The core difference between LoRa and NB-IoT is who runs the network. LoRaWAN is a low-power wide-area network (LPWAN) you build with your own gateways on unlicensed spectrum; NB-IoT is cellular IoT on a mobile operator's licensed network, paid per SIM. Both carry small, infrequent data on battery for years; 2G/4G suits larger data volumes and moving devices.
Key takeaways
- LoRaWAN runs on the unlicensed 868 MHz band; you install the gateways and pay no SIM fees, but maintenance and security are yours.
- NB-IoT runs on the operator's licensed network; it reaches indoor and underground spots well and needs a per-SIM subscription.
- LTE-M suits moving devices and more frequent data; 2G/4G suits mains-powered devices sending larger volumes of data.
- The choice rests on data volume, installation point, power source, movement, who operates the network and the need for remote updates.
- Coverage maps are not enough: log packet loss for several days at the five hardest installation points with the real antenna and enclosure.
When does the connectivity question come up in an IoT project?
Most IoT projects start with a sensor: a temperature, flow rate, level or meter reading that someone wants to monitor remotely. The sensor is chosen, the prototype works on the bench, and then the question arrives: "How does this data get back to us?" One person suggests LoRa, another says "just put an operator SIM in it", and the supplier argues for whatever module is in their catalogue. We walk through a typical water network example in remote pump station monitoring.
The connectivity decision shapes the rest of the project. Connectivity determines battery size, enclosure, antenna placement, monthly running costs and even what the installation crew does on site. A prototype that works on a desk can go silent inside a basement meter cabinet or on a concrete pole in a field. If the wrong technology is discovered after hundreds of devices are deployed, the fix usually means taking them all down again.
Why does a wrong connectivity choice get more expensive?
The Internet of Things (IoT) has long since outgrown the pilot stage. Device numbers are rising fast, and every poor connectivity decision is multiplied by that scale.
IoT Analytics expects the number of connected IoT devices to grow 14% to 21.1 billion in 2025 and reach 39 billion by 2030. (IoT Analytics — State of IoT 2025)
The cellular side is just as large. The Ericsson Mobility Report (June 2026) says cellular IoT connections reached around 4.5 billion at the end of 2025, with NB-IoT and Cat-M (LTE-M) continuing to roll out for low-cost devices with long battery lives. Businesses are using these devices more too: Eurostat reports that in 2021, 29% of EU enterprises with 10 or more employees used IoT devices or systems; of these, 30% used them to manage energy consumption and 24% for condition-based maintenance.
As IoT fleets grow, three costs grow with them: return visits to site (battery, antenna or device replacement), network or subscription fees, and decisions that cannot be made because data is missing. Connectivity is the single choice with the biggest influence on all three. The choice of radio module also affects the device's CE conformity work; see our guide to CE marking for electronic devices.
What are LoRaWAN, NB-IoT, LTE-M and cellular IoT?
LoRa / LoRaWAN. LoRa is a long-range, low-power radio modulation; LoRaWAN is the network protocol built on top of it. It uses unlicensed ISM spectrum (the 868 MHz band in Europe and Türkiye). An organisation either installs its own LoRaWAN gateways or uses a LoRaWAN network operator. It is ideal for very small packets sent infrequently. As of 2026, the LoRa Alliance describes LoRaWAN as an ITU-approved international standard (ITU-T Y.4480) offering range of up to 15 km in rural areas and battery life of more than 10 years.
Unlicensed bands carry duty-cycle limits, so LoRaWAN is not suited to frequent or large transmissions.
NB-IoT (Narrowband IoT). A 3GPP (3rd Generation Partnership Project) standard narrowband cellular IoT technology running on an operator's licensed network. Its ability to reach indoor and underground locations, together with power-saving modes (PSM and eDRX), makes it well suited to fixed, battery-powered devices. The operator runs the network and you pay per SIM. It is not designed for moving devices or high data rates.
LTE-M (LTE for Machines, Cat-M). Another 3GPP cellular IoT standard. It offers higher data rates than NB-IoT and supports handover between cells for devices on the move, which makes it a good fit for trackers and applications that send data more often. Check that the operator offers it in your area.
Standard cellular (GPRS, 4G, LTE Cat-1). Standard cellular is the familiar mobile network: broad coverage, plenty of capacity, and large payloads such as firmware updates are easy to deliver. The trade-off is power consumption, so it is best for mains-powered devices or ones that can be recharged regularly.
LoRa vs NB-IoT vs LTE-M vs 4G: comparison table
| Criterion | LoRaWAN | NB-IoT | LTE-M | 2G / 4G |
|---|---|---|---|---|
| Who runs the network? | You, or a LoRaWAN operator | Mobile operator | Mobile operator | Mobile operator |
| Spectrum | Unlicensed ISM | Licensed | Licensed | Licensed |
| Data volume | Very low, infrequent | Low | Medium | High |
| Long battery life | Excellent | Excellent | Good | Difficult |
| Indoor / underground | Depends on gateway placement | Strong | Good | Limited |
| Moving devices | Limited | Not suitable | Suitable | Suitable |
| Running costs | Gateways and maintenance (no SIMs) | Per-SIM subscription | Per-SIM subscription | Per-SIM subscription |
| Remote firmware updates | Hard, slow | Possible, limited | Suitable | Easy |
Treat the comparison table as a starting point rather than an answer. NB-IoT coverage can be strong in one district of a city and absent in the next; a LoRaWAN gateway that covers a wide area on flat, open ground may only reach a few streets in a dense urban area. On farmland, gateway positions can be planned in advance on the parcel map produced by a drone NDVI analysis.
How do you choose IoT connectivity? A six-question framework
- How much data, and how often? A few meter readings a day and a vibration spectrum every minute are very different workloads. Small, infrequent data points to LoRaWAN or NB-IoT; larger or more frequent data points to LTE-M or 4G. For industrial electricity meters in Turkey, the distribution company usually makes this call; see the automatic meter reading obligation for details.
- Where will the device sit? Basement, metal cabinet, underground chamber (such as the district flow meters used in water loss analysis), inside a concrete structure, or open ground? Hard-to-reach indoor spots are where NB-IoT is strongest; wide open areas, such as fields fitted with soil moisture sensors, favour LoRaWAN.
- Where does power come from, and for how long must the battery last? With mains power, energy use is a secondary concern. For years of operation without a battery change, plan LoRaWAN or NB-IoT together with the transmission interval. Battery loggers spread across a store, as in pharmaceutical warehouse temperature mapping, are a typical case for this calculation.
- Will the device move? Vehicle, container or asset tracking (for example cold chain monitoring in a refrigerated truck) calls for LTE-M or 4G. We trace the path from the sensor in the trailer to the control room in our guide to reefer truck temperature monitoring.
- Who will operate the infrastructure? Running your own network avoids SIM subscriptions but makes gateway installation, maintenance and security your responsibility. An operator network removes that burden in exchange for an ongoing per-device cost.
- How important are remote updates? If firmware must be updated in the field, very low-bandwidth links make that slow and awkward. With security patches in mind, this question deserves serious weight.
In most projects, the answers to these six questions do not point to a single technology. A common and sensible answer is a hybrid architecture: LoRaWAN in dense sensor areas, NB-IoT at scattered or hard-to-reach points, and 4G for the gateways' own backhaul.
How do you run a field test? Decide on site, not at the desk
The most common mistake in choosing IoT connectivity is deciding from a coverage map. Operator maps mostly show outdoor coverage; they cannot show the inside of a meter cabinet. The better method is to pick the five hardest installation points in the pilot area, measure signal with the real antenna and enclosure for each candidate technology, and log packet loss over at least several days. That measurement costs very little compared with deploying hundreds of devices on the wrong network.
However good the measurements, links will drop from time to time; making the device store data and keep working meanwhile is a separate design question. Our article on access control offline mode walks through one example. All our connectivity, protection-rating and field-device guides are gathered in the IoT & Hardware hub.
We show how this choice plays out in utility metering in our automatic meter reading guide, and how a device moves from design to series production in our article on the IoT product development process.
Monitoring herds on pasture is a good example of LoRaWAN in use where mobile coverage is weak across large areas; we describe it in our article on the livestock monitoring system.
To see how the choice plays out in the field, take agriculture: in a greenhouse with several blocks and a distant borehole, sensors can use a short-range network while the controller reaches the cloud over cellular. We describe that set-up in our guide to smart greenhouse automation.
How does Digital Bridge choose IoT connectivity?
At Digital Bridge, we do not treat connectivity as a decision separate from the device. Our electronics, antenna layout, enclosure, embedded software and management panel are designed by the same team, so the choice is made for the device as a whole.
- We start with a free technical feasibility study. We establish data volumes, installation points, power source and the software the data must reach, and set out in writing which technology, or which hybrid architecture, makes sense.
- We measure on site. We test prototypes at real installation points and log signal strength and packet loss; problems found in testing go back into the design before series production. This is the foundation of our IoT device design and manufacturing service.
- We design for the application. For metering, our AMR meter reading system runs over GPRS/4G, LoRaWAN, NB-IoT or M-Bus; we apply the same selection process to precision agriculture IoT and to cold chain monitoring in food and logistics.
- We carry the data into your systems. Device data is collected in a remote monitoring panel and passed to your ERP, SCADA or in-house software via API.
We show how these connectivity options come together at city scale, with water, lighting and waste examples, in our article on smart city applications.
Next step
If you are planning an IoT project, prepare four things: what data the device will send and how often, a description of the installation points (photos help a great deal), the power source, and the software the data needs to reach. Then get in touch through our contact page, and we will make the technology choice with you, backed by measurements on site.