NETWORK PERSPECTIVE
behind it is starting to matter far more than most conference panels acknowledge.
AI query must be processed in a data centre, delivered across networks and often supported by compute infrastructure closer to the user. The smarter our digital services become, the more electricity the infrastructure beneath them quietly consumes. None of this means the telecom industry is powerless in the face of these trends. Network design has always evolved alongside the constraints placed upon it. Engineers are already exploring ways to reduce energy consumption through more efficient hardware or playing with the physics of Joule’s Law. But efficiency improvements can only stretch so far when the overall demands continue to grow. At some point the physical limits of infrastructure begin to reassert themselves. In my view, the next constraint on Europe’s digital future may not be spectrum, regulation or even competition. It might just be electricity. And telecom executives who often measure success in gigabits per second might want to start measuring it in kilowatt hours. Of course, energy is only one of the invisible constraints shaping Europe’s telecom future, but more on that next time.
Power availability is already starting to shape decisions about where digital infrastructure is built with hyperscale data centres increasingly located in regions where electricity is cheaper or more abundant. If we apply the same logic to edge computing we end up with curious experiments like Project Natick, where Microsoft sank a data centre 117 feet to the seabed off the coast of Scotland. Asking, “where is power available to support this workload?” rather than, “where is the user driving this workload?” fundamentally changes the concept of the network edge. In the past, digital services were often described as ‘location- agnostic’. In reality, they remain deeply tied to the location of the physical systems that power them. For telecom operators and their investors, this matters for a simple reason: electricity is no longer just another line item on the operating budget. Networks are always-on systems, and as data traffic grows, so too does the energy required to move, process and store it. In a world where power is both more expensive and potentially more constrained, electricity cannot just be a monthly bill. Instead, it must be treated as a strategic input to infrastructure economics. And artificial intelligence is only going to accelerate the trend. Training large models is famously energy intensive, but the real shift comes from everyday usage. Each
Europe also happens to be entering this phase of digital expansion at an awkward moment for its electricity market. Successive geopolitical crises have triggered price volatility and exposed diverging energy policies across European governments. At the same time, the ongoing drive for electrification of entire sectors of the economy (e.g. transportation, heating or heavy industry) means that competition is hotting up for the same resource that quietly powers every piece of equipment in the telecom ecosystem. In the coming years, electrons may start attracting as much attention as contestants on Love Island. The deeper issue is not simply the price of electricity, but the infrastructure that delivers it. Much of Europe’s power grid was designed decades ago for a vastly different economy that assumed predictable industrial demand and relatively stable consumption patterns. Today those same networks are expected to support electric vehicles, heat pumps, data centres and an increasingly digital society. Unfortunately, electrical infrastructure does not scale at the speed of software. Copper and concrete tend to move at their own pace. Slowly. As we know all too well in telecoms, the lowest common denominator is the one that eventually determines the limit for us all.
Volume 48 No.2 MAY 2026
41
Made with FlippingBook - Online magazine maker