For decades, a surprising amount of Britain’s infrastructure has quietly relied on old connections. Burglar alarms dial out over copper phone lines. Lift emergency phones sit on analogue circuits. Telecare pendants worn by older people connect through landlines. Smart meters, payment terminals and asset trackers lean on 2G mobile networks that were built for phone calls in the 1990s.
Those networks are being switched off. The traditional copper telephone network is scheduled to go digital by 31 January 2027. The UK’s 3G mobile networks are already gone. The remaining 2G networks are due to close between 2029 and 2033.
Each change has been announced for years. What makes the current moment different is that the deadlines are now close enough to be concrete, and that several retirements overlap. For anyone responsible for connected devices in the UK, the question is no longer whether these networks will close, but which of their devices still depend on them.
Three Networks, One Deadline Crunch
A recent piece from IoT connectivity provider Eseye, published on IoT For All, set out why UK networks are being switched off and what the timelines look like. Three separate retirements are under way.
The Copper Phone Network
The first is the public switched telephone network, or PSTN, along with ISDN business lines and copper-based broadband. These are the traditional fixed lines that have carried voice calls and many machine connections for generations. They are expensive to maintain, use more energy than modern alternatives and cannot support today’s connectivity demands.
Openreach, which runs much of the UK’s fixed network, states on its Big Switch Off page that “by 31 January 2027, all traditional phone lines will be going digital.” The deadline was originally set for December 2025 and was pushed back. Stop-sell policies, which prevent new orders of the old services, already apply in most areas.
Openreach’s guidance is blunt: “Don’t wait until the deadline.” Customers who have not moved by then may be placed on what it calls Emergency Voice and Access Continuity, a temporary emergency voice-only service that may not support existing features.
3G: Already Gone
The second retirement is effectively complete. According to Eseye’s summary, EE and Vodafone finished switching off 3G in early 2024, Three followed shortly after, and VMO2 completed a staged regional withdrawal into early 2026. Any device that can only connect over 3G will no longer find a network.
2G: The Next To Go
The third is 2G, the oldest mobile technology still in service and the one many simple IoT devices were built on. The industry has committed to closing 2G by 2033 at the latest. Individual operators have set earlier dates: EE plans to begin its closure from May 2029, VMO2 from summer 2029, and Vodafone Three expects to switch off during 2030.
That leaves a window of roughly three to seven years for 2G devices, depending on which network they use. For equipment with a ten-year design life, installed today on 2G, that window is already too short.
Which Devices Are At Risk
Openreach and Eseye both list the kinds of equipment most likely to be affected. The overlap between their lists is telling: many of these devices perform safety-critical or revenue-critical jobs.
Alarms And Telecare
Security and burglar alarms are among the most common devices still using copper lines or 2G to call monitoring centres. Openreach specifically names “house alarms, CCTV” as equipment to check.
Telecare and health monitoring devices are the most sensitive category. Pendant alarms and health monitors worn or used by older and vulnerable people often connect through a home phone line. Openreach urges users of “telecare or health devices” to check compatibility with digital lines before switching. A device that silently fails after a network change could leave someone without help when they need it.
Lifts, Door Entry And Emergency Buttons
Lift emergency phones, door entry systems, intercoms, emergency call buttons and access control systems appear on both lists. These are often installed in buildings by one contractor and maintained by another, which makes it easy for the underlying connection to be overlooked. Many sit on analogue lines that have worked unnoticed for years.
Payment Terminals, Meters And Trackers
Eseye’s list extends to EPOS and payment terminals, smart utility meters, asset trackers and remote monitoring kiosks. Fax machines and ISDN business circuits are affected too. For retailers, a payment terminal that cannot connect is lost revenue. For utilities, meters that stop reporting mean estimated bills, site visits and regulatory headaches.
Why This Is Harder Than It Looks
On paper, the fix is simple: move each device to a newer connection. In practice, several factors make the migration difficult.
Devices Built To Outlast Networks
Many IoT devices are designed to run for ten years or more. Eseye points out that devices built for such long lifespans are now outlasting the networks they were designed for. A meter or alarm installed in 2018 on 2G may still be perfectly functional in 2029, apart from the fact that its network is closing.
Replacing or upgrading these devices usually means someone visiting each site. Across thousands of homes, shops or buildings, the logistics and labour can cost far more than the devices themselves.
Hidden Dependencies
A second problem is visibility. Some devices use an old connection only as a backup, such as an alarm that normally reports over broadband but falls back to a phone line or 2G if the main connection fails. Those fallback paths are easy to miss. Everything appears to work until the day the primary connection drops and the backup is no longer there.
Eseye highlights these silent dependencies as a particular risk. Peripheral systems, such as a lift phone connected through a building’s old line, may not appear in any central inventory.
Service Can Degrade Before The Official Date
The third issue is timing. Official switch-off dates mark the end point, not the start of the risk. Eseye warns that operators may restrict access before the formal closure. The fact that a device connects today is not a guarantee that it will connect next year.
What Replaces The Old Connections
The right replacement depends on the device, where it is and what it needs to do.
Digital Voice And IP Lines
For fixed-line equipment, the main path is a digital line: voice over IP delivered over broadband or fibre. Many alarms, lift phones and telecare devices can work over digital lines, but some older models cannot, or need adapters. Openreach’s advice is to check compatibility with the service provider and device supplier before switching.
A digital line also changes resilience. Unlike copper lines, which carried their own power, digital services depend on power at the premises. During a power cut, equipment may need battery backup to keep working, which matters for safety devices in particular.
Newer Cellular Options
For devices that use mobile networks, the replacement is usually 4G or 5G. Several flavours of 4G are designed for IoT. LTE-M and NB-IoT are low-power options for battery devices that send small amounts of data, which we covered in our explainer on the networks built to whisper. LTE Cat 1 bis, a single-antenna version of standard 4G, has become a common replacement for 2G in powered devices such as payment terminals and trackers.
New shipment data from IoT Analytics shows how quickly that shift is happening worldwide: Cat 1 bis accounted for more than half of the growth in cellular IoT module shipments in the first half of 2026, IoT Business News reported. For UK device owners replacing 2G equipment, it is one of the most widely available options.
Multi-Network And eSIM
The third part of the answer is flexibility. Eseye recommends multi-network resilience, with devices able to fall back to more than one technology, and SGP.32-compliant eSIM platforms that can switch operator profiles over the air. That allows devices to move to another network, or away from a retiring one, without a physical SIM swap. It will not rescue a 2G-only device, but it can make the next generation of equipment far less vulnerable to future network changes.
Who Is Responsible
One reason these migrations stall is that responsibility is spread thin. The organisation that owns a building may not own the lift phone. The company that installed an alarm may no longer maintain it. The provider of a telecare service may depend on a phone line paid for by the resident. A payment terminal may be leased, with connectivity bundled by the terminal provider.
Openreach addresses its guidance separately to homeowners, business owners, developers and landlords, which reflects how many parties are involved. In practice, each device needs a named owner who knows how it connects, who supplies that connection and who will pay for any replacement.
For larger organisations, that usually means bringing together facilities, IT, procurement and the suppliers of each system. For housing providers and care services, it means checking every property and every resident’s equipment, not only the devices listed in a central system. For device makers and service providers, it means contacting customers early with clear information about which products are affected and what the upgrade path is.
Without that clarity, the most likely outcome is that devices keep working until they suddenly don’t, and the search for who should have acted begins only after something fails.
A Practical Plan For The Next Three Years
For organisations with connected devices in the UK, a few steps can reduce the risk.
Start with an inventory. List every device that connects to a network, how it connects, and whether it has a backup path. Include peripheral systems such as lift phones, door entry, fire panels and alarms, which are often managed outside the IT department.
Prioritise by consequence. Telecare, alarms, lift phones and emergency buttons deserve attention first, because a failure can put people at risk. Payment terminals and revenue-generating devices come next.
Check dates for each network. A device on EE’s 2G network faces a different timeline from one on Vodafone Three’s. Devices on copper lines face the January 2027 deadline regardless of mobile plans.
Test replacements in real locations. Coverage maps give a general picture, but basements, lift shafts and rural sites can behave differently. A replacement that works in the office may not work in the plant room.
Plan for the future, not only for the next deadline. Choosing devices that support multiple networks and remote SIM management reduces the chance of repeating the exercise when today’s networks eventually change.
Begin early. Openreach’s message applies to mobile networks too. Migrations take longer than expected, installers are in demand near deadlines, and some devices will need redesign rather than a simple swap.
Lessons Beyond The UK
The UK is not alone. Operators in many countries are retiring older networks to reuse spectrum and cut costs, and fixed-line operators are moving away from copper. The details differ by market, but the pattern is the same: long-lived devices meet shorter-lived networks.
For device makers selling internationally, that makes connectivity planning part of product design rather than an afterthought. A product built today may still be in service in the late 2030s. The networks it depends on need to be around for that long, or the device needs to be able to move to something else. We looked at how industrial sites weigh these choices in our guide to industrial IoT and industrial connectivity.
Before The Lines Go Quiet
Network retirements rarely make headlines. They happen in stages, through stop-sell notices and operator announcements, and the devices that depend on old networks tend to be the ones nobody thinks about until they stop working.
The UK’s timeline is now clear. Copper phone lines go digital by the end of January 2027. 3G has already gone. 2G follows between 2029 and 2033. For alarms, telecare devices, lift phones, meters and terminals still relying on those networks, the time to act is before the deadline, not on the day the connection disappears.
