An inventory of the West's supply-chain dependence
As Donald Trump arrives in Beijing today — the first visit by a US president to China in almost nine years — a structural finding is becoming public that has lain in plain sight for more than a decade and yet gone unnoticed. The West has lost its strategic depth to China. What began as a decision for economic efficiency has turned into a geopolitical dependence that breaks every response of the Western democracies on its material preconditions. This working paper examines four fields — solar, rare earths, semiconductors and others — and asks how, over four decades, rational individual decisions gave rise to a structure that nobody wanted and that nobody can now correct without damage. It adds to this diagnosis a second dimension that has come to the fore in the past two years: the operational controllability of Chinese components even after their installation in Western infrastructure — a qualitatively different class of threat that cannot be removed by diversification alone.
For four decades, the global division of labour was regarded as an unquestionable good. Whoever extracts polysilicon where energy is cheap, cuts the wafer where wages are low and produces the cell where economies of scale take hold, lowers prices for everyone. In the course of a decade the solar cell thus became eighty-five per cent cheaper. That was a real achievement. It was celebrated. It made possible the energy transition, which would not have been possible without this fall in prices.
What nobody factored in: each single one of these location decisions was rational in the narrow sense. The buyer at a German module manufacturer was looking for the cheapest supplier. The shareholder of an American semiconductor group wanted to see quarterly results. The European regulator upheld ecological and social standards without asking where production would move to if it were made more difficult only within his own jurisdiction. The sum of all these rational individual decisions was a historically unprecedented event — the voluntary transfer of the industrial base of an entire civilisational region to another.
Nobody decided this. It was built the way a coral reef grows — through millions of individual acts, each small and reasonable in itself. That the result was not a reef but a dependence was seen only when one called into it and the answer did not come back.
The following findings come from current data surveys by the International Energy Agency, the European Commission, the European Court of Auditors and private market observers such as Wood Mackenzie and Enerdata. They are not spectacular. They are technical and sober. That is precisely what gives them their force.
Solar. China controls more than eighty per cent of global production capacity at every stage of the solar supply chain. Polysilicon, wafer, cell, module — the same dominance everywhere. For polysilicon capacity under construction, the Chinese share reaches almost ninety-five per cent. Forty per cent of the world's polysilicon production takes place in a single province, Xinjiang. One in seven modules worldwide leaves a single factory. The ten largest module manufacturers in the world are all Chinese — JinkoSolar supplies 93 gigawatts a year, Longi 78 gigawatts, Trina and JA Solar 77 gigawatts each. Between 2011 and 2021 China invested more than fifty billion dollars in PV production capacity — ten times as much as Europe over the same period. In 2023 alone, one hundred and thirty billion dollars flowed into the sector.
Rare earths and permanent magnets. For nineteen of twenty strategically important minerals, China controls seventy per cent of global processing capacity. For magnet rare earths — neodymium, praseodymium, dysprosium, terbium — the figures are about sixty per cent of mining output and ninety per cent of magnet production. Ninety-one per cent of the European Union's rare-earth magnets come from China. In April 2025 Beijing introduced export controls on seven heavy rare-earth elements. Car manufacturers in the US and Europe had to cut back production or shut down briefly. In October 2025, Announcement No. 61 brought a second escalation — a licensing requirement with global reach. If a magnet produced in Germany contains even 0.1 per cent Chinese rare-earth material, it falls under Chinese export control. This rule was suspended for one year in November 2025 — until 10 November 2026. Today's Trump–Xi summit will also decide what happens after that date.
Semiconductors. TSMC in Taiwan produces more than sixty-five per cent of the world's advanced semiconductors and ninety per cent of the chips used to train modern AI models. For the most advanced processes of all — five nanometres and below — the figure is almost one hundred per cent. This production takes place on a few square kilometres in Taiwan, less than one hundred and fifty kilometres from the Chinese coast. CIA Director William Burns said publicly in 2024 that a military disruption of TSMC's operations would be the greatest economic damage to the world economy since the Second World War.
Others. The list of concentrations is long and extends far beyond the three fields named. China dominates the lithium-ion battery value chain (CATL and BYD being two of the world's largest manufacturers), wind-turbine production (Goldwind, Envision, Mingyang) and global shipbuilding (51 per cent market share). Of the EU's magnesium supply — critical for hydrogen electrolysis — ninety-seven per cent comes from China. For gallium seventy-one per cent. For germanium forty-five. For graphite forty. The ten largest suppliers of solar manufacturing equipment are all Chinese — a double dependence, for whoever wants to build a new solar factory in Europe buys the tools from the very place whose competition he wants to challenge.
For four decades, supply-chain dependence was theoretical. Nobody had played it as a card. It was a worst case in the risk reports. In the last twelve months it has stepped out of the risk reports and migrated into operational reality.
April 2025: China imposes export controls on seven rare-earth elements. Consequence: car factories in the US and Europe cut production. It is the first concrete demonstration that the theoretical possibility can be put to use.
October 2025: Announcement No. 61 extends control to every component containing Chinese material or Chinese technology, in every country of the world — the zero-point-one-per-cent rule with extraterritorial reach. Here it becomes clear that supply-chain dependence is not merely a bottleneck but a lever of sovereignty. A power that dominates supply chains can deprive the other power of control over its own industry — not through war, not through sanctions, but by setting licence conditions.
November 2025: The October escalation is suspended for one year — as part of the Trump–Xi deal at the APEC summit in Busan. Suspension, not repeal. The blade stays in its sheath, visible.
February 2026 to today: the US–Iran war, the closure of the Strait of Hormuz, the volatility of energy prices. While the Western economies suffer from the energy shortage, China falls back on its stock of solar and wind power. What was debated in the West as climate policy had for ten years been quiet geopolitical precaution in China.
Today, 13 May 2026: Trump lands in Beijing. He does not come as the man who nine years ago wanted to keep China small. He comes as the man who asks Xi to mediate on Iran, who needs Chinese ships to secure the Strait of Hormuz, whose magnet supply hangs on the negotiating table.
Up to this point we have been speaking of a static dependence. If someone does not deliver, someone else must take their place — or supply breaks off. This logic applies to polysilicon, magnets, semiconductors. It underestimates, however, a second layer of entanglement that has come to the fore in the last two years: the operational controllability of Chinese components even after their installation in Western infrastructure.
In May 2025, US energy officials discovered undocumented communication modules in Chinese solar inverters — cellular radios that were not listed in any official product description. They allow remote access that bypasses the installed firewalls. An incident in November 2024 had shown that this capability is not theoretical. In the US and other countries, solar inverters were deactivated from within China — as a side effect of a commercial dispute between the manufacturers Sol-Ark and Deye. Controllability has been demonstrated, not merely hinted at.
Huawei and Sungrow together control more than half of the global inverter market. In Europe, more than 200 gigawatts of installed solar capacity are connected to Chinese inverters — the output of more than 200 nuclear power stations. What a climate policy began has become a constellation of leverage points.
The pattern repeats itself in other fields.
Telecommunications. Huawei and ZTE in European 5G networks. In January 2026 the EU Commission turns its recommendation, non-binding since 2020, into a legally binding obligation to exclude these suppliers. Nevertheless, Italy, Spain, Poland, Portugal, Austria and Greece continue to use Chinese equipment. Germany has a two-stage exit plan running to 2026 and 2029. Bart Groothuis, the European Parliament's rapporteur on cyber security, states the structural root precisely:
It is not a secret backdoor in hardware or software. You are always just one software update away from a new backdoor. The real problem is the ties these companies have to the Chinese state.
China's National Intelligence Law of 2017 obliges every Chinese actor — company, citizen, research institution — to cooperate with the intelligence services. This makes the question does Beijing want to use this device for espionage or sabotage legally secondary. The answer is: Beijing can order it at any time, and the company must comply. This legislation is the structural foundation, independent of the concrete intentions of any given Chinese government.
Port cranes. A congressional investigation in 2024 found cellular modems on Chinese port cranes made by ZPMC that were not documented in any contract with the US ports and served no official function. ZPMC controls about 80 per cent of all container cranes in US ports. In February 2024 Biden earmarked 20 billion dollars in an executive order for replacing these cranes. In the same month the Port of Virginia confirmed that it expected further ZPMC cranes by August 2025. This shows the operational depth of the entanglement — even after official recognition of the problem, the ports go on ordering, because there is no alternative.
Electric vehicles. Modern electric cars are, as the European Council on Foreign Relations puts it, platforms for mobility, permanently networked with cameras, microphones, GPS and data transmission. Chinese manufacturers such as BYD, NIO and Dongfeng potentially collect their drivers' data and transmit it to Chinese cloud systems. The British Special Boat Service in Poole has already banned MG, BYD, Volvo and Polestar from its site. In September 2024 the US Department of Commerce proposed excluding Chinese connected vehicles from the US market.
Drones. DJI controls about 70 per cent of the American hobby drone market and up to 90 per cent of the commercial market. The threat is twofold: the drones map the infrastructure they fly over, and they can be deactivated during a conflict. In December 2025 the FCC put all foreign drones on the Covered List. In March 2026 routers and home network devices followed — TP-Link, Netgear, Linksys, Asus, D-Link.
Other devices. The original Reuters investigation explicitly mentions that the same undocumented communication components were also found in Chinese electric-vehicle charging stations, heat pumps, battery storage units and smart-meter systems. That means: every device of Chinese manufacture connected to the power grid belongs structurally to the same risk class.
Several observations follow from this finding.
First, controllability is not a theoretical residual risk. It was operationally demonstrated in November 2024 with the inverters. Mike Rogers, former director of the National Security Agency, describes the strategic situation: We know that China believes there is value in putting at least parts of our core infrastructure at risk of destruction or disruption.
Second, controllability is a qualitatively different class of threat from supply-chain dependence. A static dependence can be mitigated by stockpiling, recycling or diversification. A controllability dependence does not disappear through diversification. A Chinese inverter installed in Germany remains a potential point of remote control for its twenty-year service life — even if, from tomorrow, not a single Chinese inverter is delivered to Europe any more.
Third point: the regulatory thresholds obscure the problem. In November 2024 Lithuania blocked remote control of energy installations above 100 kilowatts by Chinese suppliers. But hundreds of thousands of smaller household installations remain below the threshold — and add up to an order of magnitude that is decisive for grid stability. An attack on just 3 gigawatts of networked inverters, SolarPower Europe warned in a report in May 2025, could have significant effects on the power system.
Fourth, the response lags unevenly. While the US acts with FCC bans and Lithuania, Estonia and Latvia have taken tough steps, Italy, Spain, Poland and Portugal refuse. Strand Consult estimates the one-off cost of completely replacing Huawei radio equipment in Germany at 29 euros per inhabitant — roughly one or two monthly mobile phone bills. It is not primarily a question of cost. It is a question of decision and implementation.
The European Union has not overlooked the problem. As early as 2024 it adopted the Critical Raw Materials Act. On 4 March 2026 this was supplemented by the RESourceEU initiative. Three non-binding targets were set for 2030: at least ten per cent of annual consumption of strategic raw materials is to be extracted in the EU, forty per cent processed there, and twenty-five per cent obtained from recycling. No more than sixty-five per cent may come from a single non-EU country.
Three billion euros are being mobilised over the next twelve months. Most of this, however, comes from pots that were already planned and are now running under a new label — two billion euros in EIB loans, three hundred million from the Battery Booster initiative, seven hundred million from the Innovation Fund. Genuinely new money is limited.
In April 2026 the European Court of Auditors publishes its report. The self-diagnosis is sober. Despite fourteen new trade agreements, despite the Critical Raw Materials Act, despite RESourceEU, the EU is unlikely to succeed in reaching the diversification targets by 2030. Dependence has not fallen since the law was adopted. It has grown.
Stéphane Séjourné, Executive Vice-President of the European Commission, states the strategic situation precisely: Dialogue with China remains essential. In diplomatic language that means: there is no other choice. At the same time the US is building up its own raw-material stockpiles and pursuing decoupling against Europe as well.
They are trying to stockpile raw materials from the US and elsewhere. For us it is important to have the right organisation and financial instruments to be effective in derisking.
Séjourné, technical briefing, December 2025. The Europeans can rely neither on China nor on the US. They are sitting between two stools — and over the last thirty years they have economised away their own seat.
Here the diagnosis becomes sharper. The Western elites who today talk about decoupling and resilience are themselves the beneficiaries of the supply-chain constellation that they now present as a threat. Apple built its global market position on Chinese manufacturing depth. Tesla depends on Chinese battery suppliers. The German car industry stabilised its profitability by sourcing intermediate products from China. The pharmaceutical industry obtains its active ingredients largely from Chinese sources.
These elites do not really want to decouple. They want rhetorical distance from the Chinese dependence, but the operational business. They want to talk about sovereignty and at the same time keep the margins that are only possible through globalisation. This double bind explains why the political initiatives — Critical Raw Materials Act, Chips Act, Inflation Reduction Act — produce a great deal of rhetoric but little structural change.
The Court of Auditors' report does not say it outright, but it describes it: when the diversification targets are made non-binding, when the funding consists largely of relabelling, when implementation is delegated to the member states, then the policy is designed in the format of a declaration of intent, not in the format of a strategy. It is meant to appear effective without becoming effective. It prolongs the status quo by ending it rhetorically.
This is a repetition of what was described in the working paper on Deutsche Bahn as the administration of decline — the diagnosis is clear, the capacity for reform is missing. Interlocking policy-making across twenty-seven member states, Commission and Parliament produces movement without change. What appeared at the railway as a national symptom shows itself here as a continental structural feature.
In the working paper on the railway, the constellation of universal Megamachine logic and national incapacity for reform was described as negative world leadership — a special German position at the bottom of a table whose top was once the goal. What was analysed there as a national phenomenon has since shifted to the global level.
At the global level it becomes apparent that not only the railway but the entire Western industrial landscape has lost its substance to the balance sheet. What for forty years was booked as an efficiency gain was at the same time the sale of strategic depth. What stood in the textbooks as comparative advantage was the licence for the voluntary disarmament of one's own industrial base.
In the same period China did something different. It built up strategic depth while we were giving ours away. It pursued industrial policy while we declared industrial policy obsolete. It deployed patient capital while we prioritised quarterly results. It took the Megamachine — Mumford's social machine that organises our production — into its own hands and directed it, while we understood it as a naturally grown market order.
The result is a world in which the Western democracies claim rhetorical hegemony but have lost material sovereignty. Trump in Beijing is the symbolic condensation of this situation. He comes with threats of sanctions and leaves with requests. He brings the tariffs, which Xi turns round with the rare-earth lever. He demands submission and receives a trade board as a face-saving procedure.
What Maria Stöhr put pointedly in her leading article in DER SPIEGEL — The MAGA man is making China great — is not an observation about Trump but an observation about the structural constellation in which Trump is merely the most recognisable embodiment. Whoever operates a Megamachine that lies in foreign hands cannot free himself by being loud.
First. The strategic depth of an industrial society is not an economic category. It is the material precondition for a society to remain capable of acting in crises. Whoever gives it up for short-term margin optimisation sacrifices not a quarterly figure but the precondition of his own policy.
Second. The West's present supply-chain dependence is not the result of a conspiracy but the sum of rational individual decisions over forty years. That is precisely where its tenacity lies — it has no originator who can be called to account. It is emergence, not a plan.
Third. The European reform response has the format of a declaration of intent, not of a strategy. The targets are non-binding, the funding largely relabelled, the implementation delegated. The result is movement without change. In April 2026 the Court of Auditors confirms what was foreseeable as early as 2024: the targets for 2030 will not be met.
Fourth. Beyond static dependence, a second dimension has emerged — the operational controllability of Chinese components after their installation. It is not theoretical; it was demonstrated in November 2024 with the deactivation of solar inverters from within China. This second layer cannot be removed by diversification alone. Every Chinese device already installed remains a potential point of remote control for its service life — even if, from tomorrow, no comparable device were delivered any more.