Specifying cellular connectivity for an industrial device comes down to two questions, and they are easy to confuse. The first is a question of capability: how much data does this device need to move, at what power and cost? The second is a question of coverage: can it actually reach a network wherever it is deployed? eRedCap answers the first question. It says almost nothing about the second. Getting a deployment right means treating them as two separate axes, and this guide maps how they fit together.
Where eRedCap sits on the capability axis
eRedCap, or enhanced Reduced Capability, is the 5G NR device type defined in 3GPP Release 18, which was frozen in June 2024 as the first release of the 5G Advanced era. It builds on the Release 17 RedCap work with further reductions in complexity, bandwidth and power, and it is aimed squarely at the hundreds of millions of LTE Cat-1 and Cat-1bis connections that will need a successor as operators refarm 4G spectrum. Where standard RedCap supports up to 20 MHz in FR1 and peaks in the region of 150 Mbps, eRedCap caps peak throughput far lower, roughly 10 Mbps down and 5 Mbps up, and operates only in sub-6 GHz FR1 bands. That is a deliberate trade: less throughput in exchange for a smaller, cheaper, lower-power modem.
The clearest way to see eRedCap is as a rung on a ladder. Below it sit the low-power narrowband technologies for tiny, infrequent payloads. Above it sits full-fat 5G NR for genuine broadband. eRedCap and RedCap occupy the long-vacant middle, the tier that most industrial telemetry, tracking, metering and monitoring actually needs. Our full explainer on what eRedCap is covers the Release 18 specification in detail.
| Tier | Typical throughput | Best fit |
|---|---|---|
| NB-IoT / LTE-M | Tens to hundreds of kbps | Meters, low-rate sensors, long battery life |
| eRedCap (Rel-18) | Around 10 Mbps down | Cat-1 / Cat-1bis successor, mid-tier IoT |
| RedCap (Rel-17) | Up to roughly 150 Mbps | Routers, cameras, wearables, richer telemetry |
| Full 5G NR | Hundreds of Mbps to Gbps | Broadband, video, fixed wireless access |
Why capability is only half the specification
A device that can move 10 Mbps is worth nothing if it cannot see a network from where it sits. That is the trap in specifying purely on capability tier. eRedCap carries a coverage dependency that Cat-1bis does not: it requires a 5G Standalone core. In the UK, EE, Vodafone and Three are all deploying 5G SA, but coverage remains primarily urban, and the picture is similar across most of Europe. For any device shipping into rural, remote or mobile deployments, that means either an LTE fallback or a dual-mode design pairing Cat-1bis with eRedCap, so the device can keep working outside 5G SA footprint.
In other words, the moment you choose eRedCap you have inherited a coverage problem, and coverage is a different axis with its own answers. This is where a whole separate branch of the standard comes into play.
The coverage axis: terrestrial reach and non-terrestrial networks
Terrestrial 5G SA will keep filling in, but it will never reach the places where a great many industrial assets actually operate. Cellular coverage maps reflect where people live, not where pipelines, substations, farm equipment, buoys and long-haul fleets sit. For those assets, the answer is non-terrestrial networks: the 3GPP work, beginning in Release 17, that folds satellite connectivity into the same cellular standard a device already speaks, rather than bolting on a separate proprietary terminal.
NTN is the coverage half of the industrial IoT decision, and it deserves its own treatment. We recommend our companion analysis over on IoTPortal, which sets out how non-terrestrial networks bring satellite coverage to standard cellular IoT, the split between IoT-NTN and NR-NTN, and the direct-to-device commercial wave arriving through 2026. Read alongside this page, it completes the two-axis picture: eRedCap for capability, NTN for coverage.
Where the two axes converge: Ultra RedCap and satellite
The capability and coverage axes are not permanently separate. They are already starting to meet inside the standard. The RedCap family runs RedCap in Release 17, eRedCap in Release 18, and what the family calls Ultra RedCap in Release 19, the generation where NTN moves from a parallel workstream into the reduced-capability lineage itself. Release 19 introduces regenerative payloads that turn a satellite into a base station in orbit, store-and-forward operation suited to intermittently connected sensors, and device-to-device paths via satellite. For mid-tier IoT, that is the point at which a single modem can, in principle, reason about both axes at once.
The right architectural assumption for any new remote design is dual-mode by default. Assume the device will sometimes see a terrestrial network and sometimes a satellite, and that it will not know which in advance. Specify for both axes, not one.
What this means for specifying hardware now
The ecosystem is still forming. eRedCap modules began shipping in 2026, with volume availability expected from 2027, and pricing is on a downward trajectory but not yet at Cat-1bis parity. That argues for a staged approach rather than a single bet.
- Near term: specify LTE Cat-1bis for anything that has to ship and run reliably today, with an eye on 4G end-of-life planning.
- Medium term: evaluate first-generation eRedCap modules as they mature, prioritising dual-mode Cat-1bis and eRedCap designs for long-lifecycle products.
- Confirm 5G SA coverage in your actual deployment regions before committing to eRedCap as a primary bearer.
- Factor NTN into the roadmap now. Choosing NTN-capable silicon early is a small premium against the cost of a later redesign.
Specify on one axis and you will eventually be caught out by the other. The device that can move the data cannot always find the network, and the network that reaches everywhere does not always carry the throughput. eRedCap and non-terrestrial networks are the two halves of the same decision, and the deployments that age well are the ones designed with both in mind from the start.