Photonics Beyond the Hype: Where Europe Actually Stands Today
Why light is no longer a future promise, but a present infrastructure choice
From future promise to present constraint
For years, photonics has been presented as a technology of the future — elegant, powerful and perpetually just over the horizon. It appears in research agendas and innovation strategies alongside quantum computing and other long-term breakthroughs. As a result, it is still widely perceived as experimental rather than structural.
That framing is becoming increasingly misleading. The renewed relevance of photonics is not driven by scientific novelty, but by a far more pragmatic force: the physical limits of electronic infrastructure. Modern digital systems are no longer constrained by how fast chips can calculate, but by how much energy is required to move data between them — inside servers, across racks and throughout data centers.
At that point, the discussion shifts from performance to physics. Digital infrastructure, ultimately, is a physical system. And every physical system encounters limits — not economic limits, not regulatory limits, but thermodynamic ones.
“The amount of energy required to move data around inside a data center is becoming a larger problem than the computation itself. Optical interconnects are no longer a luxury; they are the only way to break through the ‘power wall’ of modern AI clusters.”
Ian Young
Senior Fellow and Director of Exploratory Integrated Circuits, Intel
The interconnect bottleneck: where electronics begin to fail
For decades, progress in computing was driven by making transistors smaller, faster and more energy-efficient. That trajectory is now colliding with a different constraint. While individual chips continue to improve, the energy cost of moving data between them is rising sharply.
In modern systems, electrons are pushed through increasingly dense copper interconnects at ever higher speeds. At scale, this results in resistance, signal degradation and heat dissipation that grow non-linearly with bandwidth. The consequence is counterintuitive but critical: adding more compute capacity no longer delivers proportional system-level gains.
This is the interconnect bottleneck. It is not a software problem, nor a design oversight. It is a physical limitation of using electrons as information carriers at scale. Once this bottleneck is reached, optimisation within the same paradigm yields diminishing returns.
Light behaves differently. Photons have no mass, generate minimal heat during transmission and can carry multiple data streams simultaneously via different wavelengths. Under the conditions that increasingly define modern infrastructure — high bandwidth, short distances, dense integration — optical interconnects are not an alternative. They are the only physically viable path forward.
“Moore’s Law for electronics is hitting physical limits. We are transitioning from an era of ‘faster electrons’ to an era of ‘smarter photons’. Photonics is the fundamental shift needed to keep global digitalisation sustainable.”
Ton Backx
Emeritus Professor and former Director, Institute for Photonic Integration, Eindhoven University of Technology
What photonics is — and what it is not
At this point, a common misconception needs to be addressed. Photonics is often conflated with the idea of a fully “optical computer” — a machine where light replaces electronics entirely. That vision, while scientifically intriguing, is neither the immediate goal nor the primary value proposition.
Photonics excels in two domains: data transport and specific forms of computation, particularly linear algebra operations that underpin artificial intelligence and signal processing. For logical decision-making — the if-this-then-that operations that define general-purpose computing — electronic transistors remain highly efficient.
The future, therefore, is not replacement but integration. Hybrid architectures combine electronic logic with photonic transport and acceleration. This marriage does not aim to outperform electronics on their own terms, but to change the system assumptions under which digital infrastructure operates.
“Light is the ultimate medium for information. It has no mass, creates minimal heat when moving and can carry vastly more data than any copper wire. It is the natural successor to the electron for everything involving scale.”
Michal Lipson
Professor of Electrical Engineering and Applied Physics, Columbia University
The three phases of photonics adoption
Seen through this lens, the development of photonics follows a clear, realistic trajectory.
Phase one: communication. Optical fibre, wavelength-division multiplexing and coherent optics have formed the backbone of global communications for decades. This phase is mature, industrialised and indispensable.
Phase two: interconnects and data centers. Optical links are now moving inside data centers — from rack-to-rack, board-to-board and increasingly chip-to-chip. This is where photonics directly affects energy consumption, cooling requirements and architectural flexibility. This phase is happening now.
Phase three: photonic computing. Optical accelerators for specific workloads, particularly AI, are emerging. These systems are still developing, but strategically unavoidable as data-intensive applications continue to scale.
Data centers and the energy wall
The importance of phase two cannot be overstated. Data centers are no longer a marginal concern of the IT sector. They are becoming central actors in national energy systems. In several European countries, new data center capacity is constrained not by capital or demand, but by grid availability.
This is the “energy wall”: a point at which digital growth collides with electrical infrastructure limits and climate targets. Under these conditions, efficiency gains are no longer incremental improvements — they determine whether expansion is possible at all.
Photonics addresses this constraint at the system level. By reducing electrical interconnects, it lowers heat generation. By enabling new architectural layouts, it offers flexibility in how compute, memory and networking are physically organised. In this context, photonics is not an optimisation. It is an enabling technology for keeping digital infrastructure within energy and climate boundaries.
Europe’s position: architect of hardware sovereignty
Europe’s role in this transition is often underestimated. While global attention focuses on AI platforms, data and algorithms, Europe holds a structurally different advantage: control over critical hardware technologies.
The continent’s strength lies in materials science, photonic integration, manufacturing equipment and system engineering. Nowhere is this more visible than in the Netherlands, where a dense ecosystem spans research, design, fabrication and tooling.
“The Netherlands holds a unique position globally. We do not only possess the knowledge, but the entire value chain: from the design of photonic chips to the machines that manufacture them. What is being built here is the foundation of the next digital generation.”
Ewit Roos
CEO, PhotonDelta
This position has geopolitical implications. Technological sovereignty is not determined solely by who owns the software stack, but by who controls the physical layer that makes digital systems possible.
“If you want to have technological sovereignty, you must master the physical layer. Photonics is one of the most critical ‘deep tech’ areas where Europe can and must lead to remain competitive in the global AI race.”
Thierry Breton
Former European Commissioner for the Internal Market
From future vision to infrastructure choice
Photonics is no longer best understood as a promise of faster machines tomorrow. It is an answer to the constraints of digital infrastructure today. As electronic systems approach their physical limits, light offers a different set of assumptions — about energy, scale and architecture.
The question, therefore, is no longer whether photonics will matter. The question is which parts of Europe’s digital infrastructure can continue to function efficiently without it.
