A deadline, not a software update
On 31 January 2027, Openreach switches off the old copper phone network for good. No extensions this time. Openreach has been signalling this move for the best part of a decade, and it has just written to every communications provider on its network making one point very plainly: there is no fallback date to slip to.
Most people reading that will file it under “my landline is going digital” and move on, and for most households that’s genuinely all it is. But for a meaningful slice of UK businesses, this isn’t a phone upgrade. It’s the day a dependency nobody wrote down stops working, all at once, on a date that’s been public for years.
What’s actually riding on that copper
Openreach’s own numbers give a sense of scale: around a million copper-based lines are still in use across the UK, roughly 300,000 of them business lines. Ofcom’s figures tell the same story from the consumer side — residential customers still on the old network fell from 5.2 million in July 2024 to 3.2 million a year later, so the move is genuinely happening. What’s left, by Openreach’s own account, is the hard tail: the complex connections. Multi-site businesses. Alarm and lift lines. Payment devices. The ones nobody wants to touch because nobody’s quite sure what’s plugged into them.
That matters because a landline isn’t just a landline anymore, in the sense most BC plans use the word. It’s whatever analogue kit has been quietly signalling over it for the last fifteen years: intruder and fire alarm panels, lift emergency phones, older card payment terminals, door entry systems, and — for around 1.8 million UK households — careline and telecare pendants. That last category is exactly why this deadline has already slipped once: government asked providers not to force vulnerable telecare users onto digital lines until their devices were confirmed to still work. The date moved. The underlying problem — nobody quite knowing which devices depend on which lines — didn’t go away. It just got three more years.
The line item that isn’t a dependency map
I’ve lost count of the BIAs I’ve reviewed where “telephony” or “landline” sits as a single row, ticked off as a dependency, with nothing underneath it. That’s fine as a heading. It isn’t a dependency map. A dependency map tells you what’s actually downstream of that line — which alarm panel, which lift, which card reader, who installed it, who holds the maintenance contract, and what happens to each one specifically if the signal stops.
What makes this a genuinely useful test, rather than just another compliance headache, is that it’s the rare resilience scenario with a fixed, published, no-surprises date. Most of the dependency failures I see — a substation outage, a third-party platform going down, a supplier’s own system failing — arrive with no warning at all. This one has been public for the best part of a decade, has a hard stop on 31 January 2027, and still, by Openreach’s own account, has roughly 300,000 business lines yet to move, plenty of which will be the complex, multi-site, multi-device kind.
If an organisation can’t get its arms around a dependency it’s had years of notice on, that tells you something uncomfortable about how it’ll cope with the one it gets no warning on at all.
What good practice looks like
For anyone advising on this, or buying BC and resilience services, the useful question isn’t “has the landline switch-off been handled?” It’s “can you show me, site by site, every piece of equipment still signalling over copper, who’s responsible for migrating each one, and what the fallback is in the gap?” That’s a different conversation, and most suppliers answering a tender won’t have it ready unless they’re asked directly.
A hundred days is enough time to do this properly, if it starts now rather than in January. It isn’t enough time if the first conversation about what’s actually plugged into those lines happens the week the network goes dark. Openreach isn’t moving the date this time. Whether an organisation finds that out through a planned site survey or a 2am call to the alarm company is entirely down to whether anyone went looking.
