Leuven Before the Fab

Why the world’s most advanced chips begin in imec’s cleanrooms

In global discussions about semiconductors, the focus tends to drift toward factories, supply chains and geopolitical leverage. Attention goes to where chips are manufactured, who controls production capacity and how nations secure access to critical technologies. Yet these debates often overlook a more fundamental question: where do future chip technologies actually originate?

The answer leads not to a megafab, but to Leuven, Belgium. There, imec operates as one of the world’s most influential semiconductor research institutes. Largely invisible to the general public, imec plays a defining role in shaping the technological possibilities that the global industry will work with years from now. Its cleanrooms do not produce commercial chips. Instead, they function as a kind of intellectual foundry, where ideas are tested long before they become products—or are abandoned altogether.

From Hardware to Understanding: The ASML Relationship

The recent opening of the High-NA EUV Lab in Veldhoven underscored Europe’s ambition to remain at the forefront of advanced lithography. ASML’s High-NA extreme ultraviolet systems represent a new generation of patterning technology, essential for extending chip scaling further into the future. Yet while the most advanced hardware experiments increasingly take place in Veldhoven, the deeper work that determines whether High-NA EUV will succeed unfolds elsewhere.

That work happens in Leuven. At imec, researchers focus on the less visible but more decisive layers of the problem: materials behavior, photoresist chemistry, pattern variability, defect mechanisms and the integration of lithography into complete manufacturing flows. These questions are not about whether a machine functions, but whether an entire process ecosystem can be made stable, scalable and economically viable.

The collaboration between ASML and imec is therefore not simply about access to tools. It is about shared understanding. imec provides a pre-competitive environment where equipment suppliers, chipmakers and designers collectively explore the limits of physics and manufacturing. In doing so, it absorbs risk that no single company could—or would—carry alone.

Silicon Photonics and the Limits of Electricity

As transistor scaling slows and system complexity increases, bottlenecks increasingly emerge outside the transistor itself. Data movement, rather than computation, has become one of the dominant constraints on performance and energy efficiency. This is particularly visible in cloud infrastructure and data centers, where electrical interconnects consume vast amounts of power and generate significant heat.

imec’s work in silicon photonics addresses this challenge at a structural level. By enabling data transmission through light rather than electrical signals, silicon photonics offers a path toward faster and significantly more energy-efficient communication within and between chips. The promise is not incremental improvement, but a rethinking of how digital systems are architected.

What makes imec’s approach distinctive is its insistence on integration. Photonics is not treated as a parallel technology, but as something that must coexist with advanced CMOS logic, memory and packaging techniques. Research therefore focuses not only on optical components themselves, but on how photonics can be manufactured at scale, integrated into existing process flows and aligned with the economic realities of the semiconductor industry.

For hyperscalers and cloud providers, these developments are closely watched. If successful, silicon photonics could reshape the physical and economic foundations of the cloud over the coming decade, enabling higher performance while curbing the relentless growth in energy demand.

Beyond Scaling: Rethinking the System

While node numbers still dominate headlines, imec’s research agenda increasingly reflects a broader realization: scaling alone is no longer enough. As physical limits approach, progress depends on new system-level strategies. This includes three-dimensional integration, advanced packaging and heterogeneous architectures that combine logic, memory, analog, RF and photonic components into tightly coupled systems.

At the same time, imec explores computing paradigms that depart from conventional architectures altogether. Neuromorphic and brain-inspired approaches, for example, aim to deliver drastic gains in energy efficiency for artificial intelligence workloads. These efforts are not driven by immediate market demand, but by the expectation that today’s computing models will struggle to scale sustainably in the long term.

Crucially, imec works on time horizons that most commercial organizations cannot afford. Much of its research targets technologies that may only mature five to ten years from now. The value lies not in certainty, but in reducing uncertainty—by identifying what is physically plausible, what is manufacturable and what is unlikely to succeed.

Leuven Through Silicon Valley’s Eyes

imec’s relevance is reflected in the composition of its partner ecosystem. Companies such as Intel, TSMC, Samsung and Apple maintain deep research collaborations with the institute. From Silicon Valley’s perspective, Leuven functions as an early-warning system for technological change, offering insight into future directions long before they appear on public roadmaps.

This has led to frequent comparisons between Leuven and Silicon Valley. Yet the resemblance is limited. imec is not a start-up factory, nor is it driven by venture capital dynamics. Its strength lies in long-term, capital-intensiveand deeply scientific research. Its influence is subtle but structural, shaping the options available to the global industry rather than competing within it.

Strategic Neutrality in a Fragmented World

As semiconductor supply chains become increasingly politicized, imec occupies an unusual and strategically important position. It is European in location and governance, global in participation and explicitly pre-competitive in mission. This allows it to serve as a neutral ground where collaboration remains possible, even as geopolitical tensions rise elsewhere.

For policymakers, technology leaders and institutions concerned with digital sovereignty, imec represents something increasingly rare: a place where technological futures are explored before they harden into dependencies. What happens in imec’s cleanrooms does not make headlines, but it quietly shapes what the world will be able to build.

In that sense, imec does more than research technology. It researches the space of future possibilities—methodically, collaboratively and always several years ahead of the market.