FiberCop plans to convert more than 100 Italian telecommunications exchanges into edge data centres, using the retirement of copper infrastructure to place computing capacity closer to businesses, public bodies, and regional users.

The first site is operating in Rome, while exchanges in Turin, Genoa, Bologna, Naples, and Palermo are due to follow. FiberCop has not published a complete deployment timetable, expected capital expenditure, individual site capacities, or customer commitments for the wider programme.

Rather than building a small number of large campuses, the company intends to distribute computing and storage through locations already connected to Italy’s fixed-access network. The exchanges would combine fibre connectivity, data processing, and storage within the same infrastructure, allowing operators to offer services that do not need to send every workload to a distant cloud region.

The programme has been under technical development since December 2025, when FiberCop began testing national edge-cloud systems. Initial demonstrations followed at Mobile World Congress in March, and the company is now presenting the Rome facility as the first operating node in a national architecture.

Old network sites gain another purpose

Telephone exchanges were designed to aggregate local connections and house switching equipment, placing them in communities where reproducing the same network presence would be difficult and expensive. As copper equipment is removed and traffic moves to fibre, portions of those buildings can potentially be reused for computing infrastructure.

FiberCop operates nearly 28 million kilometres of fibre and says its ultra-broadband network reaches more than 96% of active fixed lines. That footprint gives it buildings, fibre routes, network access, and local presence across much of the country, although not every exchange will be technically or commercially suitable for conversion.

Edge facilities require reliable power, cooling, physical security, fire protection, equipment monitoring, and sufficient space for customers’ hardware. Older telecommunications buildings vary considerably, so creating a consistent service across more than 100 locations will involve considerably more than installing servers in rooms vacated by copper equipment.

“Our exchanges are moving from network access points to become true innovation hubs, in which connectivity, computing power and storage capacity are integrated to support the country’s digital development,” Massimo Sarmi, chairman and chief executive of FiberCop, said in the programme announcement.

The reuse model could reduce some construction and connectivity costs compared with developing entirely new facilities. It also gives FiberCop a potential source of revenue from buildings whose original function is declining, while making the fibre network more useful to operators selling cloud, AI, and managed services.

Latency alone will not create demand

Edge computing is most useful where data must be processed close to the place it is generated, whether because an application needs a rapid response, moving the data is expensive, or an organisation wants information to remain within a particular location. Potential uses include transport management, industrial monitoring, healthcare systems, public safety, environmental sensing, and local government services.

Many business applications nevertheless work adequately from central cloud regions and benefit from the scale, tooling, and service breadth available there. A distributed facility must offer more than a marginal reduction in network delay if it is to attract enough sustained workloads to cover its operating costs.

FiberCop’s immediate customers are telecommunications operators rather than the businesses and public bodies ultimately using the services. Those operators will need to combine computing, software, security, and support into products organisations are willing to buy, leaving utilisation dependent on an ecosystem beyond FiberCop’s direct control.

AI provides one possible source of demand, particularly for inference workloads analysing local video, sensor, production, or mobility data. The hardware required for such work can consume substantial electricity and produce considerable heat, even where the site is much smaller than a hyperscale facility.

The company has not disclosed how much accelerated-computing capacity will be available at each exchange or whether customers will install their own equipment. Nor has it explained whether every node will carry a common technical specification, or whether larger cities will receive more capable facilities than smaller regional sites.

Distribution changes resilience and security

A national network can reduce dependence on one central location and keep services operating closer to users, although a larger number of facilities also increases the number of sites that must be secured, maintained, patched, and monitored. Consistent operating standards will be essential if public bodies or regulated industries use the network for sensitive workloads.

FiberCop says proximity will support data sovereignty, but geography is only one part of that claim. The location of a server can help an organisation control where information is processed, while ownership of the software, remote administrative access, subcontractors, and applicable legal arrangements remain equally relevant.

The architecture will also need clear resilience arrangements. Smaller sites may have less redundancy than large data centres, while local power failures, connectivity incidents, or physical damage could interrupt a regional service. Operators will need to decide whether workloads fail over to another exchange, return to a central cloud, or continue locally in a reduced mode.

FiberCop reported revenue of €3.78 billion for 2025, alongside a net loss of €205.7 million. The edge programme is therefore part of a wider effort to extract more value from an extensive fixed-network estate while the company continues investing in fibre.

Its existing exchanges give the plan a physical foundation that many edge-computing proposals lack. The strongest evidence will come from the nodes following Rome: their technical capacity, the operators using them, the workloads running there, and the proportion of available infrastructure that customers are prepared to pay for.