On the same Wednesday that Unitree's shares surged on their Shanghai debut, Germany's engineering lobby put out a statement that read less like market commentary than a demand. Humanoid robotics and physical AI, the VDMA said on August 19, must move to the top of the political agenda, and Germany and Europe must build resilient domestic supply chains for the critical parts that go inside the machines.
The association, which represents mostly mid-sized mechanical and plant engineering firms, framed it as a race that has already started. Its named remedy came from Hartmut Rauen, VDMA's deputy executive director.
“Germany and Europe must establish a comprehensive value chain featuring domestic production of critical components,” Rauen said, in remarks reported by Reuters.
The VDMA added that the crucial challenge lies in scaling. That is the sentence doing the most work. Europe has labs, precision engineering firms and one very well funded humanoid startup. What it lacks is volume, and volume is where component sourcing, unit cost and political leverage get decided.
What the ask actually contains
Read closely, the August 19 statement is a position, not a policy paper. It does not name a subsidy instrument, a procurement quota, a local-content threshold or a standards body. It asks for attention and for domestic production of critical components, which in Brussels terms is closer to a request for a strategic-technology designation than a costed programme.
That vagueness is itself informative. VDMA's underlying analysis is more concrete. Its Future Business study Humanoid Robotics 2040, published April 28, projects billions in annual revenue and unit sales in the millions by 2040, and explicitly compares the emerging sector to the automotive industry in scale. A body that believes a car-industry-sized market is forming will ask for car-industry-sized policy attention, and the component supply chain is the obvious pressure point.
It is a real pressure point. A humanoid robot is mostly joints, and joints are mostly actuators, harmonic drives, precision planetary and cycloidal reducers, ball and roller screws, bearings and rare-earth permanent magnets. Japan's Harmonic Drive Systems and Nabtesco have long dominated the high-precision reducer segment, with Harmonic Drive alone estimated to hold around 60 percent of the global harmonic drive market. China processes roughly 90 percent of the world's permanent magnets and, by supply-chain analysts' teardown estimates, supplies the large majority of component value in leading humanoid platforms. Chinese reducer makers are now taking share in the mid-precision tier on price.
Europe is not absent from that list. Schaeffler builds bearings, planetary roller screws and rotary actuators. Bosch supplies sensors and motion control. But holding component positions is not the same as holding a value chain, which is precisely Rauen's point.
Europe is not starting from zero
The counter-example sits in Metzingen, in Baden-Wuerttemberg. On June 10, NEURA Robotics announced a Series C of up to 1.4 billion dollars, described as the largest round ever raised by a full-stack robotics company. Tether led, with Qualcomm, Amazon, NVIDIA, imec.xpand, Bosch, Schaeffler, Lingotto Horizon, InterAlpen Partners and the European Investment Bank participating. Reports put the valuation near 7 billion dollars. NEURA says its orderbook and deployment pipeline exceed 1 billion dollars.
The EIB's presence is the politically interesting part.
“By backing NEURA Robotics, the European Investment Bank is putting serious European firepower behind the next wave of physical AI and cognitive robotics,” said Nicola Beer, an EIB vice president, citing the bank's TechEU programme and Europe's technological autonomy.
NEURA is also building capacity on the training side. In March it launched the TUM RoboGym with the Munich Institute of Robotics and Machine Intelligence at the Technical University of Munich, a roughly 17 million euro joint investment across 2,300 square metres at the TUM Convergence Center near Munich Airport, with NEURA contributing about 11 million euros.
Does industrial policy of this kind work?
The honest answer is: sometimes, and rarely quickly.
The case for it is strongest exactly where VDMA is pointing. Precision reducers and magnet production are capital-intensive, long-cycle, learning-curve businesses where incumbents are hard to dislodge and where a decade of patient capital genuinely changes outcomes. Europe already has the metallurgy, the machine tools and the safety-certification culture. Public co-investment of the EIB variety, alongside strategic corporates like Bosch and Schaeffler, is a defensible use of a state balance sheet.
The case against is Europe's own recent record. The bloc's battery push produced a high-profile insolvency and a great deal of announced-but-unbuilt capacity, and its chip ambitions have moved slower than the headline targets implied. Component plants built without a guaranteed order book become stranded assets. There is also an obvious interested-party problem: the organisation asking for the policy represents the firms that would receive it.
And there is a demand-side gap that nobody in the August 19 statement addressed. China's humanoid scaling is being pulled along by domestic buyers and state-linked demand. Europe has no equivalent anchor customer and no public procurement channel for humanoids. Supply-side subsidies without demand-side commitments tend to produce capacity nobody buys.
What to watch
Three things. First, whether Berlin or the Commission converts VDMA's framing into a named instrument. Robotics is a plausible candidate for the Commission's applied-AI and advanced-manufacturing workstreams, and the wording will matter more than the euro figure. Second, whether Bosch and Schaeffler follow their NEURA equity with actual European actuator and reducer capex, which would signal that the private sector believes the demand case. Third, the regulatory calendar: the EU Machinery Regulation applies from January 20, 2027, and AI Act obligations for high-risk AI embedded in machinery bite by August 2, 2028. Certification could become Europe's genuine differentiator, or its handicap, depending on how Brussels writes the guidance.
The next hard number to watch is shipment share. If Europe's slice of global humanoid units still rounds to zero when the 2027 figures land, the VDMA will be making this argument again, with less time left on the clock.
“Germany and Europe must establish a comprehensive value chain featuring domestic production of critical components.”— Hartmut Rauen, Deputy Executive Director, VDMA