AI supply-chain thesis — mapping bottlenecks, focus companies, and supply-chain exposure for investors.
**Unitree Robotics has filed for a Shanghai STAR Market IPO at roughly RMB 4.2B / $608M after shipping the world's cheapest credible humanoid — the $27,300 G1 with an estimated bill of materials under $9,000 — by making its own motors, gearboxes, lidars, and depth cameras in-house and pulling everything else from suppliers a train ride away in the Shenzhen Pearl River Delta ecosystem.** That vertical-and-local model is what blew apart the Western humanoid cost stack and is what makes Unitree the structural anchor for the 2026–2027 humanoid bill-of-materials trade. ## Where Unitree sits in the supply chain Unitree is the OEM. Upstream it touches four layers — each one a separately investable bucket. **Rare-earth feedstock and magnets.** A humanoid carries 40–50 actuators (every joint, every finger, every wrist), each wound on NdFeB sintered magnets. That is roughly 10–20× the rare-earth content of an EV per unit. The listed ex-China feedstock layer is $MP, $USAR, $ALOY, $LYC.AX, $ILU.AX, $UUUU, $TMC. JL Mag and Ningbo Yunsheng dominate inside China. **Servo motors and BLDC actuators.** Unitree designs its own motors but the named upstream actuator supplier across Chinese humanoid OEMs (Unitree, UBTech, AGIBot) is $300124.SZ Inovance — China's #1 servo-motor maker at roughly $22B mcap. Tuopu Group (601689.SS, not yet on Macroplane) supplies rotary actuators and dexterous-hand assemblies. The Western incumbents Unitree underbids: $6594.T Nidec (BLDC and coreless motors) and $MOG.A Moog (aerospace-grade actuators). **Strain-wave reducers — the precision bottleneck.** Cycloidal and harmonic gearboxes are the joint-precision moat. Global incumbent: $6324.T Harmonic Drive Systems (Japan). $6268.T Nabtesco is the cycloidal RV leader for the heavier joints. $SHA.DE Schaeffler is the bearings and planetary-screws layer. Chinese challenger Leader Drive (688017.SS, not on Macroplane) ships at roughly a third of Harmonic Drive pricing and is Tesla Optimus's stated Mexico supplier. Unitree builds its own planetary gearboxes in-house, sidestepping this layer for low-ratio joints — but the rest of the humanoid market still buys from Harmonic, Nabtesco, and Schaeffler. **Sensors and perception.** Unitree self-develops its lidars and depth cameras, which is what flipped its sensor cost vs. competitors. The merchant lidar layer — relevant for any humanoid OEM that does not verticalize — anchors on $HSAI Hesai. Magnetic position sensors and IMUs for joint feedback: $ALGM Allegro, $MELE.BR Melexis, $6762.T TDK (the latter owning InvenSense for IMUs). **Edge compute.** $NVDA (Jetson SoCs and GR00T VLA models — Nvidia is also one of Unitree's largest research-platform customers, alongside Apple and Meta) and $QCOM (RB6 robotics platform) supply the on-board brain. ## The edge Three things make Unitree structurally different from Western humanoid OEMs. 1. **In-house motors at 30–40% of Western prices.** Unitree designs and builds its own BLDC motors and the low-ratio planetary gearboxes. That removes the Harmonic Drive / Maxon margin from the BOM on every joint that does not need strain-wave precision. 2. **Shenzhen-PRD proximity to suppliers.** Prototype samples arrive next-day; iteration cycles are weeks, not quarters. Western humanoid teams sit 6–18 months behind on every BOM revision. 3. **Quadruped scale paid for the humanoid R&D.** Unitree shipped Go2 quadrupeds for $1,600–$2,800 from 2021 onward, generating cash and validating the motor/gearbox stack at scale before pivoting that stack into the G1. The result: G1 ships at ~$27,300 with an estimated 67% gross margin. Tesla Optimus and Figure are still iterating sub-1,000-unit BOMs at numbers Western chains cannot match. ## The risks - **IPO timing.** A $608M Shanghai STAR raise into a humanoid sector with little revenue and concentrated research-platform buyers (Nvidia, Apple, Meta — hundreds of units each) is priced on growth, not earnings. Any slip past the 10,000-unit industrial deployment threshold compresses the multiple immediately. - **Export controls on rare earths and lidar.** Both flow upstream into the BOM. A symmetric move (China cutting magnet exports, US restricting Jetson-class compute) hurts Western humanoid OEMs harder than Unitree, but Unitree still loses access to its Nvidia, TI, and ADI dependencies. - **$TSLA Optimus and AGIBot Walker.** Optimus has Tesla's manufacturing depth; AGIBot has Chinese capital and is already shipping Walker units. Both compress Unitree's margin path even where Unitree wins on cost. - **Verticalization cuts both ways.** Building its own motors and lidars means the structural cost edge is real, but the dependency on the merchant layer shrinks — meaning the upside for $ALGM, $MELE.BR, $HSAI from "Unitree adoption" is muted vs. the upside from Optimus, Figure, and Apptronik. ## The catalysts - **Shanghai STAR Market IPO pricing and first-day trading.** Sets the comparable for every public humanoid pure-play including KOID (KraneShares Humanoid ETF) and KSTR (STAR Market ETF). - **G1 industrial-deployment milestones.** Sub-$30/hour tote-handling labour cost was the stated viability threshold; scaling past 10,000 units is the next. - **$300124.SZ Inovance humanoid revenue split.** Earnings will start breaking out the humanoid contribution from EV/industrial servo motors — the cleanest listed proxy for Unitree volume. - **$6324.T Harmonic Drive order book commentary.** If the Chinese gearbox displacement plays out, this is where the squeeze first shows up. The picks-and-shovels of Unitree are not Unitree shares — they are the rare-earth feedstock, the listed Chinese servo-motor suppliers, the Japanese gearbox incumbents under cost pressure, and the Western sensor IC vendors who win on every humanoid that is not vertically integrated. Not investment advice; do your own research.
**Unitree Robotics has filed for a Shanghai STAR Market IPO at roughly RMB 4.2B / $608M after shipping the world's cheapest credible humanoid — the $27,300 G1 with an estimated bill of materials under $9,000 — by making its own motors, gearboxes, lidars, and depth cameras in-house and pulling everything else from suppliers a train ride away in the Shenzhen Pearl River Delta ecosystem.** That vertical-and-local model is what blew apart the Western humanoid cost stack and is what makes Unitree the structural anchor for the 2026–2027 humanoid bill-of-materials trade. ## Where Unitree sits in the supply chain Unitree is the OEM. Upstream it touches four layers — each one a separately investable bucket. **Rare-earth feedstock and magnets.** A humanoid carries 40–50 actuators (every joint, every finger, every wrist), each wound on NdFeB sintered magnets. That is roughly 10–20× the rare-earth content of an EV per unit. The listed ex-China feedstock layer is $MP, $USAR, $ALOY, $LYC.AX, $ILU.AX, $UUUU, $TMC. JL Mag and Ningbo Yunsheng dominate inside China. **Servo motors and BLDC actuators.** Unitree designs its own motors but the named upstream actuator supplier across Chinese humanoid OEMs (Unitree, UBTech, AGIBot) is $300124.SZ Inovance — China's #1 servo-motor maker at roughly $22B mcap. Tuopu Group (601689.SS, not yet on Macroplane) supplies rotary actuators and dexterous-hand assemblies. The Western incumbents Unitree underbids: $6594.T Nidec (BLDC and coreless motors) and $MOG.A Moog (aerospace-grade actuators). **Strain-wave reducers — the precision bottleneck.** Cycloidal and harmonic gearboxes are the joint-precision moat. Global incumbent: $6324.T Harmonic Drive Systems (Japan). $6268.T Nabtesco is the cycloidal RV leader for the heavier joints. $SHA.DE Schaeffler is the bearings and planetary-screws layer. Chinese challenger Leader Drive (688017.SS, not on Macroplane) ships at roughly a third of Harmonic Drive pricing and is Tesla Optimus's stated Mexico supplier. Unitree builds its own planetary gearboxes in-house, sidestepping this layer for low-ratio joints — but the rest of the humanoid market still buys from Harmonic, Nabtesco, and Schaeffler. **Sensors and perception.** Unitree self-develops its lidars and depth cameras, which is what flipped its sensor cost vs. competitors. The merchant lidar layer — relevant for any humanoid OEM that does not verticalize — anchors on $HSAI Hesai. Magnetic position sensors and IMUs for joint feedback: $ALGM Allegro, $MELE.BR Melexis, $6762.T TDK (the latter owning InvenSense for IMUs). **Edge compute.** $NVDA (Jetson SoCs and GR00T VLA models — Nvidia is also one of Unitree's largest research-platform customers, alongside Apple and Meta) and $QCOM (RB6 robotics platform) supply the on-board brain. ## The edge Three things make Unitree structurally different from Western humanoid OEMs. 1. **In-house motors at 30–40% of Western prices.** Unitree designs and builds its own BLDC motors and the low-ratio planetary gearboxes. That removes the Harmonic Drive / Maxon margin from the BOM on every joint that does not need strain-wave precision. 2. **Shenzhen-PRD proximity to suppliers.** Prototype samples arrive next-day; iteration cycles are weeks, not quarters. Western humanoid teams sit 6–18 months behind on every BOM revision. 3. **Quadruped scale paid for the humanoid R&D.** Unitree shipped Go2 quadrupeds for $1,600–$2,800 from 2021 onward, generating cash and validating the motor/gearbox stack at scale before pivoting that stack into the G1. The result: G1 ships at ~$27,300 with an estimated 67% gross margin. Tesla Optimus and Figure are still iterating sub-1,000-unit BOMs at numbers Western chains cannot match. ## The risks - **IPO timing.** A $608M Shanghai STAR raise into a humanoid sector with little revenue and concentrated research-platform buyers (Nvidia, Apple, Meta — hundreds of units each) is priced on growth, not earnings. Any slip past the 10,000-unit industrial deployment threshold compresses the multiple immediately. - **Export controls on rare earths and lidar.** Both flow upstream into the BOM. A symmetric move (China cutting magnet exports, US restricting Jetson-class compute) hurts Western humanoid OEMs harder than Unitree, but Unitree still loses access to its Nvidia, TI, and ADI dependencies. - **$TSLA Optimus and AGIBot Walker.** Optimus has Tesla's manufacturing depth; AGIBot has Chinese capital and is already shipping Walker units. Both compress Unitree's margin path even where Unitree wins on cost. - **Verticalization cuts both ways.** Building its own motors and lidars means the structural cost edge is real, but the dependency on the merchant layer shrinks — meaning the upside for $ALGM, $MELE.BR, $HSAI from "Unitree adoption" is muted vs. the upside from Optimus, Figure, and Apptronik. ## The catalysts - **Shanghai STAR Market IPO pricing and first-day trading.** Sets the comparable for every public humanoid pure-play including KOID (KraneShares Humanoid ETF) and KSTR (STAR Market ETF). - **G1 industrial-deployment milestones.** Sub-$30/hour tote-handling labour cost was the stated viability threshold; scaling past 10,000 units is the next. - **$300124.SZ Inovance humanoid revenue split.** Earnings will start breaking out the humanoid contribution from EV/industrial servo motors — the cleanest listed proxy for Unitree volume. - **$6324.T Harmonic Drive order book commentary.** If the Chinese gearbox displacement plays out, this is where the squeeze first shows up. The picks-and-shovels of Unitree are not Unitree shares — they are the rare-earth feedstock, the listed Chinese servo-motor suppliers, the Japanese gearbox incumbents under cost pressure, and the Western sensor IC vendors who win on every humanoid that is not vertically integrated. Not investment advice; do your own research.
The Unitree Robotics IPO Thesis: China's $608M Humanoid Bet and the Supply Chain Behind It thesis on Macroplane focuses on Shenzhen Inovance Technology Co., Ltd. (300124.SZ), Harmonic Drive Systems Inc. (6324.T), Tesla, Inc. (TSLA).
It covers Rare Earth Elements, Permanent Magnets, Servomotors, Harmonic Drives / Precision Reducers, LiDAR, Sensors & MEMS, Edge Computing, Actuator Subsystems, EV OEMs, Industrial Robotics, and 1 more.
Browse all AI supply-chain theses · Macro trends · Industries · Product categories · Baskets