The EU’s deep integration into global supply chains has delivered substantial economic benefits. However, recent crises and trade tensions have revealed that reliance on foreign suppliers can pose serious risks, especially for essential goods. Disruptions in key regions have cascaded through supply chains, limiting access to critical inputs, driving up costs, and halting production. Recent EU policy initiatives – including the Competitiveness Compass, the Critical Raw Materials Act, and the Chips Act – all emphasise the need to identify and address such dependencies. Doing so requires answering three questions: Which products are truly at risk? Which dependencies are structural? And which supply chains can realistically be diversified?

A growing body of research addresses these questions. The European Commission (2021) identified 390 highly dependent products. Arjona et al. (2023, 2025) refined the analysis by correcting trade data for re-exports and assessing recent shifts towards partners with trade agreements. Méjean and Rousseaux (2024) added a domestic-absorption measure, Vicard and Wibaux (2023) refined the time dimension, and Borin et al. (2025) used firm-level data to detect vulnerabilities and quantify exposure to external risks, such as sudden supply disruption from Chinese suppliers. However, two key gaps remain. First, previous studies proxy domestic production with export data, which can misrepresent actual capacity. Second, they largely ignored the geopolitical context: a dependency only becomes a strategic risk if the supplier’s political alignment, institutional stability, and willingness to weaponise trade make the EU vulnerable. 

Our work developed in Bonnet et al (2026) fills both gaps by developing a multi-layer analytical framework that combines improved dependency identification, explicit geopolitical risk assessment, and an understanding of structural constraints on diversification. 

Layer 1: What is new in the identification of foreign-dependent products?

We define foreign-dependent products as those for which the EU relies on a limited number of suppliers and has limited domestic production capacity. This is captured by combining import concentration (Herfindahl–Hirschman Index) and exposure (the share of extra-EU imports in total EU supply). Our approach introduces three innovations. First, we use Eurostat Prodcom data to measure domestic production directly, avoiding biases from using exports as a proxy. Second, we determine thresholds for identifying ‘dependent’ products through cluster analysis rather than arbitrary cut-offs, allowing the data to guide classification. Third, we track dependencies over time (2017–2023), distinguishing structural from temporary dependencies. As a matter of fact, a dependency that has persisted for years without adjustment is far more urgent than one that has only recently emerged.

China is the EU primary supplier and has doubled its influence in renewable energy

Applying this approach identifies 835 foreign-dependent products in 2023, accounting for 16% of EU imports and 17% of their value across five strategic ecosystems: energy-intensive industries; digital; aerospace & defence; health; and renewable energy.

China is the EU primary supplier for 47% of dependent products and roughly half of their total import value (€206 billion out of €404 billion). While its role has stabilised or declined in most sectors, it has strengthened markedly in renewable energy, where the share of products dependent on China has doubled from 40% to 80%, driven by its dominance in photovoltaics.

The US is the second-largest supplier, accounting for 11% of import value. Its role is particularly strong in aerospace & defence and health, where it supplies more than half of EU foreign-dependent imports. While its share in renewable energy has slightly declined, the US shows a gradual upward trend in the energy-intensive sector, although remaining a relatively minor supplier.

Figure 1 Dependency by sourcing country

Dependency, once established, tends to persist

The number of dependent products remains stable over time, but the key question is whether this aggregate stability conceals underlying changes. It does not; there is a strong persistence at product level: 45% of dependencies are structural, remaining present throughout 2017–2023. These include smartphones and laptops in thedDigital ecosystem, pharmaceutical products and medical instruments in health, and raw materials such as gallium and germanium in energy-intensive industries. Even among the rest, most reappear over time. In other words, once the EU becomes dependent on a product, it tends to remain so.

Layer 2: Geopolitics turns trade vulnerabilities into critical dependencies

Not all dependencies carry the same risk. Reliance on a stable, politically aligned partner differs fundamentally from reliance on a geopolitically distant or unstable one. To capture this dimension, we construct three indicators: geopolitical risk to measure political instability in sourcing countries (Caldara–Iacoviello index); political distance (UN voting alignment combined with governance quality); and exposure to export restrictions (Global Trade Alert).

Geopolitical exposure reduces 378 strategic dependencies to 206 critical ones

These are concentrated in specific ecosystems. In energy-intensive industries, dependencies such as gaseous hydrocarbons and steel derivatives are largely sourced from Russia, combining high geopolitical risk and political distance. In digital and in renewable energy, China dominates supply of key products such as smartphones, laptops, and photovoltaic cells, amplifying exposure through both concentration and geopolitical risk.

A different pattern emerges for the US. Many dependencies linked to the US are primarily exposed to trade-restriction risk, including aerospace components and industrial chemicals. While the US is politically aligned and a major supplier, it also records the highest frequency of restrictive trade measures in recent years. 

Figure 2 Country distribution by geopolitical risk measures and trade restrictions

This highlights a key point: political alignment does not eliminate trade policy friction. Even like-minded partners can weaponise trade dependencies by applying trade restrictions, while suppliers such as China combine concentration with broader geopolitical exposure, making these dependencies particularly critical. By contrast, a group of stable partners – including Canada, the UK, Japan, South Korea, and Australia – scores low across all risk dimensions and represents the most credible pool for diversification.

Layer 3: Why diversification is harder than it looks

A natural response to critical dependencies is to seek alternative suppliers. At first glance, this appears feasible. Critical products are often not technologically complex, trade relationships are not unusually persistent, and about half are downstream goods that should be easier to substitute.

However, this intuition does not hold. Network centrality analysis shows that global production is often dominated by a very small number of global exporters. The issue is not identifying alternatives in principle, but that suppliers with sufficient scale do not exist in practice. 

Even simple products can be impossible to diversify when global supply is highly concentrated, and the EU is asymmetrically dependent on them. 

This challenge is particularly acute in the renewable energy ecosystem, where supply is concentrated in a handful of countries whose output cannot be easily replicated. Importantly, the EU is not uniquely exposed; all global importers face the same structural bottleneck.

Asymmetric interdependence further amplifies risk. Where the EU depends more on a supplier than the supplier depends on the EU, bargaining power shifts away from Europe. This is most visible in digital and in renewable energy, where China dominates supply. In contrast, for some energy-intensive products sourced from Russia, exposure is partly mitigated by Russia’s own reliance on EU demand and the availability of alternative suppliers with whom the EU has trade agreements.

Policy conclusions

Not all dependencies require intervention, but for a subset of critical products, inaction is not an option. The framework developed here helps prioritise responses by combining three dimensions: exposure, geopolitical risk, and structural constraints on diversification. The results point to differentiated policy needs. Where supply is geopolitically risky and highly concentrated, diversification is limited, making stockpiling and strategic supply agreements necessary. Where alternatives exist, diversification and partnerships with reliable suppliers can reduce exposure. Over time, targeted industrial policy and innovation are needed to address structural dependence.

Persistence also matters. Dependencies that have remained unchanged for years are unlikely to resolve through market forces alone and require active policy intervention. Monitoring should therefore move beyond static snapshots to track structural vulnerabilities over time.

Finally, the role of the US deserves attention. While much of the debate focuses on China, the US is a key supplier in several strategic sectors and the most frequent user of trade restrictions in our data. Political alignment does not preclude trade friction. Recognising this is essential for a realistic EU resilience strategy.

Authors’ note: The views expressed are purely those of the author and may not in any circumstances be regarded as stating an official position of the European Commission. 

References

Arjona, R, W Connell, and C Herghelegiu (2023), “The EU’s strategic dependencies unveiled”, VoxEU.org, 23 May. 

Arjona, R, W Connell, and C Herghelegiu (2025), “EU supply chain tectonics”, VoxEU.org, 3 April. 

Bonnet, P, E Di Girolamo, A Pagano, B Velazquez, and E Zaurino (2026), “EU foreign dependencies: Assessment of trade and geopolitical vulnerabilities”, DG Trade Chief Economist Note 2, March. 

Borin, A, F P Conteduca, F Leone, M Mancini and P Zi (2025), “Hidden exposures in domestic supply chains: The spread of foreign trade risks”, VoxEU.org, 2 December. 

European Commission (2021), “Strategic dependencies and capacities”, Staff Working Document SWD(2021)352 final.

Méjean, I and P Rousseaux (2024), “Identifying European trade dependencies”, in J Pisani-Ferry, B Weder di Mauro and J Zettelmeyer (eds), Europe’s Economic Security, Paris Report 2, CEPR Press.

Vicard, V and P Wibaux (2023), “EU strategic dependencies: A long view”, CEPII Policy Brief 41.