this page matters because it shows the concrete device surface behind the policy discussion. Use it to verify whether Digit combines sensors and connectivity in a way that could change the in-home data footprint, and compare the listed capabilities such as Box Carrying (16kg), Stair Navigation, and Warehouse Operations with any cloud, app, or voice layers.
G1 is tracked on ui44 as a available humanoid robot from Unitree. The database currently records a listed price of $13,500, a release date of 2024, ~2 hours battery life, Not disclosed charging time, and a published stack that includes Depth Camera, 3D LiDAR, and 4 Microphone Array plus Wi-Fi 6 and Bluetooth 5.2.
For privacy-focused reading, this page matters because it shows the concrete device surface behind the policy discussion. Use it to verify whether G1 combines sensors and connectivity in a way that could change the in-home data footprint, and compare the listed capabilities such as Bipedal Walking, Object Manipulation, and Dexterous Hands (optional Dex3-1) with any cloud, app, or voice layers.
Database context
Manufacturer context behind the article
Check whether this is one product story or a broader company pattern
Manufacturer pages add the market context that individual product pages cannot show on their own. They help you check whether the article is centered on a brand with a deep lineup, whether that brand spans several categories, and how much of its ui44 footprint depends on one flagship model versus a broader product strategy. That matters for topics like privacy, warranty terms, setup friction, and launch promises because the surrounding lineup often reveals whether a pattern is isolated or systemic.
AGIBOT
ui44 currently tracks 6 robots from AGIBOT across 2 categorys. The company is grouped under China, and the current catalog footprint on ui44 includes A2 Ultra, X2, Expedition A3.
That wider brand context matters because privacy questions rarely stop at one FAQ page. A manufacturer route helps you see whether the article is centered on one premium model or on a company that has several relevant products and therefore more than one place where the same policy or app assumptions might matter. The category mix here currently points toward Humanoid, Quadruped as the most useful next route if you want to see whether this article reflects a wider pattern inside the brand.
1X Technologies
ui44 currently tracks 2 robots from 1X Technologies across 1 category. The company is grouped under Norway, and the current catalog footprint on ui44 includes NEO, EVE.
That wider brand context matters because privacy questions rarely stop at one FAQ page. A manufacturer route helps you see whether the article is centered on one premium model or on a company that has several relevant products and therefore more than one place where the same policy or app assumptions might matter. The category mix here currently points toward Humanoid as the most useful next route if you want to see whether this article reflects a wider pattern inside the brand.
Boston Dynamics
ui44 currently tracks 3 robots from Boston Dynamics across 2 categorys. The company is grouped under USA, and the current catalog footprint on ui44 includes Atlas (Electric), Spot, Stretch.
That wider brand context matters because privacy questions rarely stop at one FAQ page. A manufacturer route helps you see whether the article is centered on one premium model or on a company that has several relevant products and therefore more than one place where the same policy or app assumptions might matter. The category mix here currently points toward Humanoid, Commercial as the most useful next route if you want to see whether this article reflects a wider pattern inside the brand.
Agility
ui44 currently tracks 1 robot from Agility across 1 category. The company is grouped under USA, and the current catalog footprint on ui44 includes Digit.
That wider brand context matters because privacy questions rarely stop at one FAQ page. A manufacturer route helps you see whether the article is centered on one premium model or on a company that has several relevant products and therefore more than one place where the same policy or app assumptions might matter. The category mix here currently points toward Humanoid as the most useful next route if you want to see whether this article reflects a wider pattern inside the brand.
Database context
Broaden the scan without leaving the database
Categories, components, and countries add the wider context
Category framing
Category pages are useful when the article touches a buying pattern that shows up across brands. A category route helps you confirm whether the linked products sit in a narrow niche or whether the same question should be tested across a larger field of alternatives.
Humanoid
The Humanoid category page currently groups 61 tracked robots from 44 manufacturers. ui44 describes this lane as: Full-size bipedal humanoid robots designed to work alongside humans. From factory floors to household tasks, these machines represent the cutting edge of robotics.
That makes the category route a practical follow-up when you want to check whether the products linked in this article are typical for the lane or whether they sit at one edge of the market. Useful starting examples currently include NEO, EVE, Mornine M1.
Research
The Research category page currently groups 21 tracked robots from 16 manufacturers. ui44 describes this lane as: Academic and research robotics platforms pushing the boundaries of what machines can learn and do.
That makes the category route a practical follow-up when you want to check whether the products linked in this article are typical for the lane or whether they sit at one edge of the market. Useful starting examples currently include HRP-4C, HRP-5P, NAO6.
Country and ecosystem context
Country pages give extra context when support practices, launch sequencing, regulatory posture, or manufacturer mix matter. They are not a substitute for model-level verification, but they do help you see which ecosystems cluster together and which manufacturers sit in the same regional field when you broaden the search beyond the article headline.
China
The China route currently groups 46 tracked robots from 14 manufacturers in ui44. That gives you a useful regional lens when the article points toward support practices, launch sequencing, or brand clusters that may share similar ecosystem assumptions.
On the current route, manufacturers like AGIBOT, Roborock, Unitree Robotics make the page a good way to broaden the scan without losing the regional context that often shapes availability, documentation style, and adjacent alternatives.
Norway
The Norway route currently groups 2 tracked robots from 1 manufacturers in ui44. That gives you a useful regional lens when the article points toward support practices, launch sequencing, or brand clusters that may share similar ecosystem assumptions.
On the current route, manufacturers like 1X Technologies make the page a good way to broaden the scan without losing the regional context that often shapes availability, documentation style, and adjacent alternatives.
USA
The USA route currently groups 16 tracked robots from 12 manufacturers in ui44. That gives you a useful regional lens when the article points toward support practices, launch sequencing, or brand clusters that may share similar ecosystem assumptions.
On the current route, manufacturers like Boston Dynamics, Figure AI, Tesla make the page a good way to broaden the scan without losing the regional context that often shapes availability, documentation style, and adjacent alternatives.
Database context
Questions to answer before you move from reading to buying
A follow-up FAQ built from the entities already linked in this article
Frequently Asked Questions
Which page should I open first after reading “Humanoid Robot Payload Specs Explained”?
Start with X2. That gives you a concrete product anchor for the article’s main claim. From there, branch into the manufacturer and component pages so you can tell whether the article is describing one specific model, a repeated brand pattern, or a wider technology issue that affects multiple shortlist options.
How do the manufacturer pages change the buying decision?
AGIBOT help you zoom out from one article and one product. On ui44 they show lineup breadth, category spread, and the neighboring robots tied to the same company. That context is useful when you are deciding whether a risk belongs to a single model, whether it shows up across a brand’s portfolio, and whether you should keep looking at alternatives before committing.
When should I switch from reading to side-by-side comparison?
Move into Compare X2, NEO, and Atlas (Electric) as soon as you understand the article’s main warning or promise. The article explains what to watch for, but the compare view is where you can check whether price, status, battery life, connectivity, sensors, and category fit still make the robot a good match for your own home and budget.
Database context
Where to go next in ui44
Keep the research chain inside the database
If you want to keep going, these follow-on pages give you the cleanest expansion path from article to research session. Open the comparison route first if you are deciding between products today. Open the manufacturer, category, and component routes if you still need to understand the broader pattern behind the claim.
Written by
ui44 Team
Published April 26, 2026
Share this article
Open a plain share link on X or Bluesky. No embeds, no widgets, no cookie baggage.