LoRaWAN in 2026: When It's Still the Right Call

LoRaWAN has been pronounced dead annually for the past five years. It is also still shipping in volume — and in 2026, for the right deployments, it remains the most pragmatic choice. The category just got narrower as alternatives matured.

Where LoRaWAN still wins

Three deployment patterns where LoRa remains the default:

1. Customer-premises private networks.

A customer who owns the gateway, controls coverage in their facility, and does not want recurring per-device data fees is exactly the LoRa target. Industrial sites, agricultural operations, large commercial properties. The capex of a gateway pays back quickly when you avoid carrier fees on hundreds or thousands of devices.

2. Very low-data, multi-year battery.

LoRa’s air-time efficiency at low payloads beats cellular alternatives when the device sends a few bytes a few times a day and needs to run on a coin cell or small battery for years. NB-IoT closes the gap, but LoRa still leads on the absolute lowest power profile.

3. Outdoor coverage where cellular is patchy.

Remote infrastructure — rural pumps, distributed solar, environmental sensors — often sits in cellular dead zones. A LoRa gateway placed strategically reaches kilometers, including indoors and through obstructions, where cellular drops.

Where newer alternatives win

Three areas where the case for LoRa has weakened:

Mobile assets across regions. Cellular IoT (LTE-M, NB-IoT) handles roaming and mobility better than LoRa. Asset trackers, logistics, vehicle telemetry — these have mostly moved off LoRa in the last few years.

Higher-data products. Anything beyond a few hundred bytes per message, or anything more frequent than once an hour, runs into LoRa’s air-time limits. The duty-cycle restrictions on European bands in particular are a hard ceiling.

Public-network reach in dense urban environments. Public LoRa networks (TTN, Helium, regional carriers) have uneven coverage. Cellular IoT — even NB-IoT — wins in cities almost everywhere. LoRa shines on private networks; public LoRa is a coverage gamble.

What changed in the last few years

The protocol itself has not. The ecosystem has:

• NB-IoT and LTE-M coverage has filled in much of the geography LoRa used to dominate. Carriers price NB-IoT in cents per device per month, removing one of LoRa’s classic advantages.
• Satellite IoT (Swarm, Kineis, others) is now economically viable for remote-only deployments. A device that uplinks to satellite once a day at $0.08-0.20 per message changes the math for some agricultural and infrastructure use cases.
• Helium’s pivot away from a pure LoRa play has reshaped the public-network landscape. Coverage in many regions is patchier than it was in 2022.
• The Things Network and Stack remain strong for private-network deployments, with mature device profile management, ADR, and network-server tooling.
• Class B and Class C use cases remain niche. Most production deployments are Class A — uplink-driven, infrequent, battery-budget-led.

The decision framework in 2026

A short flow to decide LoRa vs alternatives:

1. Is the device static, on customer premises, with reliable line-of-sight to a controllable gateway location? → LoRa with private gateway, almost always.
2. Is the device mobile, possibly crossing regions, needing reliable connectivity? → NB-IoT or LTE-M, with cellular roaming agreements.
3. Is the device truly remote, where neither cellular nor LoRa gateway placement is realistic? → Satellite IoT.
4. Is the device in a building, with reliable Wi-Fi or Ethernet? → Wi-Fi or wired. LoRa on a building network is rarely the right answer.

The intermediate cases — semi-mobile, semi-rural, mid-data — are where the choice is genuinely close and depends on operational details.

What good LoRa deployments look like

The deployments that hold up in 2026:

• Gateway placement is surveyed, not assumed. Range is verified with the actual device, not predicted with a coverage tool.
• ADR (adaptive data rate) is enabled and trusted. Devices climb to higher data rates when conditions allow, conserving air-time.
• Class A only unless there is a specific reason. Class B and C burn battery and complicate the architecture.
• Network server is decoupled from application server. Clean separation lets you swap one without touching the other.
• Devices have a join lifecycle that handles re-joins gracefully. Devices that lose their session keys and cannot recover are a support burden.
• Payload codecs are versioned. When the device firmware updates, the network server must understand the new codec without breaking historical data.

What we typically do

For 2026 IoT projects, our default for static, low-data, customer-premises deployments is LoRaWAN with a controllable gateway and The Things Stack as the network server. For mobile or coverage-uncertain deployments, NB-IoT or LTE-M. The choice is deliberate, made in the first week of the project, after a power and coverage study.

If you are weighing LoRa against cellular for your next deployment, we have shipped both more than once.