NB-IoT and LTE-M: Cellular Connectivity for Low-Power Devices
A water meter buried in a concrete pit, three floors underground, sends one tiny reading a day. It has no power outlet, no Wi-Fi in range, and it needs to keep doing this for ten years on a battery you’ll never replace.
Wi-Fi can’t reach it. Bluetooth and Thread can’t travel that far. LoRaWAN could — if someone built and ran a gateway nearby. But there’s already a network that blankets the entire city, penetrates deep underground, is professionally maintained, and charges pennies per device per month. It’s the cellular network — the same one your phone uses — running two special low-power modes built for exactly this.
Those modes are NB-IoT and LTE-M. They’re how a battery-powered sensor talks to a cell tower without the power drain, cost, or complexity of a smartphone modem — and they’re quietly becoming the default way to connect the millions of devices that live too far from a wall socket to use anything else. 📶
The Gap These Two Fill
To see why these technologies exist, line up the options a device has for talking to the world, and watch where each one fails.
The short-range mesh protocols — Zigbee, Thread, Bluetooth mesh — are brilliant inside a home or building, but they need a nearby hub and their range is measured in rooms, not kilometres. Wi-Fi reaches further but is power-hungry and stops at the edge of your network. These are the technologies I’ve covered across this series, and none of them can help a sensor sitting alone in a field or a basement across town.
The unlicensed long-range option — LoRaWAN — solves the distance problem beautifully and costs almost nothing to run, but someone has to own and operate the gateways. If there’s no gateway near your device, there’s no network. It’s superb when you control the coverage and awkward when you don’t.
And then there’s full cellular — the LTE modem in your phone — which has coverage everywhere but was built to stream video and hold calls. Drop that into a battery sensor and the module is expensive, the power draw murders your battery, and you’re paying for bandwidth a water meter will never use.
NB-IoT and LTE-M sit precisely in the gap: the reach and reliability of the cellular network, stripped down to sip power and cost almost nothing. They belong to a category called LPWAN — Low-Power Wide-Area Network — and specifically the licensed side of it, running on the mobile operators’ own spectrum rather than the free-for-all unlicensed bands LoRaWAN uses. That licensed spectrum is the whole difference: it’s clean, protected, and already deployed nationwide.
Same Family, Different Jobs
NB-IoT and LTE-M are often mentioned in the same breath, and they do share DNA — both are 3GPP standards, both run on licensed cellular spectrum, both deliver years-long battery life and SIM-grade security. But they were tuned for genuinely different jobs, and the differences decide which one you should use.
NB-IoT (Narrowband IoT) is the extremist. It operates on an ultra-narrow 180–200 kHz slice of spectrum, engineered for one thing above all: tiny, infrequent payloads reaching devices in brutal locations. Its data rate is deliberately minuscule — tens to low hundreds of kilobits per second at best, often far less — because a sensor sending a few bytes a day doesn’t need more. In exchange it buys the deepest indoor penetration of any cellular technology, the lowest power draw, and the lowest module cost. The catch: it’s built for devices that stay still and don’t mind latency. Handing off between towers at speed is not its game.
LTE-M (LTE for Machines, or Cat-M1) is the all-rounder. It offers meaningfully higher data rates — up to around 1 Mbps — enough for firmware updates, occasional voice, and richer telemetry. Crucially, it supports mobility: an LTE-M device can move between cell towers cleanly, which NB-IoT can’t reliably do. It also has lower latency. The trade-off is modestly higher power consumption and cost than NB-IoT — still a fraction of full LTE, but not quite as frugal as its narrowband sibling.
The mental model that actually works: if your device is a stationary meter or sensor that wakes up occasionally to whisper a few bytes from somewhere hard to reach, that’s NB-IoT. If it moves, needs to be commanded and updated over the air, or sends more than trivial amounts of data, that’s LTE-M. One is a specialist; the other is a generalist.
The Tricks That Make Batteries Last a Decade
A ten-year battery life on a cellular connection sounds impossible — your phone barely lasts a day. The trick is that NB-IoT and LTE-M devices spend almost all of their life asleep, and 3GPP built two specific power-saving mechanisms into the standards to let them do it.
PSM (Power Saving Mode) lets a device tell the network, “I’m going to sleep, and I won’t be reachable for a while — don’t bother paging me.” The device powers down its radio almost completely while staying registered on the network, so it doesn’t have to do a costly re-attach when it wakes. For a sensor that reports once a day, the radio can be dark for 23 hours and 59 minutes.
eDRX (Extended Discontinuous Reception) is the lighter-touch cousin. Instead of sleeping completely, the device agrees with the network on a stretched-out schedule for when it will briefly wake to check for incoming messages — maybe every few seconds or minutes instead of constantly. It’s the right choice when a device needs to remain loosely reachable but still wants to save power between check-ins.
Together these are why the battery math works. The radio — by far the biggest power draw — is off the overwhelming majority of the time, waking only to transmit its tiny payload or to peek for a command. A coin cell or a modest battery pack genuinely can carry a device for a decade, which is the entire premise of deploying something you’ll never physically touch again.
This is the same design philosophy as Thread’s sleepy end devices or Bluetooth’s low-power nodes, scaled up to the cellular network: do the expensive thing (talk on the radio) as rarely as possible, and stay asleep the rest of the time.
Why 2026 Is the Migration Moment
There’s a specific reason cellular IoT is having a moment right now, and it’s not hype — it’s a deadline. The old networks these devices used to run on are being switched off.
For years, a huge amount of machine-to-machine connectivity — alarm panels, older trackers, vending telemetry — ran on 2G and 3G. Those networks are now being shut down worldwide to free up spectrum for 4G and 5G. Every one of those legacy devices needs a new home, and NB-IoT and LTE-M are the designated replacement: they offer the same extended battery life and deep building penetration that made 2G popular, but with better security, lower latency, and — crucially — long-term carrier commitment.
That commitment matters enormously when you’re deploying a device meant to live for a decade. Both NB-IoT and LTE-M now carry longevity commitments extending well into the 2030s, so the thing you install today won’t be orphaned by another sunset halfway through its life. And the economics have crossed a threshold: module costs have dropped to under $5 in volume, which makes cellular viable even for low-margin applications that could never have justified it before.
Get Uladzislau Bayouski’s stories in your inbox
Join Medium for free to get updates from this writer.
The result is a genuine inflection. By 2026, NB-IoT and LTE-M modules are expected to account for the majority of new cellular IoT module shipments for battery-powered, multi-year applications. This isn’t a pilot-project technology anymore — it’s the default answer for a very large class of devices.
Where RedCap Fits — and Where It Doesn’t
Any honest 2026 discussion of cellular IoT has to address the new arrival everyone’s talking about: 5G RedCap. It’s important, and it’s widely misunderstood as a replacement for NB-IoT and LTE-M. It isn’t.
RedCap (Reduced Capability), standardised in 3GPP Release 17, was created to fill a gap above the low-power technologies, not to absorb them. NB-IoT and LTE-M cover the low-rate floor — meters, trackers, simple telemetry. Full 5G covers the high end. In between sat a class of devices that needed more than about 1 Mbps but couldn’t justify a full 5G module: video surveillance, advanced wearables, richer industrial sensors. RedCap is that middle rung, offering up to roughly 150 Mbps while consuming far less power than a full 5G module.
There’s an even leaner variant, eRedCap (Release 18), simplified further down toward a maximum 10 Mbps for cost-sensitive mid-tier devices. But note the direction of travel: RedCap and eRedCap add rungs between LPWA and full 5G. They don’t remove the bottom rung.
The distinction that matters for builders: RedCap is lower-complexity 5G, not a low-power LPWA technology. It won’t match NB-IoT’s coverage or battery profile, its module cost in early 2026 sits around $50 versus under $5 for LPWA, and it requires 5G standalone network infrastructure that’s still rolling out. If your device sends a few bytes a day from a basement and must last a decade on a coin cell, RedCap is the wrong tool and NB-IoT is right. If your device streams moderate data and has a real power budget, RedCap earns its place. They’re neighbours, not rivals.
What They’re Actually Used For
The abstract strengths become obvious the moment you look at real deployments, because the technology and the use case fit together so tightly.
NB-IoT’s heartland is stationary, deep-coverage, tiny-payload metering and monitoring. Smart utility meters — water, gas, electricity — buried in pits and basements. Environmental sensors. Parking-space detectors embedded in the tarmac. Agricultural soil probes in remote fields. Anything that sits still, reports rarely, lives somewhere signals struggle to reach, and must run for years untouched.
LTE-M’s heartland is anything that moves or needs a conversation. Asset and vehicle trackers that hand off between towers as they travel. Wearables and personal-safety devices. Connected-health monitors that need firmware updates and occasionally more than a trickle of data. Fleet telematics. Alarm and security panels migrating off 2G that need reliable two-way command, not just one-way reporting.
A well-designed smart-city estate often uses both by policy: NB-IoT for the sleepy sensors that just report, LTE-M for the controllers and gateways that need to be commanded and updated. They’re complementary tools, and mature deployments treat them that way rather than picking a side.
The Honest Trade-offs
Cellular LPWAN is powerful, but it’s not free of caveats, and pretending otherwise helps no one.
You don’t own the network. Unlike LoRaWAN, where you can run your own gateways, NB-IoT and LTE-M mean depending on a mobile operator — their coverage, their pricing, their roadmap. That’s a feature (someone else maintains nationwide infrastructure) and a constraint (you’re a tenant, not a landlord).
Coverage is uneven and operator-specific. Not every carrier offers both technologies, and support varies by country and even by region within a country. Some operators run NB-IoT but not LTE-M, or vice versa. Roaming across borders — vital for asset trackers — is better on LTE-M than NB-IoT and still needs checking operator by operator.
There are recurring costs. Every device needs a SIM and a subscription. The per-device cost is tiny — often pennies a month — but it’s an ongoing operational expense that unlicensed options don’t impose, and at fleet scale it adds up and needs managing.
NB-IoT’s constraints are real. The low data rate and weak mobility aren’t bugs — they’re the deliberate cost of its efficiency — but they will bite you if you pick NB-IoT for something that later needs to move or send more data. Choosing the wrong sibling is the most common and most avoidable mistake.
For Builders and Buyers
Match the technology to the device’s life, not just its data. Ask whether it moves, whether it needs over-the-air commands and updates, how much data it really sends, and where it physically lives. Stationary + tiny payload + deep indoor → NB-IoT. Mobile or command-and-update + moderate data → LTE-M. Moderate streaming + real power budget → look at RedCap instead.
Verify coverage before you commit the fleet. Check that your target operators actually offer your chosen technology in every region you’ll deploy, and if devices cross borders, confirm the roaming story explicitly. This is the step that quietly sinks projects when skipped.
Design around PSM and eDRX from the start. The decade-long battery life only materialises if your firmware uses the sleep modes properly. Treat wake schedule, payload size, and reporting frequency as first-class design decisions, because they determine whether you hit your battery target or miss it by years.
Don’t over-buy connectivity. The pull of “future-proofing” with 5G/RedCap is strong, but a water meter does not need 150 Mbps, and paying for it in module cost and power is a real penalty. Buy the rung of the ladder your device actually stands on.
Plan for the long life you’re designing for. These devices are meant to run into the 2030s. Favour the technologies with stated longevity commitments, keep firmware updatable where the technology allows it (LTE-M makes this easier than NB-IoT), and don’t tie a ten-year device to any network anyone is already planning to sunset.
💡 Final Thought
There’s something quietly elegant about NB-IoT and LTE-M. They didn’t require building a new network from scratch or inventing exotic radio physics. They took the cellular infrastructure that already blankets the planet — the most expensive, most maintained communication system humanity has ever built — and carved out two lanes on it narrow enough for a device that sips power and speaks in whispers.
That’s the whole trick, and it’s why these two matter more than their low profile suggests. The dramatic connectivity stories go to 5G and satellite, but the unglamorous reality is that most of the tens of billions of devices we’re connecting don’t need drama. They need to send a few bytes, from somewhere awkward, for years, cheaply, without anyone ever visiting them again. NB-IoT and LTE-M are the answer to that unglamorous, enormous question.
As 2G and 3G go dark and module prices fall through the floor, these two are inheriting the vast installed base of the world’s simplest connected things — and quietly becoming the connective tissue for the meters, trackers, and sensors that will define IoT’s next decade. Not the loudest technology in the room. Just the one holding most of it together. 📶
If you found this article insightful, follow for more deep dives into IoT, connectivity, and the tech shaping our connected world! 🔬
