Astribot T1: What a $13K Humanoid Buys

It is also not a finished home butler. The useful way to read T1 is as a lower entry price for a manipulation platform, not as proof that a household can buy a safe, supported, autonomous chore robot today.

The short version: Astribot T1 is one of the most interesting affordable humanoid signals of 2026 because it combines a real published price band with a useful payload and wheeled stability. But the gaps matter. Battery life, sensor stack, safety certifications, support model, regional delivery, and the split between autonomy, teleoperation, and scripted demos are still not clear enough for a normal buyer.

What did Astribot announce?

Astribot, also known as Stardust Intelligence, announced T1 in late May 2026 as a compact wheeled humanoid follow-on to Astribot S1. In ui44's database, T1 is tracked as a pre-order robot at ¥89,900 CNY, with a note that public configuration, deposit, delivery timing, and regional availability terms are not fully disclosed.

The published and corroborated headline specs are unusually specific for this price tier:

That mix is why T1 is worth covering. A 5 kg single-arm payload is meaningful for household-scale objects: pots, small grocery bags, bottles, folded laundry, small appliances, and many tool or lab items sit below that line. The wheeled base also avoids the hardest bipedal reliability problem. In a home, wheels can be a practical advantage if the robot stays on one floor and does not need to climb stairs.

But published specs are not the same as home readiness. A robot can be tall, strong, and affordable on paper while still being unsuitable for kitchens, children, pets, carpet edges, clutter, privacy-sensitive rooms, and warranty support.

What does the $13K price actually buy?

It buys a lower-cost entry point into embodied-AI hardware. That is important. For years, humanoid pricing has split into two unsatisfying categories: research or enterprise platforms that are too expensive for almost everyone, and consumer concepts with vague pricing and vague delivery. T1's starting price is useful because it makes the comparison sharper.

The price appears to buy the physical platform: a 155 cm, 66 kg wheeled body; dual-arm manipulation; a cable-driven architecture; 23 degrees of freedom before end effectors; and compatibility with grippers or five-finger hands. It also buys access to Astribot's direction of travel: learning from human demonstrations and turning manipulation demos into deployable workflows.

It probably does not buy the thing many consumers imagine when they hear "home humanoid." It does not yet prove that T1 can unload a dishwasher without breaking glass, fold a basket of mixed clothing unsupervised, identify what belongs in which drawer, recover when a pet blocks the path, or explain exactly who is liable if it drops a pan.

That distinction matters because a low price can create a false sense of maturity. A cheaper robot is not automatically a consumer robot. It may simply be a cheaper developer platform.

How T1 compares with Unitree G1, 1X NEO, and other home-adjacent robots

The closest price comparison is Unitree G1, which ui44 tracks as available from $13,500. G1 is smaller and lighter than T1 at 132 cm and 35 kg, and it is a biped rather than a wheeled robot. Its standard arm load is around 2 kg, with about 3 kg on the EDU configuration. That makes G1 a remarkable research humanoid for the money, but not a supported household helper.

T1 goes after a different trade-off. It is heavier, wheeled, and claims a higher single-arm payload. If your task is moving around a flat indoor workspace and manipulating objects, wheels plus 5 kg per arm may be more immediately useful than bipedal locomotion. If your task requires stairs, uneven outdoor terrain, or studying humanoid walking, G1 remains the clearer research reference.

1X NEO is a different kind of comparison. ui44 tracks NEO as a $20,000 pre-order home-focused humanoid with a soft, lightweight body, 167 cm height, 30 kg weight, depth sensors, tactile skin, and roughly 4 hours of battery life. NEO is more explicitly designed around private homes, while T1 is framed across home, commercial, research, and industrial scenarios. If you are evaluating consumer readiness, NEO's home-first design is relevant. If you are evaluating price-to-manipulation hardware, T1 is the sharper new signal.

NEURA 4NE-1 Mini starts at €19,999 before taxes and shipping, with a €29,999 Pro tier that adds dexterous hands, teleoperation, SDK access, and digital-twin tools. It is also 132 cm tall and 36 kg, with a listed 3 kg payload and roughly 2.5 hours of battery life. That makes it another serious lower-cost humanoid reference, but still mostly a research, education, and developer proposition.

Then there is Hello Robot Stretch 4, which is not a humanoid, yet may be more honest for some home-assistance work. ui44 tracks Stretch 4 at $29,950, available now, with an 8-hour light-load runtime, a 160 cm working height, self-charging, ROS 2/Python access, and a telescoping arm rated for 2.5 kg extended or 4 kg retracted. It looks less futuristic than a humanoid, but it is built around bounded mobile manipulation in real indoor spaces.

The lesson is simple: price only answers one question. Home usefulness comes from the combination of manipulation, mobility, autonomy, safety, software, support, and repeatability.

The good signal: wheeled manipulation at a lower price

T1's most credible home-relevant signal is not that it has a face or a humanoid outline. It is that Astribot is emphasizing manipulation.

The launch materials and independent launch coverage describe cooking demos, laundry folding, laboratory operations, automotive parts sorting, and EV charging tasks. Those are not all home tasks, but they share the same underlying problem: the robot has to bring an arm to an object, choose a contact, apply the right force, move without snagging or dropping, and recover when the scene is not perfect.

The wheeled base helps here. A wheeled humanoid can be less romantic than a walking one, but wheels are often the practical choice indoors. They reduce the number of failure modes, make the platform more stable during arm motion, and can stretch battery and compute budget toward perception and manipulation instead of whole-body balance. For a kitchen, lab bench, laundry counter, or workshop, that may be a smart compromise.

T1's 5 kg single-arm payload also puts it above many small companion robots and some lower-cost humanoids. Payload is not dexterity, but it defines the work envelope. A robot that cannot lift the object cannot complete the task no matter how good its AI model is.

The weak signal: home autonomy is still mostly unproven

The largest missing piece is not another spectacular demo. It is boring proof. How long can T1 work unplugged? What sensors are in the shipping configuration? What compute is included at the ¥89,900 price? Which demos are autonomous, which are learned from human demonstration, and which require supervision or scripted setups? How often does it fail, and what happens after a failed grasp?

These are not cynical questions. They are the difference between a platform and a product. A developer or lab may accept missing information because the goal is to experiment. A home buyer needs something else: clear delivery terms, repair options, safety limits, privacy controls, software updates, insurance clarity, and a way to return or service a 66 kg robot.

The word "home" also needs discipline. T1 may be home-adjacent because the tasks shown or claimed overlap with household chores. That does not mean it is ready for unsupervised use in an ordinary apartment. Homes contain people who did not sign up to be robot operators: children, guests, caregivers, roommates, pets, and neighbors. A home robot must understand when not to act.

Should you preorder Astribot T1 for a home?

For most households, no. At least not yet.

A serious robotics developer, AI lab, service-robot integrator, or company building manipulation workflows may reasonably watch T1 closely. The price could make experimentation easier, especially if Astribot publishes clearer SDK, compute, end-effector, and support details. For that audience, T1 could be a useful bridge between expensive research humanoids and narrow industrial arms.

For an ordinary home buyer, the bar should be higher. Before treating T1 as a home robot, ask for five things:

1. Exact configuration: what hardware, compute, hands, grippers, sensors, battery, charger, and software access are included at ¥89,900?
2. Delivery and support: where does Astribot ship, when, with what warranty, repair process, spare parts, and updates?
3. Autonomy disclosure: which tasks run without remote human control, and how does the robot ask for help?
4. Safety and privacy: what force limits, emergency stops, logs, camera controls, and household data policies are in place?
5. Repeatability: can the robot perform the same useful household task many times across messy rooms, not just once in a launch clip?

If those answers are vague, treat T1 as a promising platform rather than a consumer purchase.

What T1 means for the home-humanoid race

T1 matters because it pressures the market. A published starting price near Unitree G1 changes the conversation from "humanoids are someday expensive" to "which compromises make a lower-cost humanoid useful?" That is healthy. It forces buyers to compare payload, stability, runtime, dexterity, sensors, and software instead of only watching viral videos.

It also reinforces a pattern ui44 keeps seeing: the first real path toward home robots may not look like a perfect walking servant. It may look like a wheeled manipulator, a developer humanoid, a teleoperated assistive robot, or a platform that proves one narrow workflow before expanding. Zeroth M1 shows the small companion end of that spectrum. Stretch 4 shows the practical mobile-manipulator end. T1 sits closer to the humanoid-developer middle.

Compared with Figure 03, which has no consumer price and is focused on industrial evidence, T1 is more accessible but less proven. Compared with 1X NEO, it is less clearly home-first but potentially more interesting as a lower-cost manipulation platform. Compared with Unitree G1, it gives up bipedal walking in exchange for wheeled stability and a stronger claimed arm payload.

That is a real product strategy, not just a gimmick. But the buyer conclusion is still cautious.

ui44 take

Astribot T1 is a strong signal for affordable humanoid hardware and a weak signal for immediate household usefulness.

If Astribot follows the launch with detailed specs, transparent configuration pricing, delivery terms, developer documentation, safety information, and repeatable task metrics, T1 could become one of the most important affordable humanoid platforms in the database. If the evidence stays at the level of launch videos and broad scenario lists, it remains a fascinating pre-order robot rather than a home helper.

For now, the smartest reading is: T1 makes the affordable humanoid race more real, but it does not make the home-humanoid problem solved. Compare it in the ui44 robot database, check alternatives in /compare, and wait for the boring details before treating any $13K humanoid as a safe household purchase.