When the copper goes, the power cut becomes the operator's problem

For most of a century a copper telephone line in Italy drew its power from the exchange. When the electricity failed in a house, the phone on the hall table usually kept working, because the exchange had batteries and generators and the line carried enough current to ring a bell and carry a voice. Nobody marketed this, and few customers knew it. Yet a great deal of quiet dependence was built on it: the elderly customer whose telecare alarm dials out over the line, the farm that calls for help when a storm has cut the power, the small shop whose alarm panel reports a break-in at night.

My position is that as operators move customers off copper, that dependence does not disappear. It moves, and it moves onto the operator. Fibre to the home and fixed-wireless access both need power in the customer's premises, and a mobile site needs power at the mast. An operator migrating customers off copper should tell each customer plainly how long the new service lasts without mains power, should provide backup equipment to customers who depend on the line in an emergency, and should size the batteries at its own sites by what each site serves rather than by one standard.

What the copper line did quietly

At Corvalta Mobile we resold the incumbent's wholesale copper lines to about 205,000 fixed customers. Between 2024 and 2026, copper was switched off in exchange areas covering 168,000 of them, under the procedure supervised by AGCOM, the communications regulator. Our task was to move those customers to wholesale fibre to the home where it was available, or to our own 5G fixed-wireless service where the radio capacity allowed.

When we analysed the customer base before the migration, we found that about 9,400 customers had told us at some point of a telecare device, a medical alarm, or a lift or security line on their connection. That was almost certainly an undercount, because customers rarely tell a broadband provider such things and our records were patchy. None of those devices would work during a power cut on fibre or fixed wireless unless the router or terminal had a battery unit. On copper they would have kept working as long as the exchange held.

Tell customers the number of minutes

The first thing I would ask of any operator is honesty about time. We now state, in the migration letter, on the order page and on a label on the equipment, how long the service runs without mains power: zero minutes for a fibre router without a battery unit, about four hours with the unit we supply, and for fixed wireless the same figures for the terminal, with a note that the service also depends on our mast having power, which in most places it will for several hours.

The wording took weeks of argument. Our commercial colleagues worried that a figure of zero would frighten customers into leaving us. Some did leave; of the 168,000, about 17,000 did not migrate with us, and the exit surveys mention power among other reasons. I think the alternative, letting customers discover the figure during a storm, is worse for them and in the end worse for us.

The second thing is to equip the customers who depend on the line. We offered a battery unit free to every customer who had declared a care or alarm device. In the first quarter the offer was opt-in and only 41% took it. We then made it opt-out, and 88% kept the unit. I should have started with opt-out. The cost across the migration was about €1.3 million, a small part of the migration budget.

Batteries sized by what a site serves

The third point concerns our own network. A mobile site's battery autonomy is usually set by an engineering standard, often a few hours, applied to every site regardless of what it serves. When a site carries fixed-wireless homes that have just lost an exchange-powered line, the site has taken on part of a role the exchange used to play. We now classify sites by role: sites carrying fixed-wireless homes, sites that are the only coverage for a village, sites serving hospitals or care homes, and the rest. The first three classes receive between six and eight hours of autonomy, and the rest stay at the standard. Where we could, we combined this work with our radio replacement programme, which saved the cost of separate visits.

Our finance team objected that we were taking on a cost the incumbent had carried and that nobody had asked us to carry. That is a fair description. It is not, in my view, a reason to leave customers less safe than they were before a migration they did not request. Resilience is also becoming a competitive question. After the summer storms of 2025, customers in two southern provinces asked us directly how long our service would last in a blackout. An operator that can answer with a figure earns a kind of trust that advertising cannot buy.

I do not argue that operators should carry unlimited obligations, or that every site should run for days on batteries. Generators and longer autonomy cost money a challenger has to find elsewhere, and beyond a certain point a long regional blackout is a matter for civil protection rather than for a telecom company. My argument is narrower. The copper line did something for vulnerable customers that nobody wrote down, and as it is switched off somebody has to write it down and decide who does it now. Operators are best placed to answer first, because they know which customers are being moved and what the new service needs in order to stay on.

There is a reasonable objection that public rules on emergency access should settle these questions, and I would welcome clearer ones. Until they exist, an operator should not wait for them. The customers being migrated this year cannot wait for a consultation, and the records needed to protect them already sit in the operator's own systems.

I would ask any board overseeing a copper migration three questions. How many of our migrating customers depend on the line in an emergency, and how do we know? What does the migration letter say about power, in minutes? And which of our sites now carry customers who used to rely on an exchange battery, and have we changed those sites' batteries to match? If management cannot answer all three, the migration plan is not finished.