Deutsche Bank Report: China's Humanoid Robot Production May Reach 100,000 Units by 2026
0xBroomberg
Deutsche Bank's latest report shows Chinese government officials estimate 100,000 humanoid robots will be produced in 2026 — 2.5× the bank's own prior forecast; both the U.S. and China are accelerating, opening an investment window across components and full machines.
What does 100,000 units mean — and why is it so far above Deutsche Bank's own estimate?
Deutsche Bank previously forecast about 40,000 units for China in 2026. The government's latest figure is 2.5× higher.
This means → capacity expansion is outrunning sell-side models, and supply-chain procurement timelines may need recalibrating.
In plain terms = even the most bullish analysts underestimated how fast China is moving.
Both sides are accelerating — where does each stand?
U.S. side: Agility Robotics plans a SPAC listing in Q4 2026; Meta acquired an embodied-AI model startup; OpenAI is hiring robotics engineers; Nvidia is expanding its robotics team in China.
Agility's parts localization rate has hit 75%, with a target to cut per-unit bill of materials from $125,000 to $30,000. This means → U.S. full-machine makers are shifting from "lab demo" to "cost accounting" mode.
China side: Unitree is ramping mass production; BYD and Li Auto disclosed robotics progress; Alibaba released a robotics model; Kepler is in the process of being acquired. Unitree and Figure AI have live-streamed robots working in factories and warehouses, validating commercial viability.
What are governments betting on?
Shanghai: deploy 100,000 humanoid robots by 2030. Japan: deploy 10 million AI robots by 2040.
The U.S. may impose export restrictions on Chinese robotics products. This reflects that humanoid robots have entered the arena of great-power industrial competition — no longer just a tech story.
UBTECH and DOBOT have launched bionic humanoid robots with emotional companionship features — in plain terms = use cases are expanding from factory logistics to home services.
Which names does Deutsche Bank flag — and what is the investment logic?
Components (Asia-Pacific): Hengli (buy, closing price RMB 110.58), ShuangHuan Driveline (buy, RMB 42.39), Harmonic Drive Systems (buy, ¥7,430), Yaskawa Electric (buy, ¥5,490).
Full machines (U.S.): Tesla (buy, $394.46), Mobileye (buy, $9.43 — exposure via Mentee Robotics).
This means → Deutsche Bank's logic is clear: a doubling of output pulls component suppliers first; full-machine brands come second.
Blackstone is betting too — what does the smart money see?
Blackstone has taken a major position in Futronic, an auto-parts supplier that has pivoted its motion-control technology — the core tech that makes robotic arms move precisely — into humanoid-robot actuators.
This reflects top-tier PE's judgment: the most certain value in humanoid robots lies not in "building the whole machine" but in core components like motion control — consistent with Deutsche Bank's picks.
In plain terms = no matter which brand wins, the parts sellers make money. Classic "selling shovels" logic.
Can 100,000 units actually be delivered — what should investors watch?
Whether China's 100,000-unit production target is met on schedule is the key checkpoint for validating this acceleration thesis.
Unitree and peers have entered mass-production phase, but moving from "demonstrating feasibility" to "100,000 units per year" still carries capacity-ramp uncertainty.
This means → for investors, actual 2026 shipment data will determine whether this supply chain is "delivering on expectations" or "front-running them."
Content is for reference only, not financial advice.