Headlines warn that helium shortages – caused by the ongoing war in Iran and the wider region – are threatening semiconductor production. The situation is yet another example of the complexity of the global chip supply chain, and how vulnerable it is to geopolitical disruption.
That reminder was hardly needed. Some 60% of all chips, and over 90% of the world’s advanced chips – critical to so much modern technology – are produced in Taiwan, with industry leader TSMC at the forefront. Yet Taiwan is earthquake-prone and has long been under threat from China, making it a perennial supply chain concern.
With problems like these, it’s easy to see why policymakers across the world are now thinking hard about how to create more resilience on their own terms, by building chipmaking within their own sovereign ecosystems.
In Europe, that may be easier said than done.
“While Europe remains a world leader in equipment, materials, and research, its share of global chip manufacturing has fallen over the past two decades, exposing critical dependencies on foreign production,” notes a report from late last year by industry organisation SEMI.
As the EU prepares a new Chips Act, expected to be announced at the end of May, important questions remain about the bloc’s approach to boosting its semiconductor supply chain resilience.
In the wake of the pandemic and as European manufacturers reported shortages of the semiconductors they needed, policymakers started to wake up to this issue.
In Feb 2022, the EU introduced the Chips Act as a response to supply chain disruptions due to the pandemic. It entered into force in 2023, aiming to increase production capacity for the leading-edge semiconductors needed by European industries, particularly the automotive sector. With EU funding for advanced semiconductor research and a new framework for state aid for semiconductor manufacturing facilities, the overarching aim of this legislation is for the EU to produce 20% of world production by value by 2030, up from 12.7% in 2023.
Though it will take years for the impact on overall production and supply chains to show, it has clearly catalysed significant investment: a total of €69 billion in public and private investments by October 2025, according to SEMI.
Some of the results may take years to emerge, but a few major projects serve as the clearest examples of how production might increase on the continent in the years ahead. Interestingly, one theme throughout much of it is that Taiwan is figuring in a lot of the enterprises – a signal of how its industry is also looking for more supply chain optionality for the future.
Perhaps the most significant European semiconductor project right now is the new fab for ESMC – a joint venture between TSMC and three European companies – which is under construction in Dresden, Saxony and due to be completed next year. This will produce non-cutting edge chips for European industry, particularly automotive producers.
Saxony has a history of semiconductor expertise going back to the days of East Germany. It now manufactures almost one-third of Europe’s semiconductors. Saxony has also developed close links with Taiwan – so close that it has its own Science Liaison Office in Taipei.
Neighbouring Poland and Czechia are also both close partners of Taiwan in diplomacy and drone technology, and they hope to become important players in the supply chain of a theorised “Silicon Triangle.” Various initiatives offer students from different parts of Europe, particularly Saxony but also elsewhere, opportunities to study and work in Taiwan, with the aim of incubating the skills needed in a future semiconductor workforce.
Meanwhile, other projects aim to boost Europe’s capabilities in manufacturing and other parts of the supply chain. Taiwanese tech company Foxconn, alongside Thales and Radiall – two aerospace and defense industry firms – have announced plans for an outsourced semiconductor assembly and testing (OSAT) plant at an undisclosed location in France. These are expected to make more specialised chips needed for space and defence applications.
But the aim of producing 20% of the world’s chips by 2030 is widely seen as a pipe dream.
That is partly because of the staggering growth elsewhere: global semiconductor manufacturing capacity is forecast to grow 7% a year from 2024 to 2028, according to SEMI – . The European Court of Auditors concluded last year that the EU Chips Act was “very unlikely” to achieve the 20% goal, citing a forecast of 11.7% by 2030.
What flavor of chips?
A broader issue is for Europe to work out which kinds of chips it should be able to produce.
The Chips Act emphasises the importance of the most advanced chips, wanting to make sure that Europe is not left behind in the industries of the future and dependent on outsiders to be able to analyse data.
But some question the logic of this approach. Unlike the legacy chips needed by its car industry – “if it still exists” in a few years, jokes one insider with a grimace – Europe has comparatively little demand today for the cutting edge chips used in AI data centres, an area dominated by the US and China. This could be a chicken-and-egg problem, and it could also be distracting policymakers and the industry from existing chip needs.
“The EU is pushing for technological leadership: frontier AI, quantum and so on and so forth,” notes Riccardo Bosticco, a PhD Researcher at the Centre for Security, Diplomacy and Strategy (CSDS) at the Vrije Universiteit Brussel (VUB). “If the purpose is supply chain resilience, a realistic assessment of the industry would also say that if we need to secure our own industries, then the chips that we need the most are not the most advanced ones.”
According to the European Chips Survey conducted by the European Commission in 2022, three quarters of European chip demand is for 65-90nm nodes, three or four generations behind the cutting edge.
Demand for advanced chips is likely growing but remains “marginal” in comparison, as the AI race is dominated by Chinese and American companies.
In February last year, the European Commission said it would invest €20 billion to create AI gigafactories in Europe, allowing the continent to train more of its own models. That would also offer an obvious source of demand for advanced semiconductors.
But the construction has been delayed, as a key business question remains: which companies would use it?
Europe’s leading LLM company Mistral is building its own AI infrastructure, and has shown little interest in using the EU’s hardware (so far). Others have far more limited resources. Jörg Bienert, president of the German AI Association, told Science Business in October that it’s unclear whether any European firms currently have the capital to train an LLM using a gigafactory.
“The question is, who will come,” said Bienert. “I don’t see any single corporation that will be able to spend the money in order to train a large LLM on its own.”
‘Be part of the game by being in the value chain’
A higher-level question is about what Europe should be optimising for in its chip policy.
A report co-authored by Bosticco from the Chips Diplomacy Support Initiative notes that the Chips Act lays out “multiple goals—digital sovereignty, resilience, security of supply, innovation, and competitiveness—but these objectives can pull in different directions.”
The researchers discussed different possible aims for 2035, concluding that Europe should aim to make itself indispensable in the chip supply chain.
“The idea is to leverage the strengths that Europe already has,” explains Bosticco. “Europe wants to become globally relevant and less vulnerable to coercion, not by eliminating every sort of dependency that it may have, but actually reinforcing the strengths that it already has.”
Bosticco highlights IMEC, a world leading research institute in Belgium, and Zeiss, a specialised optics company, as examples of Europe’s existing strengths.
The paper he co-authored envisages moving supply chain bottlenecks out of China, combined with Europe indispensability in a supply chain with its allies. The main challenge, says Bosticco, is a political one. Getting European and other partners to sustain backing for, and cooperate on, such a strategy will not be easy.
Nevertheless, officials in Europe seem to be thinking along these lines. Franck Paris, director of the French de facto embassy in Taipei, told journalists at a briefing in February that it is “not a realistic project” for Europe to replicate the whole supply chain.
But, he said, “we need Europe to be part of the game. So we need at some stage to be indispensable in the [value] chain.”
He highlighted ASML, the ultra-specialised Dutch lithography machine maker that is essential to advanced chip making, but “it’s not enough”, he argued.
Yet even if Europe can make itself indispensable, East Asia – particularly Taiwan – remains vital to those supply chains. Any push to become less dependent on Taiwan for cutting edge chips would simply be unrealistic. “There’s no intention – and there is no actual possibility – of completely eliminating Taiwan from the supply of semiconductor chips in Europe,” says Bosticco.
As a result, Europe will continue to depend on East Asia’s stability while, the Chips Diplomacy Support Initiative paper argues, the continent has “little influence to determine the course of events in East Asia.”
