It's the connectivity layer that's been keeping smart meters reporting, trackers pinging, and vending machines talking to head office since the early 2000s. It's also, in a growing number of countries, on borrowed time.
This guide covers what GPRS actually is, why it became the default choice for low-data IoT, where it falls short, and, because this is the part that actually matters for your roadmap, what to do as operators switch it off.
What Is GPRS?
GPRS stands for General Packet Radio Service. It's an enhancement to GSM (Global System for Mobile Communications), the core standard behind second-generation (2G) mobile networks, and it's often referred to informally as 2.5G: a step up from 2G, though not quite a generation of its own.
Since it first rolled out in the early 2000s, GPRS has become the most widely adopted cellular technology for connecting low-data-volume devices. It was purpose-built for exactly the kind of intermittent, small-payload communication that defines most IoT use cases: a sensor that wakes up, sends a few bytes, and goes back to sleep.
How GPRS Works: From GSM to Packet Switching
The real difference between GSM and GPRS isn't a marketing distinction; it's a fundamentally different approach to moving data across the network.
GSM uses circuit switching (CS). A dedicated circuit is reserved between two endpoints for the full duration of a connection, regardless of whether data is actively flowing. That's fine for a phone call, where the line needs to stay open continuously. It's wasteful for a device that only has something to say once an hour.
GPRS uses packet switching. Data is broken into packets and sent in short bursts, sharing network capacity with other devices and users rather than monopolising a channel. The network only allocates resources when there's actually something to transmit.
On 2G and 3G networks, these two technologies run side by side, each doing a different job: circuit switching carries voice calls and SMS, while GPRS carries data. For IoT, GPRS is the piece that matters. It's the technology actually moving your telemetry, your status updates, and your commands.
GPRS transmissions are protected by cyclic and convolutional error correcting codes, and devices fall into one of three classes depending on how they handle voice and data simultaneously, from Class A devices that can run both at once, to Class C devices that switch manually between the two.
The Benefits of GPRS for IoT
Running Costs Based on Data, Not Time
This is the big one for IoT economics. With circuit-switched connections, you pay for how long the line is open. That's a poor fit for a device you want to leave always on but which only needs to transmit occasionally, such as an asset tracker, a remote sensor, or a smart meter. GPRS flips the billing model: you pay for the volume of data sent, not the duration of the connection. A device can sit connected around the clock and only cost you money when it actually has something to say.
A Meaningful Speed Jump Over GSM
GPRS is a considerable step up from basic GSM data transfer, which topped out around 9.6 to 14.4 kbit/s. GPRS pushed theoretical rates up to roughly 40 to 115 kbit/s, depending on how many of the eight available timeslots a device could use. That ceiling is rarely hit in real-world conditions, but it's still a meaningful jump. On the consumer side, this is what first made streaming, gaming, and richer mobile browsing possible. For business and industrial use, it opened the door to applications like remote surveillance and basic videoconferencing.
Mature, Wide-Reaching Coverage
Two decades in the field means GPRS coverage is about as mature and geographically broad as cellular IoT connectivity gets. For low-power, low-data devices deployed across large or remote areas, that maturity has been hard to beat, at least until recently.
The Limitations of GPRS
GPRS is a strong technical match for slow, infrequent data connections. If your device wakes up occasionally to send a small payload, 2G or 3G with GPRS has generally done the job well.
Where it falls short is anything data-heavy. Real-time video, high-frequency telemetry, large firmware-over-the-air updates, or applications with strict latency requirements need later-generation technologies. GPRS simply wasn't designed for that kind of throughput, and pushing it there tends to mean dropped packets, retransmissions, and unpredictable performance.
The bigger limitation isn't technical, though. It's a deadline.
2G and 3G Sunsetting: Why GPRS Has an Expiry Date
For two decades, 2G and 3G, and GPRS specifically, have quietly supported an enormous share of the world's IoT deployments. But mobile operators are under real pressure to reclaim that spectrum. Radio frequency is finite, and every megahertz still allocated to legacy 2G/3G infrastructure is a megahertz not available for 4G and 5G capacity that operators can monetise far more effectively. That trade-off is why sunsetting has picked up pace, not slowed down.
The timeline varies a lot by country
There's no single cut-off date. Shutdown timing is being decided operator by operator, market by market, and the spread is wide:
Region |
Status (as of 2026) |
|---|---|
| Sweden | Tele2, Telenor, and Tre switched 2G off from December 2025; Telia Sverige has set December 2027 |
| Norway | Telia Norge is already offline; Telenor Norge has set December 2027 |
| Denmark | No shutdown dates set yet, citing ongoing M2M/IoT dependency |
| Finland | Telia Finland has issued no shutdown date |
| France | Orange and Free Mobile targeting October 2026; SFR targeting November 2026 |
| Germany | Deutsche Telekom targeting mid-2028; Vodafone Deutschland targeting September 2028, with critical IoT exceptions running to 2030 |
| Spain | Movistar targeting 2027; MasOrange targeting December 2030 |
Even within the Nordics, Com4's home turf, you can see how fragmented this is: two operators in the same country, months apart on 2G, and years apart in some cases. If your fleet is deployed across multiple markets, or roams between them, there's no single date you can plan around. You need visibility into every operator you rely on, not just the market leader.
The risk shows up before the official shutdown date
This is the part that catches teams out. Devices don't necessarily fail cleanly on the announced shutdown date. They can start failing well before it, as operators reduce 2G/3G capacity, reallocate spectrum in stages, and thin out roaming agreements ahead of the final switch-off. The risks aren't limited to "no signal" either. Firmware fallback behaviour, SMS dependency for provisioning or alerts, roaming maturity on newer networks, and the practical lifespan of installed hardware all affect whether a device quietly keeps working or quietly stops.
In short: waiting for the announced date to plan your migration is waiting too long.
What Should Replace GPRS in IoT Projects?
The right replacement depends on what your device actually needs. Data volume, mobility, power budget, and expected device lifetime all point in different directions.
LTE-M is a strong fit for battery-powered or mobile devices that need lower power consumption than a full 4G modem, while still supporting voice fallback and reasonable throughput. A natural step for trackers and mobile assets.
NB-IoT is built for static, low-data deployments that need deep indoor or underground penetration, such as utility meters. It offers excellent power efficiency, though roaming support still needs checking market by market.
LTE Cat 1 / Cat 1bis is currently the most dependable baseline for fleets that need broad global coverage and mature roaming, without over-engineering for applications that don't need 5G-level throughput.
5G RedCap is designed for IoT rather than smartphones, and is a genuinely promising long-term option, but is still maturing in terms of availability and roaming agreements. Worth watching, not yet a default choice for most deployments.
Satellite/NTN is a complement rather than a replacement, useful for remote assets outside terrestrial coverage rather than as a primary connectivity layer.
There's no universal right answer here. The honest answer is that it depends on the device, and that's a conversation worth having before you commit to a chipset.
How to Prepare Your IoT Connectivity for the Future
A few practical steps, in the order we'd actually recommend tackling them:
- Audit what's actually deployed. Know which of your devices depend on 2G/3G/GPRS today, in which countries, and on which operators, not just which technology your SIMs are provisioned for.
- Check firmware behaviour, not just network specs. Understand how your devices fall back (or fail) when their primary network degrades, and whether that fallback logic still makes sense on newer networks.
- Match the replacement technology to the device, not the trend. LTE-M, NB-IoT, LTE Cat 1/1bis, and 5G RedCap each solve a different problem. The fastest or newest option isn't automatically the right one.
- Plan around the earliest realistic shutdown, not the announced one. Given how coverage degrades ahead of a formal sunset, build in margin rather than cutting it close.
- Work with a connectivity partner who can see the whole map. Multi-network, multi-country deployments need someone tracking timelines operator by operator on your behalf. That's a lot to keep on top of internally, especially across markets.
Where Com4 Fits In
Migrating a live IoT fleet off GPRS isn't a single switch to flip. It's a question of matching the right technology to each device, tracking sunset timelines across every market and operator you touch, and doing it without disrupting a service that's already running. That's the kind of groundwork we help customers work through every day, market by market, device by device.
If GPRS is quietly running somewhere in your fleet, it's worth finding out exactly where, before an operator decides the date for you.
