LTE-M vs LoRaWAN: Where the Real Costs Hide

Most LTE-M vs LoRaWAN comparisons give you a table of specs and leave you to figure out the rest. Peak data rates your sensor will never hit. Range figures that assume flat terrain with no buildings. Battery claims based on a workload nobody actually runs in the field.

After 15+ years deploying IoT across airports, container yards, MRO hangars, and industrial fleets, I can tell you the protocol matters less than the operating model it locks you into. One choice means you rent connectivity from an operator. The other means you own infrastructure. Both have costs people don’t talk about until the invoice arrives.

This is the comparison I wish someone had handed me before my first large-scale tracking deployment. Not specs. Operating outcomes.

Two Protocols, Two Operating Models

Before anything else, understand the architecture. Everything that follows (cost, battery, risk) is a consequence of this single fork in the road.

LTE-M (also labeled Cat-M1 on module datasheets) runs on licensed cellular spectrum. Your device talks directly to an operator’s tower. The operator manages infrastructure, allocates spectrum, and handles cell-to-cell handovers when assets move. You pay a subscription per device. Think of it as renting connectivity: predictable access, broad coverage, but dependent on your landlord’s decisions about bands, roaming, and feature support.

LoRaWAN uses unlicensed sub-GHz bands (902-928 MHz in the U.S., for example). Devices transmit to gateways you or a third party install. Those gateways forward packets to a network server that deduplicates, authenticates, and routes the data. You can self-host that server, pay a managed provider, or mix both. Think of it as owning infrastructure: more control over coverage and data, more responsibility for maintenance and uptime.

This table is the one every comparison article publishes. And on its own, it’s misleading. Peak data rate is irrelevant when your device sends 50-byte sensor messages every 15 minutes. Connected-mode mobility adds nothing if your assets sit stationary in a warehouse. The row that actually drives your 10-year cost is infrastructure: who owns it, who fixes it when it breaks, and who decides to shut it down.

Battery Life Claims Deserve Skepticism on Both Sides

Every vendor on both sides will promise 10+ years of battery life. And under the right conditions, both can deliver it. The problem is that “right conditions” rarely matches your actual deployment.

LTE-M uses two power-saving mechanisms. Power Saving Mode (PSM) drops the device off the network entirely between transmissions, while Extended Discontinuous Reception (eDRX) lengthens listening intervals. Both can bring consumption into the microamp range during sleep. Here’s the catch: your device requests these modes, but the network decides whether to grant them. On your home operator, PSM might work perfectly. On a roaming network in another country, the visited operator may not support it at all. Your battery model just broke.

LoRaWAN handles it through device classes defined in the protocol specification. Class A (the default for battery-powered sensors) only opens receive windows right after transmitting. That’s inherently efficient. But move to Class B for scheduled downlinks or Class C for near-continuous listening, and the energy profile changes substantially.

The Sentinum Juno cellular tracker is a good example of honest specification. Nordic Semiconductor’s announcement qualifies its multi-year battery claim with a specific workload: one daily uplink and one daily GNSS fix. Double the reporting frequency, add geofencing logic, or enable more frequent location fixes, and “multi-year” compresses fast.

The honest answer: battery life is a system-level outcome. It depends on payload size, reporting schedule, retry behavior, positioning requirements, and (for LTE-M) what the operator actually grants. Compare identical workloads on the actual network. Marketing sheets are starting points, not commitments. See also the article: How Long Do Active RFID Tags Last? The Honest Math.

Total Cost of Ownership Reverses the Obvious Winner

Module price differences between LTE-M and LoRaWAN are single-digit dollars. Over a 10-year deployment of hundreds or thousands of devices, that difference is a rounding error. The real cost divergence lives in operations—which is why measuring asset tracking ROI requires looking beyond upfront hardware costs to total operational impact.

LTE-M operating costs are simpler to forecast but harder to control. Connectivity subscriptions can be surprisingly low: 1NCE lists $14 for 10 years and 500 MB per device, which works out to $1.40 per device-year before SIM-format charges. But add roaming for cross-border assets, SIM lifecycle management, and the occasional coverage gap requiring a site survey or operator negotiation, and the operational overhead grows.

LoRaWAN operating costs have a different shape. No per-device subscription if you run a private network. But gateways need purchasing, installation, power, and backhaul (often cellular, which means you’re paying a connectivity subscription anyway, just at the gateway level instead of the device level). The network server needs hosting: ChirpStack is open-source and eliminates licensing fees, but not the server hardware and ops labor. Commercial hosted LoRaWAN server plans typically start around EUR 170–190/month for up to 1,000 end devices. And someone needs to handle RF planning, interference troubleshooting, and gateway replacement when hardware fails.

The pattern I see repeatedly in the field: for a fleet of 500 mobile assets crossing borders, LTE-M’s operator model usually wins on operational simplicity. For 5,000 sensors concentrated in a single facility, LoRaWAN’s private-network model can cut the per-data-point cost dramatically. The trap is assuming one model fits everything.

Your Network Provider Can Exit. Plan for It.

Protocol viability and operator viability are different things. Most comparisons skip this distinction. It’s one of the most expensive oversights in IoT.

In late 2024, Objenious (a Bouygues Telecom subsidiary) shut down its LoRaWAN network and directed customers toward cellular IoT alternatives. Devices on that network needed migration, new connectivity agreements, or replacement.

Meanwhile, in the same country, Veolia’s Birdz division was scaling. Birdz deployed millions of smart water meters on Orange’s LoRaWAN network. By 2026, the LoRa Alliance reported over 4 million active Veolia smart meters in France. The system delivered over 60,000 leak notifications in a single year, helping save over 3 million cubic meters of water.

Same protocol. Same country. One operator exited. Another built one of the largest LPWAN deployments on the planet.

This risk isn’t unique to LoRaWAN. LTE-M users face it too. If your cellular provider deprecates a frequency band, restructures IoT pricing, or changes roaming agreements, your devices need a path forward. Dual-mode modules, eSIM capability, and multi-operator agreements all provide exits. But only if you negotiate them upfront.

Three questions worth asking before any connectivity contract:

1. If this provider exits, can my devices connect to an alternative network without hardware replacement?
2. Who owns the device credentials and provisioning data?
3. What’s the contractual migration path, and has anyone actually tested it?

When to Choose LTE-M, LoRaWAN, or Both

Start with LTE-M when:

• Your assets move (vehicles, containers, ULDs, ground support equipment, pallets in transit)
• You need over-the-air firmware updates to devices in the field
• Your assets cross borders and need roaming connectivity
• You need low-latency command/response (geofence alerts, remote configuration)
• You prefer not to own or maintain RF infrastructure

Start with LoRaWAN when:

• You have a dense concentration of sensors in a defined area (factory floor, campus, port yard)
• Data payloads are small and infrequent (temperature, humidity, fill level, presence)
• Network independence from telecom operators matters to your organization
• You need to optimize cost per data point at scale (thousands of endpoints)
• Data sovereignty or private-network control is a requirement

Consider both when:

• You have a mixed estate of fixed sensors and mobile assets
• LoRaWAN gateways already need cellular backhaul
• You want connectivity redundancy across different failure modes

Telia’s smart meter program in Sweden is a practical example. Ericsson describes Telia and partners converting more than 2 million Swedish electricity meters with support for both NB-IoT and Cat-M1. They purchased a managed connectivity capability and preserved radio-mode flexibility rather than betting on one mode for every location.

For asset tracking specifically (the world we live in at Datanet), LTE-M is usually the stronger starting point for anything that moves. Moving assets need reliable handovers, periodic GNSS fixes, and sometimes firmware updates. All of those favor cellular connectivity. Devices like the Oyster Edge or the ruggedized trackers in our asset tracking catalog are built around this reality. But if those same assets spend dwell time in fixed locations (a container yard, an MRO facility, a bonded warehouse), a LoRaWAN overlay for zone-level presence detection can complement cellular tracking cost-effectively. See also the article: Active RFID vs BLE: Pick the Wrong Radio, Pay for Years. See also the article: Do RFID Tags Work on Metal? Yes, If You Pick Right.

Both Protocols Are Moving Targets in 2026

If you’re reading an LTE-M vs LoRaWAN comparison written before 2025, most of its data points are stale. Both protocols have evolved enough to invalidate assumptions that were accurate two years ago.

On the LoRaWAN side:

• November 2025: RP2-1.0.5 added SF6 and SF5 data rates of 9.4 and 15.6 kbps. This doesn’t make LoRaWAN a broadband replacement, but it widens what’s practical for richer telemetry and denser networks.
• August 2026: new standards (TS014, TS018) enable automated device-profile retrieval and QR-code-based onboarding. This targets the installation labor bottleneck that quietly inflates LoRaWAN deployment costs.
• The Alliance’s June 2026 roadmap plans migration procedures, zero-touch onboarding, and a Standard Application Data Format through 2028.
• 150 million LoRaWAN devices deployed globally as of September 2026, growing at a reported 25% CAGR.

On the LTE-M side:

• GSMA explicitly places LTE-M within the 3GPP 5G evolution strategy. It’s not being replaced by 5G. It’s part of 5G.
• 129 LTE-M network deployments counted as of October 2025, with continued expansion.
• GSMA reported one billion combined NB-IoT and LTE-M active connections at end-2025.
• The broader cellular IoT market grew 13.3% in 2025 to 4.7 billion connections and $20.8 billion in connectivity revenue.

The trajectory on both sides is toward easier operations, not just faster radios. Faster onboarding, better fleet management, fewer manual configuration steps. That’s where the real competitive pressure is. Radio performance improvements matter, but labor cost and configuration errors dominate most deployment budgets. When evaluating either option, ask what the vendor has done about Day 2 operations, not just Day 1 installation.

Frequently Asked Questions

Can LoRaWAN track moving assets?

Yes. LoRaWAN doesn’t require gateway-to-gateway handover for moving devices. But tracking quality depends on gateway coverage along the route and the achievable reporting frequency within duty-cycle limits. For most mobile asset tracking use cases (cross-border logistics, fleet management, aviation GSE), LTE-M’s network-controlled mobility and higher data budget make it the more practical choice. LoRaWAN works well for zone-level presence detection within covered areas like yards and facilities. See also the article: How Far Can RFID Tags Be Read? Real Numbers & Range Guide.

Is LoRaWAN free to operate?

Unlicensed spectrum means no per-device radio license. It does not mean free infrastructure. Gateways need hardware, mounting, power, and backhaul. Network servers need hosting and maintenance. Open-source options like ChirpStack reduce software licensing, but not operational labor. Calculate total cost of ownership over your deployment’s expected life before comparing to LTE-M subscriptions.

Which protocol offers better security?

Neither wins by default. LoRaWAN provides AES-128 encryption with separate network and application session keys, but the LoRa Alliance itself warns that poor implementation (exposed keys, reused nonces) compromises security. LTE-M inherits cellular access security, but application-layer protection still requires deliberate design. Evaluate the full chain from credential provisioning through data delivery.

Will LTE-M become obsolete when 5G arrives?

No. GSMA positions LTE-M within the 3GPP 5G standards evolution. LTE-M coexists with and benefits from 5G network investments. That said, standards longevity does not guarantee that every operator will maintain LTE-M service indefinitely. Include migration clauses in your connectivity agreements, and favor hardware that supports multiple cellular modes where possible.

How do I choose between LTE-M and LoRaWAN for a mixed fleet?

Segment your assets by mobility, reporting frequency, and geography. Mobile assets crossing sites or borders point to LTE-M. Dense clusters of fixed sensors in a single facility point to LoRaWAN. Many real-world deployments benefit from both, with LoRaWAN gateways using cellular backhaul. Optimize each connectivity layer for its workload instead of forcing one protocol across everything.

What happens if my LoRaWAN (or LTE-M) network provider shuts down?

Your outcome depends on your contract and device architecture. Objenious shut down its French LoRaWAN network in December 2024, while Birdz scaled to over 4 million meters on a different provider in the same country. Before signing, verify that you own device credentials, have tested migration to an alternative network or self-hosted server, and have contractual exit rights that don’t require replacing hardware in the field.

If you’re weighing LTE-M vs LoRaWAN for an asset tracking program, the protocol choice is only the first layer. The hardware, the integration, and the operating model around it determine whether you get real visibility or just data points. We help teams work through that full stack. Reach out at info@datanetiot.com.