LinkerBot Hands: Cheaper Humanoid Dexterity

The most interesting recent signal is LinkerBot. In a company-issued PRNewswire release, the Beijing company says it has shipped its 10,000th dexterous hand, raised $150 million since April 2025, and is producing more than 1,000 high-DoF hands per month. Those are company claims, not audited industry statistics, but they point to a serious shift: high-degree-of-freedom hands may commoditize before full home humanoids become useful.

That is good news for robotics. It is not proof that a $20,000 home humanoid will fold laundry next year. Cheaper hands solve only one layer of the problem. The robot still needs arms strong enough to use those hands, tactile and force data, robust autonomy, safe behavior near people, and a service network that can replace broken fingers without turning the whole machine into expensive e-waste.

Why are robot hands becoming a supply-chain story?

For years, the practical end of home robotics avoided human-like hands. Robot vacuums used brushes. Delivery robots used cargo boxes. Companion robots used faces, wheels, speakers, and cameras. Even serious mobile manipulators often used a simple gripper because two or three controlled surfaces are easier to model, cheaper to repair, and less likely to snag on the world.

A humanoid changes that trade-off. If the sales pitch is "works in a human home," the robot must eventually operate human tools. Door handles, cabinet pulls, cloth, buttons, spray bottles, cutlery, mugs, and folded packaging were not designed for a vacuum dock. They were designed around hands.

That is the supply-chain opening LinkerBot is trying to occupy. Its official Chinese site describes LinkerBot as a provider of high-performance dexterous robotic hands, including L10, L20, and L30 product families, while the PRNewswire release says the lineup spans 6 to 42 degrees of freedom and includes hardware plus an "Open TeleDex" teleoperation system and "LinkerSkillNet" software ecosystem. A native-language Sina/Titanium Media interview with founder Zhou Yong frames the same thesis more bluntly: hands are the core interface for robots entering human environments, and scaling hands is also a way to scale skill data.

The key buyer-facing point is simple: if hands become modular components bought from specialists, humanoid companies do not all have to invent fingers from scratch. That can lower cost and speed iteration. It can also make robot hands more replaceable, which matters in homes where drops, jams, pets, kids, and ordinary wear will punish fragile hardware.

What LinkerBot appears to have proven

Treat LinkerBot's numbers as claimed milestones, not neutral market data. Even with that caveat, the pattern is notable.

The mistake would be to read this as "cheap hands make home robots solved." The better reading is: hand hardware is becoming a purchasable subsystem. That is how mature hardware markets usually develop. Smartphone makers do not build every camera sensor. Automakers do not machine every component in-house. Home humanoid companies may eventually differentiate on autonomy, safety, service, software, and task proof while buying hands, actuators, batteries, cameras, or teleoperation systems from specialist suppliers.

What the ui44 database says about dexterity today

The ui44 database already shows a split between three groups: available mobile manipulators, developer humanoids, and high-dexterity platforms that remain commercial, research, or early-access products rather than ordinary home buys.

This is why cheaper hands could matter first for developers, labs, factories, service pilots, and early adopter platforms. They reduce the cost of trying manipulation. They do not automatically make the robot safe, autonomous, quiet, repairable, or useful in a kitchen at 7:30 a.m.

The cost curve can move before home usefulness does

A lower-cost dexterous hand can still change the home-robot timeline in three ways.

First, it increases experimentation. If a five-finger hand costs tens of thousands of dollars, far fewer labs and startups will attach one to a mobile base and collect contact data. If usable hands fall into the low-thousands range for some configurations, more teams can try laundry, drawers, dish racks, food packaging, appliance buttons, and tool use.

Second, it separates the robot from the end effector. A future home robot may not ship with one universal hand forever. The hand could be a replaceable module with a stronger gripper for chores, a softer hand for eldercare, or a cheaper replacement assembly for repairs. That is especially important if fingers become consumable parts rather than museum-grade mechanisms.

Third, it pressures full-stack humanoid companies. If an outside supplier can make hands cheaply and reliably, then a humanoid maker has to justify why its in-house hand is better. The answer might be tighter integration, better tactile skin, higher force, safer compliance, or better autonomy data. But it needs to be an answer buyers can see in tests, not just in a launch video.

The important caveat is that a robot hand is not a whole manipulation system. A hand with many degrees of freedom can still fail if the wrist is weak, the arm cannot reach the shelf, the robot cannot detect slip, the planner cannot recover from a bad grasp, or the warranty makes every broken finger a painful support case.

What cheaper hands still do not solve

There are at least six unsolved buyer questions that cheaper hands do not remove.

1. Arm payload and reach. The hand may be capable, but the arm still needs to lift the object, keep it stable, and reach useful places. Unitree G1's database entry is a good reminder: a $13,500 starting price is exciting, but standard arm payload is still around 2 kg and the platform is aimed at development.

2. Tactile proof. "Dexterous" can mean finger count, joint count, or real contact awareness. A buyer should ask whether the robot can detect slip, force, object deformation, and bad grasps in live tasks.

3. Recovery behavior. The hard part is not picking up one mug once. It is noticing when the mug is slipping, choosing not to pour, setting it down safely, and trying again.

4. Durability. Hands touch everything. They will hit counters, pinch fabric, jam against drawers, drag on tables, and get dusty. A cheap hand that breaks often is not cheap in a home.

5. Safety near people. Home robots need compliant motion, pinch-point limits, force caps, emergency stops, and conservative behavior around children, pets, and fragile objects.

6. Service and parts. If a finger module fails, can the owner replace it? Is there a regional repair path? Is the hand calibrated after replacement? Is the warranty void if the robot grabs the wrong thing?

These questions matter more than whether a robot has five fingers in a photo. A gripper-based system like Hello Robot Stretch 4 may be less anthropomorphic, but it is available, open, self-charging, and built around real mobile manipulation. A five-finger humanoid can be more general in principle while being less useful in practice if the autonomy and support layer are immature.

What buyers should watch next

For the next wave of home humanoids, ignore the hand photo for a moment and ask for evidence in four areas.

Repeated task tests. A single clip of a hand folding cloth is not enough. Look for dozens or hundreds of trials with failure rates, reset counts, and intervention rates.

Task variety. Good hands should help across rigid, soft, slippery, fragile, and tool-use tasks. If every demo uses the same object on the same table, the skill may be narrow.

Repair terms. A home robot hand is a wear component. Buyers should expect finger modules, skins, tendon/linkage parts, and calibration service to be clear before trusting a robot with daily chores.

Supplier transparency. If a humanoid company uses an outside hand supplier, that is not automatically bad. It may be a strength. But the buyer should know whether the hand is proprietary, replaceable, field-serviceable, and supported in their region.

Bottom line: cheaper hands are necessary, not sufficient

LinkerBot's rise is a strong robotics signal because hands sit at the exact boundary between AI hype and physical usefulness. A cheaper, higher-volume robot hand supply chain could make experimentation faster and lower the bill of materials for future humanoids. It could also create more standardized parts, more teleoperation data, and more pressure on companies to prove real manipulation instead of just walking.

But for home buyers, the conclusion is still cautious. A good hand does not make a good home robot. The useful robot is the one that combines hand hardware with safe arms, tactile feedback, recovery behavior, software updates, clear service, and honest task evidence.

So yes, cheaper dexterous hands matter. They are one of the reasons home humanoids may eventually move from staged demos to practical chores. Just do not confuse the hand supply chain arriving with the home robot problem being solved.