While AI discussions often focus on GPUs and capacity, the technology cannot scale without high-performance connectivity. But apart from AI, connectivity is becoming a strategic asset, driven by increasing data volumes, resilience requirements, and sovereignty.
This blog provides key takeaways from a panel discussion at Level X featuring Tom De Clercq (Belnet), Andreas Herden (nLighten), and Kris Warreyn (Eurofiber).
As AI is reshaping network infrastructure, Belgium is strengthening its position as a strategic connectivity hub. What can we expect from the next generation of digital infrastructure connecting data centres, AI, cloud, and edge environments?
1. Belgium is the new sweet spot
Belgium has long been one of Europe’s crossroads for fibre infrastructure, but it has historically not been regarded as a Tier-1 data centre location. Surrounded by major Internet Exchange Points (IXPs), network traffic was often routed via Amsterdam or other neighbouring hubs before reaching a data centre in Brussels. These indirect routes increased latency and, over the years, created inefficiencies.
That is now changing. Belgium is emerging as a strategic hub for AI-driven digital infrastructure. Investments around Brussels, Antwerp, and Ghent are accelerating, while demand is shifting away from more saturated markets. Rather than experiencing a sudden boom, Belgium is seeing steady, scalable growth, supported by its central location and strong connectivity.
2. AI is changing network requirements
As AI adoption accelerates, the demands placed on network infrastructure are changing. While low latency remains essential for certain applications, AI workloads introduce a new challenge: moving and processing enormous volumes of data. As a result, bandwidth and scalability are now just as important as speed. Networks must be capable of handling continuous growth in traffic.
Today, research networks are planning for massive capacity increases, from hundreds of gigabits to multiple terabits per second within the next decade.
3. Edge computing brings AI closer to users
In addition to AI, there is a growing shift towards edge computing, where data is processed closer to users, devices and businesses. This reduces the distance data needs to travel, enabling faster and more responsive applications.
For data centres, this means that demand is no longer concentrated in a handful of major hubs. Instead, regional and edge data centres are becoming an essential part of the landscape. Rather than replacing centralised infrastructure, edge computing creates a more distributed ecosystem in which hyperscale facilities and regional data centres work together to deliver reliable services.
4. Connectivity should be treated as a strategic asset
Rather than being an operational necessity, connectivity is rapidly becoming a strategic asset. This shift is driven by the growing importance of resilience, data sovereignty, and security. For many organisations, particularly in highly regulated sectors, keeping data local has become a critical requirement. For example, hospitals and government agencies are increasingly seeking data centres that are European-owned and operated to ensure greater control over sensitive data.
The LevelX panel also expects continued growth in the use of private fibre networks and dedicated connections, enabling external data centres to operate as an extension of an organisation’s own infrastructure. At the same time, resilience is no longer achieved through simple redundancy alone. As digital infrastructure becomes increasingly business-critical, organisations are adopting multiple physically diverse network routes to minimise risk, improve business continuity, and strengthen overall operational resilience.
5. The infrastructure behind the network must evolve too
As AI workloads grow, the infrastructure supporting connectivity must evolve alongside. In fact, increasing bandwidth is only part of the challenge. Higher-performance networking equipment requires significantly more power and a new approach to data centre design.
Data centre operators are already preparing for higher rack densities and the widespread adoption of liquid cooling to support AI-ready infrastructure. At the same time, networking equipment itself is becoming more power-hungry, with next-generation routers and switches pushing the limits of existing facilities.
Therefore, connectivity is about much more than cables or fibre. Delivering the next generation of AI services requires an ecosystem in which networks, data centres, cooling systems and power infrastructure all evolve together.
Do you want to learn more? Click here to watch the full panel discussion at Level X.
