Chips, batteries, and the things you cannot buy

Europe's dependency on foreign supply chains for critical tech — and what happens when supply stops

By VastBlue Editorial · 2026-03-26 · 24 min read

Series: The Chessboard · Episode 2

The Part You Cannot Make

Somewhere inside every European car rolling off a production line in Wolfsburg, Stuttgart, or Munich sits a microcontroller — a chip roughly the size of a fingernail, manufactured on a process node that the semiconductor industry considers mature, even old. It is not a cutting-edge processor. It does not run artificial intelligence workloads or render graphics. It manages the engine timing, or controls the anti-lock braking system, or coordinates the power window motors. It costs between one and five euros in normal times. It is the kind of component that purchasing managers at automotive OEMs order in bulk on annual contracts, barely thinking about it. It is, in the language of supply chain management, a commodity.

In the autumn of 2020, this commodity vanished. The sequence was prosaic in its mechanics and devastating in its consequences. COVID-19 lockdowns had suppressed automotive demand in the first half of the year, and car manufacturers — following the lean, just-in-time inventory philosophy that had governed automotive supply chains since Toyota systematised it in the 1970s — had cancelled or deferred their chip orders. Simultaneously, demand for consumer electronics surged as hundreds of millions of people began working from home, buying laptops, webcams, monitors, and networking equipment. Foundries that had been producing automotive microcontrollers reallocated their capacity to higher-margin consumer chips. When automotive demand recovered faster than expected in late 2020, the manufacturers went back to their chip suppliers and discovered that the capacity was gone. Not reduced. Gone. Allocated to other customers, booked for months ahead, with lead times stretching from the usual twelve weeks to over fifty.

The consequences rippled through the European automotive industry with a brutality that exposed decades of strategic negligence. Volkswagen, the largest car manufacturer in Europe, cut production at its Wolfsburg headquarters and idled plants across Germany, Spain, and Portugal. Stellantis shut down lines at facilities in France, Italy, and Slovakia. BMW rationed chips across its model range, prioritising higher-margin vehicles while delaying production of entry-level models. Daimler, Renault, Volvo — every major European automaker was affected. The industry lost an estimated seven to eight million vehicles in global production in 2021 alone. In Europe, the financial impact was measured in tens of billions of euros. Factories designed to operate continuously, employing thousands of workers each, sat idle because a component costing less than a cup of coffee could not be sourced at any price.

The chip shortage was not a surprise. It was a consequence so predictable that it had been predicted — repeatedly, specifically, and in writing. The European Commission's own assessments of strategic dependencies, published as early as 2017, had identified semiconductor supply concentration as a critical vulnerability. Industry analysts had warned for years that the automotive sector's reliance on just-in-time ordering for components with months-long manufacturing lead times was a structural mismatch that would eventually produce exactly this outcome. The warning signs were visible to anyone who bothered to look at the geography of semiconductor manufacturing. But looking at the geography of semiconductor manufacturing requires confronting a fact that European industry and European policymakers have been reluctant to confront for thirty years: Europe does not make chips.

The Geography of Fabrication

The semiconductor supply chain is the most geographically concentrated critical industry on earth. It is more concentrated than oil, more concentrated than rare earth mining, more concentrated than pharmaceutical precursor chemistry. At every stage of the value chain — from design tools to raw wafer production to fabrication to advanced packaging — a handful of companies in a handful of countries control chokepoints that the rest of the world cannot route around.

Begin with fabrication, the most visible chokepoint. The Taiwan Semiconductor Manufacturing Company — TSMC — fabricates approximately 90 per cent of the world's most advanced semiconductors, defined as chips manufactured on process nodes of 7 nanometres or below. This is not a market share figure that fluctuates meaningfully from year to year. It is a structural reality that reflects two decades of compounding capital investment. TSMC's most advanced fabrication facility, Fab 18 in Tainan, contains extreme ultraviolet lithography machines that cost over €150 million each and can pattern features smaller than a virus onto silicon wafers with a precision measured in individual atoms. Building a comparable facility from scratch would require five to seven years and an investment on the order of €20 billion — and even then, the facility would lack the accumulated process knowledge, yield optimisation expertise, and supply chain relationships that TSMC has built over decades.

Samsung, based in South Korea, fabricates most of the remainder of advanced chips. Intel, the American champion, has been attempting to re-enter the advanced foundry business with investments exceeding $100 billion, but has repeatedly delayed its most advanced process nodes and currently trails TSMC by at least one full generation in manufacturing capability. China's most advanced foundry, SMIC, can produce chips on roughly a 7-nanometre equivalent process — impressive given the export controls it operates under, but two to three generations behind the leading edge.

But fabrication is only part of the picture. The deeper chokepoint — the one that even TSMC depends on — is lithography. The machines that pattern circuits onto silicon wafers are manufactured by exactly one company on earth: ASML, headquartered in Veldhoven, the Netherlands. ASML's extreme ultraviolet lithography systems are arguably the most complex machines ever built by human beings. Each one contains over 100,000 components, a tin-droplet plasma light source that generates EUV radiation at a wavelength of 13.5 nanometres, and optical elements polished to a flatness measured in individual atoms — supplied by Carl Zeiss in Oberkochen, Germany, one of only two companies in the world capable of manufacturing optics at this precision. A single EUV system weighs approximately 180 tonnes, requires multiple 747 freighter loads to transport, and takes months to install and calibrate. ASML delivers roughly 50 to 60 of these machines per year to the world's leading foundries. Without them, advanced chip manufacturing is physically impossible.

This creates a peculiar strategic geometry. Europe does not manufacture advanced chips. But Europe manufactures the only machine that can manufacture advanced chips. ASML is a European company, dependent on European optics from Zeiss and European precision engineering from dozens of suppliers across the Netherlands, Germany, and beyond. The irony is precise: Europe controls the most critical chokepoint in the semiconductor value chain and yet remains dependent on Asian foundries for the output of its own machines. It is as if a country manufactured every oil drill on earth but imported all of its petroleum.

The Battery Problem

If semiconductors represent a dependency that Europe allowed to develop through three decades of strategic inattention, batteries represent a dependency that Europe is creating in real time, with full awareness, while spending billions of euros trying to prevent it.

The arithmetic of the European Green Deal demands the electrification of transport. The EU's 2035 ban on new internal combustion engine vehicles — confirmed despite fierce lobbying from German and Italian automakers — means that Europe's automotive industry must transition entirely to battery electric vehicles within a decade. This transition requires lithium-ion batteries. Enormous quantities of them. The European Battery Alliance, launched by the Commission in 2017, estimated that Europe would need approximately 550 gigawatt-hours of annual battery cell production capacity by 2030 to meet projected demand for electric vehicles alone — before accounting for grid-scale energy storage, which could double or triple that figure.

As of early 2026, Europe's operational battery cell manufacturing capacity stands at approximately 120 gigawatt-hours per year. The gap between where Europe is and where it needs to be is not a rounding error. It is a chasm that is widening, not narrowing, as demand projections are revised upward. And the capacity that does exist is overwhelmingly dependent on Asian technology, Asian equipment, and Asian-controlled raw material supply chains.

The supply chain for a lithium-ion battery cell begins in mines, and the geography of those mines is as concentrated as the geography of chip fabrication. Lithium — the element that gives the battery its name — is extracted primarily from brine deposits in Chile, Argentina, and Australia, or from hard-rock spodumene mines in Australia and, increasingly, Africa. China controls approximately 65 per cent of global lithium refining capacity, meaning that even lithium mined in Australia or South America is predominantly processed in Chinese chemical plants before it can be used in battery cathodes. Cobalt, used in several common cathode chemistries, is extracted overwhelmingly from the Democratic Republic of Congo — which produces over 70 per cent of global supply — and refined predominantly in China. Nickel, graphite, manganese: the story repeats with minor variations. For virtually every critical battery material, the mining may happen in diverse locations, but the refining and processing is concentrated in China to a degree that makes diversification a decade-long project, not a policy announcement.

The cell manufacturing process itself is dominated by a handful of Asian companies: CATL and BYD from China, LG Energy Solution and Samsung SDI from South Korea, and Panasonic from Japan. These companies have spent a decade scaling production, driving down costs through process optimisation and vertical integration, and building relationships with the mining companies that supply their raw materials. European battery start-ups — Northvolt in Sweden was the most prominent — have struggled to compete. Northvolt, despite securing over €13 billion in investment and €50 billion in customer orders, filed for bankruptcy in November 2024 after repeated production delays, quality problems, and cost overruns at its flagship gigafactory in Skellefteå. The failure was not caused by a lack of ambition or funding. It was caused by the sheer difficulty of scaling a complex electrochemical manufacturing process from laboratory to gigafactory scale while competing against incumbents who had spent fifteen years learning how to do exactly that.

Northvolt's failure was not the failure of a company. It was a data point about the cost of arriving late to a manufacturing race where the leaders have a fifteen-year head start on process learning, supply chain integration, and yield optimisation.

Editorial observation

The European Commission's response to the battery challenge has followed a pattern that is by now familiar: identify the dependency, express alarm, announce a strategy, allocate funding, and discover that the structural advantages of the incumbent suppliers are deeper and more durable than any policy instrument can quickly overcome. The European Battery Alliance has produced Important Projects of Common European Interest — IPCEIs — that have channelled billions of euros in state aid to battery manufacturing and research. Individual member states have offered additional incentives: Germany subsidised CATL's gigafactory in Erfurt and Tesla's battery production at its Grünheide facility; Hungary attracted investments from CATL, Samsung SDI, and Eve Energy; France and Germany jointly supported ACC, the Automotive Cells Company backed by Stellantis, Mercedes-Benz, and TotalEnergies. The result is a patchwork of factories, joint ventures, and subsidy agreements that are gradually adding capacity — but not fast enough, not cheaply enough, and not with sufficient vertical integration to alter the fundamental dependency on Asian-controlled supply chains.

When Supply Becomes a Weapon

The question that transforms supply chain dependency from an economic inconvenience into a strategic emergency is simple: what happens when a supplier decides to stop supplying? This is not a hypothetical question. It has been answered repeatedly in the last five years, and the answers have been instructive.

In July 2023, China announced export controls on gallium and germanium — two metals that are essential for compound semiconductors used in 5G telecommunications equipment, military radar systems, fibre optic infrastructure, and satellite communications. China produces approximately 80 per cent of the world's gallium and 60 per cent of its germanium. The export controls required Chinese producers to obtain government licences before exporting these materials, effectively giving Beijing a veto over who received them and in what quantities. The controls were widely interpreted as a retaliatory measure against US and Dutch restrictions on the export of advanced semiconductor equipment to China, but their impact was felt globally. European buyers found themselves competing for reduced supply allocations, scrambling to identify alternative sources, and discovering — as they had with lithium and cobalt — that alternative sources did not exist at anything close to the required scale.

In December 2023, China extended export controls to graphite — the primary anode material in lithium-ion batteries. Synthetic graphite, which offers superior performance characteristics for EV battery applications, is produced overwhelmingly in China, which controls approximately 90 per cent of global anode material processing. The timing was notable: the controls coincided with a period of rapidly increasing European demand for battery materials as the continent's gigafactory plans entered their construction and commissioning phases. European battery manufacturers suddenly faced the prospect that the material they needed to fill their new factories might not be available — or might be available only at prices set by a government that had demonstrated its willingness to use supply restrictions as a geopolitical instrument.

The Russian invasion of Ukraine provided an even more dramatic demonstration of supply chain weaponisation, albeit in the energy rather than the technology domain. Europe's dependence on Russian natural gas — which accounted for approximately 40 per cent of EU gas imports before the invasion — was understood intellectually as a strategic risk long before February 2022. Studies had been published. Warnings had been issued. Diversification had been discussed at every European Council meeting for a decade. When the supply was actually cut, the speed and severity of the economic impact exceeded the worst-case scenarios that most policy models had contemplated. Gas prices increased by a factor of ten. Industrial energy costs rendered entire sectors temporarily uncompetitive. Fertiliser plants, glass factories, aluminium smelters, and chemical plants across Europe reduced output or shut down entirely. The energy crisis demonstrated something that economic models often underestimate: the difference between understanding a dependency in theory and experiencing it in practice is the difference between reading about drowning and being underwater.

The semiconductor and battery supply chains are not identical to the gas supply chain, but the structural pattern is the same. A critical input, sourced from a concentrated geography, controlled by actors whose strategic interests may diverge from Europe's. The specific scenarios vary — a Chinese embargo on critical minerals in the context of a Taiwan crisis, a disruption to TSMC fabrication due to military conflict or natural disaster, a trade policy escalation that restricts the flow of battery cells or battery-grade chemicals — but the underlying vulnerability is constant. Europe consumes what it cannot produce, from suppliers it cannot replace, on timelines that do not accommodate disruption.

The European Chips Act: Subsidy as Strategy

The European Chips Act, adopted in September 2023, represents the most ambitious attempt in European industrial history to reverse a strategic dependency through public investment. The Act mobilises €43 billion in public and private investment with the stated goal of doubling Europe's share of global semiconductor production to 20 per cent by 2030. It combines three pillars: a research and innovation pillar centred on the existing Chips Joint Undertaking, a "security of supply" pillar that relaxes state aid rules to permit direct subsidies for semiconductor fabrication facilities, and a crisis response pillar that gives the Commission emergency powers to intervene in the chip supply chain during shortages.

The centrepiece of the Act is the subsidy mechanism, and the centrepiece of the subsidy mechanism is Intel's proposed fab complex in Magdeburg, Germany. Intel's original plan, announced in March 2022, called for two advanced fabrication modules producing chips on Intel's most advanced process nodes. The German federal government committed approximately €10 billion in direct subsidies — roughly one-third of the project's estimated €30 billion total cost. It would have been the largest single subsidy commitment to a private enterprise in German industrial history. The project was positioned as the cornerstone of European semiconductor sovereignty — proof that Europe could attract leading-edge chip manufacturing through determined policy action and public investment.

In September 2024, Intel announced an indefinite pause on the Magdeburg project. The company was in the midst of a financial crisis — declining revenue, failed process technology transitions, and mounting losses in its foundry division. Intel's CEO, Pat Gelsinger, was under pressure to cut costs, and the Magdeburg fab, which would not generate revenue for years, was an obvious candidate for deferral. The suspension was presented as temporary, contingent on improved financial conditions, but few analysts expected the project to resume on anything close to its original timeline. The €10 billion subsidy commitment — the single largest bet in Europe's semiconductor strategy — was effectively frozen.

TSMC's fabrication facility in Dresden, announced in partnership with Bosch, Infineon, and NXP under the European Semiconductor Manufacturing Company (ESMC) joint venture, represents a more cautious but potentially more durable European investment. The facility will produce chips on relatively mature process nodes — 12 to 28 nanometres — targeting the automotive, industrial, and IoT markets where European companies have significant existing demand. The total investment is approximately €10 billion, with the German government contributing roughly €5 billion in subsidies. The first chips are expected in 2027.

The Dresden fab is strategically sensible precisely because it is modest. It does not attempt to replicate TSMC's leading-edge capability. It targets the mature-node chips that caused the 2021 automotive crisis — the exact components that European industry demonstrated it could not live without. But it also illustrates the limits of the subsidy approach. Even with €5 billion in public money, the Dresden fab will produce a fraction of Europe's total chip consumption. It does not address the dependency on leading-edge chips for AI, high-performance computing, or advanced telecommunications. And it relies on TSMC — a Taiwanese company operating in an increasingly precarious geopolitical environment — to provide the manufacturing expertise, process technology, and operational management. Europe gets a factory on European soil, but the knowledge and the strategic dependency travel with the operator.

The Priority Queue

There is a question that European industry has been reluctant to ask aloud, because the answer is uncomfortable: when supply is constrained, who gets priority? The chip shortage of 2021 provided an empirical answer. TSMC, faced with more demand than it could fill, made allocation decisions based on a hierarchy that reflected its own strategic and commercial interests. Apple, TSMC's largest customer by revenue, received its full allocation. Qualcomm, MediaTek, and other major mobile chip designers received close to their contracted volumes. AMD and Nvidia, designing chips for the surging data centre market, were prioritised. The automotive microcontroller customers — companies like Infineon, STMicroelectronics, and NXP, which designed the chips but relied on TSMC and other foundries to manufacture them — found themselves at the back of a queue that extended beyond the horizon.

The prioritisation was rational from TSMC's perspective. A leading-edge Apple chip generates far more revenue per wafer than a mature-node automotive microcontroller. A customer ordering millions of advanced mobile processors per quarter is strategically more important than a customer ordering a few hundred thousand legacy-node microcontrollers. The foundry's capacity allocation algorithm optimises for revenue and strategic relationship value, not for the downstream economic impact of unfilled orders. The fact that an unfilled order for two-euro automotive chips could idle a factory producing forty-thousand-euro cars was not TSMC's problem. It was Europe's problem.

The priority queue extends beyond commercial relationships into geopolitical ones. In a crisis scenario involving Taiwan — whether a military conflict, a blockade, or a coercive campaign short of war — the question of who receives TSMC's remaining output becomes a question of alliance structures, defence agreements, and strategic alignment. The United States has invested over $52 billion through the CHIPS and Science Act to establish TSMC fabrication capacity on American soil, in Arizona, precisely to ensure that American military and civilian chip supply is not entirely dependent on a facility located 160 kilometres from the Chinese mainland. Japan has similarly subsidised TSMC fabs in Kumamoto. Europe's equivalent investment — the Dresden ESMC facility — is smaller, later, and focused on less strategically critical chip categories.

The priority queue is the hidden architecture of every concentrated supply chain. In normal times, it is invisible — everyone gets served. In a crisis, it determines who functions and who does not. Europe has spent decades assuming it would always be served. It has not spent equivalent effort ensuring it would be prioritised.

Editorial observation

The same priority logic applies to critical minerals. When China restricts exports of gallium, germanium, or graphite, the allocation of remaining supply is determined by existing contractual relationships, strategic partnerships, and political considerations. Chinese battery manufacturers — CATL, BYD, Eve Energy, Gotion — have guaranteed access to domestically refined materials by virtue of being Chinese companies operating under Chinese jurisdiction. South Korean and Japanese manufacturers have spent years building diversified supply chains and strategic reserves precisely because they recognised the geopolitical risk of Chinese mineral dependency earlier than their European counterparts. European manufacturers, late to the game, find themselves negotiating for supply allocations from a position of structural weakness — dependent on the goodwill of suppliers who have other, more strategically aligned customers to serve first.

What Sovereignty Actually Requires

The word "sovereignty" appears with increasing frequency in European policy documents — technological sovereignty, digital sovereignty, industrial sovereignty, strategic autonomy. The word is used so often that it risks becoming meaningless, a rhetorical decoration that substitutes for the industrial capabilities it is supposed to describe. But sovereignty, stripped of its rhetorical function, has a specific and demanding meaning: the ability to produce, or reliably access, the inputs your economy requires to function, without depending on the permission or goodwill of actors whose interests may conflict with yours.

By this definition, Europe is not sovereign in semiconductors, not sovereign in batteries, not sovereign in critical mineral processing, and not sovereign in several adjacent technology supply chains — including the active pharmaceutical ingredients that underpin its healthcare system, the polysilicon and wafer capacity that underpin its solar energy transition, and the high-purity neon and other specialty gases that are essential to the lithography process that ASML's own machines require. The list of dependencies, when assembled comprehensively, is not a list of gaps. It is a description of an industrial base that has been systematically hollowed out over three decades, not by accident or misfortune, but by the rational operation of market forces that optimised for cost over resilience, efficiency over redundancy, and quarterly returns over strategic positioning.

Rebuilding genuine sovereignty — not the rhetorical kind, but the kind that survives a supply chain disruption — requires investments that are measured in decades and tens of billions of euros per sector. A complete semiconductor fabrication ecosystem, from wafer production through advanced packaging, would require cumulative investment on the order of €100 billion over fifteen years, plus the training of a workforce that currently does not exist in sufficient numbers. A battery supply chain that is genuinely independent of Chinese refining would require building mineral processing capacity that takes five to ten years to permit and construct, in jurisdictions where environmental regulations — rightly — impose costs and timelines that Chinese competitors do not face. A critical minerals strategy that provides genuine resilience would require either the development of European mining and processing operations — politically difficult in a continent where mining is associated with environmental degradation — or the construction of a diversified network of bilateral agreements with resource-rich countries in Africa, Latin America, and Central Asia, competing against China's two-decade head start in exactly these relationships.

None of this is impossible. All of it is slow. And the gap between the speed at which dependencies can materialise and the speed at which they can be resolved is the gap that European policymakers must learn to operate inside. A chip fabrication facility takes five to seven years to build. A lithium refinery takes four to six years. A cobalt processing plant, three to five years. A mining operation, from exploration to first ore, seven to fifteen years. Each of these timelines assumes a favourable regulatory environment, available workforce, secured financing, and stable political support — assumptions that are individually reasonable but collectively demanding. In the time it takes to build one European lithium refinery, China can commission ten.

The European Critical Raw Materials Act, adopted in 2024, sets targets for domestic extraction, processing, and recycling of strategic materials. By 2030, the EU aims to extract at least 10 per cent of its annual consumption of strategic raw materials domestically, process at least 40 per cent domestically, and recycle at least 25 per cent. These targets are calibrated to be achievable without requiring the kind of mining expansion that would trigger intense local opposition, but they are also calibrated at levels that leave the majority of supply still sourced from outside Europe, still subject to the geopolitical risks that the Act was designed to mitigate. The Act is a step. It is not a solution.

The honest strategic assessment, which European policy documents approach but rarely state plainly, is this: Europe cannot achieve full supply chain independence in critical technologies within any timeline that matters for the current geopolitical cycle. The dependencies are too deep, the investment requirements too large, the timelines too long, and the political will — measured in sustained, multi-decade commitment rather than announcement-day enthusiasm — too uncertain. What Europe can do is reduce the severity of its dependencies, diversify its supplier relationships, build strategic reserves, invest in substitution research, and develop the diplomatic and economic leverage to ensure that when supply is constrained, Europe is not last in the priority queue.

This is not the narrative that policymakers prefer. It lacks the decisive clarity of a sovereignty declaration. It does not photograph well. But it reflects the operational reality of a continent that has spent thirty years outsourcing its industrial foundations and now has perhaps ten years to build enough domestic capacity to avoid the worst consequences of that outsourcing. The chips and batteries that Europe needs are being manufactured, right now, in facilities controlled by companies and governments whose strategic interests are not aligned with Europe's. The question is not whether Europe can change this. The question is whether it can change it fast enough — and what it is prepared to sacrifice in order to try.