Thursday, September 17, 2026Published from Singapore
Asia Economic ReviewAsia's economy, considered.
Home / Briefings / Policy
Briefings · Policy

Does the EU’s AI drive risk deepening external chip dependencies?

The argument

The European Union’s strategy to expand domestic artificial intelligence infrastructure risks intensifying its reliance on advanced semiconductors from external suppliers, creating a contradiction in its pursuit of technological sovereignty.

By Lena Ho1 September 20264 min read
Photo: marcinjozwiak / Pixabay

The EU's semiconductor sovereignty objective faces a core paradox

The European Union has articulated a goal of achieving technological sovereignty, particularly within the semiconductor sector. The original Chips Act, adopted in 2023, established an objective to elevate Europe’s share of global semiconductor production to 20 percent by 2030.

However, this ambition faces a fundamental contradiction as the EU simultaneously accelerates the deployment of AI factories, gigafactories, and new data centres. These initiatives generate a substantial demand for advanced semiconductors, a category of components where Europe produces less than 10 percent of the world’s supply.

The European Court of Auditors has cautioned that the 2030 production target is unlikely to be achieved, with the European Commission's own projections, issued in an undated report, forecasting a market share of approximately 11.7 percent.

To address perceived weaknesses in the initial legislation, which prioritised supply expansion, Chips Act 2.0 is expected to introduce measures aimed at stimulating demand, including public procurement tools and enhanced coordination between producers and industrial users.

Domestic AI infrastructure expansion creates significant external dependency

The scale of the EU’s planned AI infrastructure is considerable. The European Commission’s AI Continent action plan details 19 AI factories and proposals for up to seven AI gigafactories. Furthermore, the Cloud and AI Development Act aims to at least triple the bloc’s data centre capacity within five to seven years.

This expansion necessitates a substantial supply of advanced AI processors. The Center for European Policy Studies (CEPS), in an undated analysis, estimates that each planned AI factory site requires up to 25,000 advanced chips, while a gigafactory requires a minimum of 100,000.

Almost all of these processors are expected to originate from Nvidia, which supplies the majority of graphics processing units deployed in Europe and whose proprietary CUDA software forms the foundation of much of the AI software ecosystem.

CEPS warns that this reliance could establish an 'Nvidia dependency trap,' where European computing infrastructure remains technologically reliant on a single external supplier.

Recent projects illustrate this pattern: Mistral secured 13,800 Nvidia GPUs for a data centre near Paris, Deutsche Telekom’s Munich Industrial AI Cloud is being constructed with nearly 10,000 Nvidia Blackwell GPUs, and Nscale stated its Sines deployment for Microsoft will commence with over 12,600 Nvidia Blackwell Ultra GPUs, expanding to more than 66,000 by 2027.

Europe's limited control over the global semiconductor value chain

The challenge of external reliance extends beyond a single company. Even with expanded domestic semiconductor manufacturing, the global supply chain structure limits the autonomy any single region can achieve.

Toni Roldán-Monés, economist and assistant professor of public policy at IE University, noted in an undated statement that Europe relies on both the United States and Asia at different stages of the value chain. The United States holds a dominant position in chip design, intellectual property, and certain frontier equipment.

Concurrently, the manufacturing of the most advanced semiconductors is concentrated in Asia, particularly Taiwan and South Korea, while China plays a fundamental role in various materials, industrial processes, and critical minerals. This reliance is particularly evident in fabrication, where Taiwan produces approximately 90 percent of the world’s most advanced chips.

In packaging, assembly, and testing, the EU accounts for just 4 percent of the global market and remains highly dependent on Asia, according to Laith Altimime, President of SEMI Europe, in an undated comment.

Claire Godfrey, executive director of the Balanced Economy Project, observed that no top 20 assembly, test, and packaging company is headquartered in the EU, and China dominates several inputs for the semiconductor supply chain.

While Europe possesses strengths, such as ASML in extreme ultraviolet lithography systems and Imec in semiconductor research, these do not translate into autonomy across the entire semiconductor value chain, as Roldán-Monés concluded.

The pursuit of resilience over self-sufficiency and its implications for Asia

The consensus among experts is that complete semiconductor self-sufficiency is not a realistic objective. Instead, the focus should be on mitigating strategic vulnerabilities rather than eliminating international interdependence.

Laith Altimime of SEMI Europe stated in an undated comment that global collaboration is essential, as no single country can rebuild the entire supply chain. The EU’s demand-side strategy, while intended to foster domestic investment in chip design and manufacturing, will initially direct significant capital towards external suppliers.

This is particularly relevant for Asian manufacturers of advanced semiconductors. The accelerated development of AI infrastructure in Europe, driven by initiatives like the Cloud and AI Development Act and Chips Act 2.0, will sustain and potentially increase demand for high-end chips from Asian fabrication facilities.

For instance, the projected need for hundreds of thousands of advanced GPUs for AI factories and gigafactories will primarily benefit manufacturers in Taiwan and South Korea in the near term.

This ongoing reliance underscores the continued strategic importance of Asian semiconductor production for global technological ambitions through at least 2027, as demonstrated by Nscale’s planned expansion with Nvidia Blackwell Ultra GPUs in Sines.

Asian chip manufacturers should anticipate sustained order volumes from European AI initiatives, solidifying their critical role in the global technology supply chain.

This analysis is journalism, not investment advice; consult a licensed professional before making financial decisions.

Pieces are credited to the desk that commissioned and edited them. Our editorial standards, and the desks behind them, are set out on the Editorial Standards and Team pages.

Further reading