Colman Deegan, CEO at Zayo Europe, discusses how AI is reshaping European network infrastructure, driving regional connectivity expansion and increasing the need for strategic investment and collaboration.
From towers and small cells through to backhaul and front-haul, fibre connectivity is in demand. Global networks and their associated infrastructure have consistently grown over the years. In 2010, the total amount of digital data generated, captured, and consumed globally was roughly two zettabytes. Today, that number has seen a 110x increase to 220 zettabytes annually.
The fact that data needs have grown isn’t something to be surprised by. The growth of cloud adoption, enterprise digitisation, online services, as well as general internet traffic has seen consistent growth year on year. And it’s that consistency that has made it relatively predictable when it comes to scaling and bandwidth requirements.
Operators could plan around this organic growth – expanding networks and increasing capacity. Data Centre (DC) ecosystems naturally developed to address this growth. As a result, major digital hubs were established, becoming increasingly dense and interconnected.
AI’s effect on network demand
But now we’re seeing a step change in the market. With McKinsey reporting that AI adoption has jumped from 50% to 88% in the past four years, the prolific use of this technology is naturally having an impact on network infrastructure. While broader markets may be arguing about AI disruption, when it comes to networks, it’s more a matter of pace.
AI hasn’t replaced those existing drivers for network demand. Instead, AI, and the much broader growth in compute that comes with it, has materially changed the scale of the wireless infrastructure needed. The amount of compute being deployed is increasing, as has the capacity required to connect it. When the level of demand changes, the infrastructure required to support that demand changes also.
Regionalised connectivity
To meet this demand, DC supply has become more regional, which, in turn, will regionalise connectivity. When it comes to AI, the technology is not just changing how much wireless infrastructure markets need, but it’s geographically changing the shape of the market.
In Europe, for example, the continent’s digital infrastructure landscape has historically developed around a very established set of centres. Europe’s digital infrastructure landscape developed around a very established set of centres of gravity: London, Frankfurt, Paris, Amsterdam, and Dublin (the FLAP-D).
To cope with demand, networks are increasingly looking outside these markets to establish data centres, enable edge computing and meet the demand of AI. As such, regions like Madrid and Milan are emerging as new hubs for data and connectivity.
However, this does not mean that FLAP-D markets are declining – quite the opposite. Significant demand continues to be concentrated in those markets – especially for major enterprises such as financial services, tech and media. These sectors will continue to require enormous amounts of connectivity – especially given FLAP-D markets also serve as major commercial locations for international business.
Compute follows power; connectivity follows compute
The next phase of networks is best viewed as the expansion of existing hubs rather than a decline of traditional metros. Demand has shifted to broader areas adjacent to these established markets. Evidence for this can already be found in key markets. Between August 2025 and April this year, 91% of incremental capacity in the London market was in Greater London rather than Central London. Factors such as power availability, land and permitting are determining where large-scale compute can be delivered. These constraints are subsequently pushing supply to make a wider regional footprint, which means that network requirements will follow.
In the case of London, this means extending the metro’s presence, but it also means connecting regional campuses such as Hertfordshire & Bedfordshire. This enables the UK to tap into AI growth zones and leverage hubs that are supporting gigawatts of AI compute. Of course, London is just one example, and the trend being seen may differ in scale and location. However, the underlying dynamic remains the same: the intersection of power, land and permitting is creating a broader map of demand.
For network operators, this broadens the addressable scope. Operators still need density, resilience and customer access in established urban cores, but they also need regional routes that connect peripheral data centre clusters, new compute campuses and adjacent metros back into those core hubs.
The investment challenge
This is a trend that can be seen across the whole of Europe. Central metros are being increasingly connected to regional clusters, data centres connecting to data centres, and regional compute being linked into established cloud and enterprise ecosystems. This trend goes beyond creating more capacity, but becomes a broader point around having the right architecture across a larger geography.
Naturally, the resulting question that comes from this trend is: how can investment for this connectivity be secured? Demand may be there, but on the supply side, the scale of investment required to meet this need is significantly higher than before. As the market is moving away from a relatively predictable, incremental network expansion towards larger infrastructure programmes, the cost inevitably rises. These programmes can require tens, if not hundreds, of millions of pounds.
As such, operators need to rethink supply. History has shown that speculative building is not an efficient use of resources, while overbuilding existing infrastructure can lead to a significant waste of capital. Instead, to meet the increase in investment, the next phase of expansion will need to be collaborative and disciplined. The best way to serve the market will be a combination of having a strong existing footprint and strict capital deployment. A strong asset base gives operations options such as existing routes to extend from and established data centre connectivity. These options allow operators to respond to new demand faster and with more precision.
This flexibility will not remove the need for more investment, but it will make it easier to identify the best way to deploy capital. Operators can focus on where investment will create the greatest value and avoid overinvesting. Disciplined investment is not simply about operators spending less, however. Instead, it’s about ensuring that operators avoid duplication and increase value. In some cases, the best strategy will be to build, but other times it will be to acquire, lease, use wholesale or work with a partner.
Long-lasting partnerships
It’s that sense of partnership and collaboration that will truly show value in the future of networks. Better collaboration on infrastructure sharing can help the industry respond faster and derive better returns on investment. This shouldn’t be to the detriment of market competition but more about cooperation on the supply side. If operators can partner on the infrastructure layer, they can compete on the service layer.
Edge computing is also affecting how partnerships are changing. New workloads, sites and use cases are increasing, meaning operators now have a challenge in deciding who they partner with. The rapid growth of edge computing means that operators simply can’t partner with everyone, and this naturally means strategic partnerships will be a huge consideration going forward.
With demand increasing at the pace being seen, there will be no shortage of opportunities to invest. However, if operators strike out on their own, the result will simply be undisciplined duplication, which will only destroy value before the market has even matured. Investments that expand reach, improve resilience and strengthen the customer proposition – especially when approached collaboratively – will see the greatest returns.
Wireless infrastructure is going through a significant step change in the market. Established hubs, while still a priority, are now widening their footprint and becoming increasingly interconnected. As a result, investment is spiking, leading to a rethink in how operators think about engaging with one another. Competition will continue, but it will be more focused on the service layer rather than infrastructure.
Please note, this article will also appear in the 28th edition of our quarterly publication.
