Supply chain
Market position: Highest-cost and hardest-reliability pool for humanoid motion.
Supplier examples: Leaderdrive, Harmonic Drive, precision transmission / servo vendors
Customer signal: Humanoid OEM disclosure is still limited; filings and component revenue are the public proxy.
Market position: China has strong battery supply-chain depth; humanoid-specific design is still early.
Supplier examples: Cell / pack ecosystem, power electronics suppliers
Customer signal: Watch runtime per shift, charging workflow and field replacement cost.
Market position: Cost may fall with commodity sensors, but safety/reliability can require redundancy.
Supplier examples: Camera, lidar/depth, tactile and force-sensing vendors
Customer signal: Watch sensor-count reduction vs safety-case requirements.
Market position: US frontier stack is stronger; NVIDIA / model-layer platforms are key references.
Supplier examples: NVIDIA Isaac / GR00T, robot foundation-model companies, edge-compute vendors
Customer signal: Watch software revenue, deployed robot count, intervention-rate reduction.
Market position: China manufacturing depth can matter if humanoids move from prototypes to repeatable production.
Supplier examples: Machining, casting, harnessing, thermal, testing and EMS partners
Customer signal: Watch yield, serviceability, parts count, accepted units and warranty cost.
Market position: Likely bifurcates: general dexterous hands for demos, task-specific end effectors for early deployment.
Supplier examples: Leaderdrive micro harmonic reducers for hands/wrists, dexterous-hand vendors, gripper/tooling suppliers, OEM in-house designs
Customer signal: Watch task success, finger/tool replacement rate, tactile sensing and named OEM/customer attach.
绿的谐波 is useful because it shows both sides of the supply-chain question: filing-backed acceleration is real, but product fit and stock-price moves are not customer-economics proof.
2025 revenue RMB 570.714m (+47.31%), net profit RMB 124.367m (+121.42%), industrial / embodied-intelligence robot-component revenue RMB 422.528m (+52.61%, 34.88% GM), harmonic reducer sales 425,158 units (+72.48%), and 2026 Q1 revenue RMB 140.134m (+42.96%).
This is stronger than theme exposure: public filings show revenue, profit and unit-volume acceleration. Product breadth across reducers, micro reducers, rotary actuators, frameless motors, PRS / linear actuators, joint modules and DD motors gives architecture optionality.
Still missing: named humanoid OEM, contract value / backlog / order volume, humanoid-specific revenue and margin, validated BOM attach, single-robot value content, customer ROI/payback, repeat orders and margin durability at scale.
Do not call the RMB 422.528m line pure humanoid revenue; the filing wording mixes industrial and embodied-intelligence robot components.
Do not infer Tesla, Figure, Unitree, AgiBot, UBTECH or any named OEM without primary confirmation, filing evidence, teardown or identifiable production-part proof.
Do not treat the 2026-06-04 to 2026-06-08 share-price move of +36.00% as order evidence; the 2026-06-09 announcement said there was no undisclosed major matter.
Evidence map only. No trade recommendation. Product fit ≠ customer proof; stock-price move ≠ order evidence. Read the full case study at /robotics/supply-chain/leaderdrive/.
2025 harmonic reducers produced; robot / embodied-intelligence component revenue is the key public proxy.
Global reducer benchmark; relevant for price / quality / lifetime comparison.
Isaac / GR00T anchors simulation, compute and foundation-model tooling.
Motion-control suppliers may benefit if humanoid actuators become a scaled category.