eRedCap: The Cost-Reduced 5G IoT That Changes Everything (Release 18)

An automotive factory in Bavaria is deploying 10,000 vibration sensors across its production lines. Each sensor reports 50 KB of data every 30 seconds. Industrial WiFi was ruled out: too many access points, too much interference in a 40,000 m2 metal-walled production hall. The decision is to go cellular. The CTO wants 5G NR for the latency. The CFO looks at the per-unit module price: EUR 35 per standard 5G NR chipset, multiplied by 10,000 devices. EUR 350,000 in silicon alone. For sensors that need 500 kbps.

This is precisely the problem the 3GPP set out to solve with the cellular IoT hierarchy. And this is precisely where eRedCap, introduced in Release 18, changes the equation.

The cellular IoT hierarchy: from water meters to industrial robots

Cellular IoT is not monolithic. The 3GPP has built a technology ladder where each rung represents a trade-off between throughput, latency, cost, and power consumption. Understanding this hierarchy is the prerequisite for any architecture decision.

NB-IoT (Release 13, 2016)

NB-IoT operates within a 200 kHz band — a single LTE PRB. Maximum theoretical throughput is 250 kbps downlink, 20 kbps uplink (single-tone). The design targets stationary or near-stationary devices that transmit very small data volumes: water meters, temperature sensors, leak detectors. NB-IoT does not support handover or voice. The device attaches to a cell and stays there. In return, chipset complexity is minimal and battery life can exceed 10 years.

LTE-M / eMTC (Release 13, 2016)

LTE-M expands the bandwidth to 1.4 MHz and throughput to 1 Mbps. The fundamental difference from NB-IoT is support for mobility (handover) and voice (VoLTE). LTE-M is the natural choice for medical wearables, GPS trackers, connected alarms — any device that moves and may need to place a voice call. Chipset cost is higher than NB-IoT but remains in the EUR 5-8 range.

RedCap (Release 17, 2022)

RedCap (Reduced Capability) marks IoT’s entry into the 5G NR world. A RedCap device uses a 20 MHz bandwidth (FR1), with throughput capped at 150 Mbps downlink. It is a complete but simplified 5G NR device: a single receive antenna (1Rx instead of 2Rx or 4Rx), no carrier aggregation support, no multi-layer MIMO. RedCap targets IP surveillance cameras, service robots, high-end wearables, and high-throughput industrial sensors. Chipset cost sits around EUR 15-20 — significantly less than a full 5G NR modem, but still too expensive for high-volume low-throughput sensor deployments.

eRedCap (Release 18, 2024)

This is where the architectural innovation comes in. eRedCap (enhanced RedCap) caps throughput at 10 Mbps downlink and uplink. But the throughput reduction is not the central point. What changes everything is the dual bandwidth architecture.

The dual bandwidth innovation: how eRedCap halves the cost

To understand the eRedCap trick, you first need to understand why a 5G NR modem is expensive. The costliest component in a radio chipset is the baseband processing chain — specifically, the FFT (Fast Fourier Transform) processor that converts the RF signal into usable OFDM symbols. The wider the bandwidth to be processed, the faster and more complex the FFT processor must be, and the more the silicon costs.

A RedCap device (R17) receives and processes 20 MHz of bandwidth. Its FFT processor must handle the entirety of those 20 MHz. A standard 5G NR device in FR1 can process up to 100 MHz. Silicon cost is directly proportional to this bandwidth.

eRedCap introduces a split between RF bandwidth and data bandwidth:

RF bandwidth: 20 MHz. The eRedCap device’s RF frontend receives the full 20 MHz of the cell. This is necessary to decode the control and synchronization channels that occupy the full cell bandwidth:

• PSS/SSS (Primary/Secondary Synchronization Signals): initial time and frequency synchronization
• PBCH (Physical Broadcast Channel): basic system information (MIB)
• PDCCH (Physical Downlink Control Channel): scheduling, resource assignments, DCI
• SIBs (System Information Blocks): cell parameters, access configuration

Data bandwidth: ~5 MHz. For data traffic — PDSCH (Physical Downlink Shared Channel) on reception, PUSCH (Physical Uplink Shared Channel) on transmission — the device only processes a sub-band of approximately 5 MHz extracted from the 20 MHz band.

In practice, the device’s FFT chain processes the entire 20 MHz signal received by the RF frontend. But instead of decoding the full band for data, a sub-band extraction mechanism isolates the ~5 MHz assigned by the scheduler for the data channel. The downstream baseband processor — demodulation, channel decoding, HARQ — only processes this narrow sub-band.

The result: the RF frontend and FFT processor are sized for 20 MHz (required for signaling), but the data processing chain is sized for 5 MHz only. It is this data chain that dominates silicon cost in mid-range throughput designs. By reducing it from 20 MHz to 5 MHz, total chipset cost drops by over 50% compared to RedCap, targeting a price of EUR 7-10 per module.

Cellular IoT standards comparison

The table below summarizes the characteristics of the four 3GPP cellular IoT standards. Each row represents a selection criterion that the engineer must evaluate for a given deployment.

Criterion	NB-IoT (R13)	LTE-M (R13)	eRedCap (R18)	RedCap (R17)
Bandwidth	200 kHz	1.4 MHz	~5 MHz (data) / 20 MHz (RF)	20 MHz
Max DL throughput	250 kbps	1 Mbps	10 Mbps	150 Mbps
Max UL throughput	20 kbps (single-tone)	1 Mbps	10 Mbps	50 Mbps
Typical latency	1-10 s	100 ms - 1 s	10-50 ms	5-20 ms
Rx antennas	1	1	1	1
Mobility (handover)	No	Yes	Yes	Yes
Voice (VoNR/VoLTE)	No	Yes (VoLTE)	No	Optional
Duplex	HD-FDD	HD-FDD / FD-FDD	HD-FDD / FD-FDD	HD-FDD / FD-FDD
Estimated chipset cost	EUR 3-5	EUR 5-8	EUR 7-10	EUR 15-20
Radio technology	LTE (in-band / guard-band / standalone)	LTE	5G NR	5G NR
Primary use cases	Meters, static sensors, agriculture	Wearables, trackers, alarms	Semi-broadband sensors, advanced meters, industrial IoT	IP cameras, robots, advanced wearables
PSM / eDRX	Yes	Yes	Yes	Optional
Target battery life	10-15 years	5-10 years	3-7 years	1-3 years (often mains-powered)

PSM and eDRX: the key to battery longevity

For battery-powered IoT deployments, two 3GPP mechanisms are decisive: PSM (Power Saving Mode) and eDRX (Extended Discontinuous Reception). Both are available on NB-IoT, LTE-M, and eRedCap.

PSM (Power Saving Mode)

PSM was introduced in Release 12. The principle is straightforward: after a data transmission, the device enters a deep sleep state where it is effectively powered off from a radio perspective, but remains registered on the network. The device does not respond to network paging for the duration of the extended T3412 timer.

The maximum value of the extended T3412 timer is 413 days (timer unit = 320 hours, timer value = 31, yielding 31 x 320 = 9,920 hours = 413.3 days). During this sleep, power consumption is on the order of a few microamperes — comparable to the battery’s self-discharge rate.

The trade-off is clear: the device is unreachable for the entire PSM duration. The network cannot page it to push data. Only the device can initiate communication upon waking. For a water meter that uploads a daily reading, this is perfect. For a sensor that must be queryable on demand, it is incompatible.

eDRX (Extended Discontinuous Reception)

eDRX (Release 13) provides a compromise. The device alternates between sleep periods and brief paging listening windows. The maximum eDRX cycle is 43.69 minutes in connected mode (LTE-M) and 2.91 hours in idle mode (NB-IoT).

During the Paging Time Window, the device wakes up, listens to the paging channel, and goes back to sleep if no message is addressed to it. The listening window duration is configurable (typically 1.28 to 2.56 seconds).

The PSM + eDRX combination is the most common configuration for long-life battery IoT deployments: the device uses eDRX to remain reachable at regular intervals, and switches to PSM between scheduled transmission cycles.

Decision tree: which standard for which use case

Selecting the right cellular IoT standard is not a pure technology choice. It is a multi-variable optimization: required throughput, acceptable latency, mobility, voice needs, deployment volume, battery life, and operator coverage.

Choose NB-IoT if:

• The device is stationary or near-stationary
• Data volume per transmission is under 1 KB
• Multi-second latency is acceptable
• The deployment involves tens of thousands of devices at minimal cost
• Target battery life exceeds 10 years
• No voice or handover requirement

Examples: water and gas meters, temperature/humidity sensors in agriculture, leak detectors, parking sensors.

Choose LTE-M if:

• The device is mobile (trackers, vehicles, wearables)
• Voice is a requirement (alarms with voice callback, medical devices)
• Required throughput is between 100 kbps and 1 Mbps
• Handover between cells is necessary
• Sub-second latency is needed

Examples: medical smartwatches, logistics asset trackers, personal alarms, mobile payment terminals.

Choose eRedCap if:

• The device needs 1-10 Mbps throughput
• Integration into the 5G NR ecosystem is desired (same frequencies, same sites)
• Deployment volume is large enough for chipset cost to be a determining factor
• Mobility and latencies in the 10-50 ms range are required
• Battery operation is planned (3-7 years)

Examples: industrial vibration sensors with frequent reporting, next-generation smart meters, point-of-sale terminals, semi-broadband urban environmental sensors.

Choose RedCap if:

• Required throughput exceeds 10 Mbps (up to 150 Mbps)
• Sub-20 ms latency is needed
• The device transmits video streams or large data volumes
• The device is typically mains-powered or has a rechargeable battery

Examples: IP surveillance cameras, indoor service robots, AR/VR wearables, telemedicine gateways.

What this means for the field engineer

For the RF engineer or network engineer deploying and validating IoT networks, eRedCap introduces a new verification checkpoint in test procedures.

Dual bandwidth configuration verification. The eRedCap device must receive the full 20 MHz RF bandwidth for signaling but should only process a sub-band for data. When analyzing L3 traces, the engineer must verify that the RRC Reconfiguration correctly assigns the initial BWP (Bandwidth Part) for signaling and a reduced BWP for data traffic. If the device attempts to decode the full band for data, the behavior is non-compliant and power consumption will exceed specifications.

PSM/eDRX mechanism validation. The engineer must confirm that the timers negotiated during Attach/Registration (T3324 for active duration after transmission, extended T3412 for PSM cycle, eDRX cycle length) match the values requested by the device. A network that rejects the requested eDRX values and imposes shorter cycles will degrade battery life well below deployment specifications.

Infrastructure compatibility. eRedCap operates on the same 5G NR cells as standard devices. No new radio infrastructure is required. However, the gNB scheduler must support reduced BWPs for eRedCap, which implies a RAN software update. The field engineer must validate that the gNB software version supports eRedCap profiles before declaring a site operational for this device category.

The eRedCap ecosystem in 2026

First eRedCap chipsets are expected to enter sampling in the second half of 2026, with commercial availability projected for 2027. Qualcomm, MediaTek, and Unisoc have all announced roadmaps that include eRedCap. Operators deploying 5G SA today need to ensure that their RAN infrastructure will be eRedCap-compatible via software update — a point to negotiate in equipment vendor contracts now.

For the Bavarian factory from the beginning of this article, eRedCap changes the equation. Instead of EUR 35 per 5G NR chipset, the cost drops to roughly EUR 8. Across 10,000 sensors: EUR 80,000 instead of EUR 350,000. And the sensors are natively 5G, on the same private network as the robots and cameras running RedCap and standard 5G NR.

---

The technical terms used in this article — eRedCap, RedCap, NB-IoT, LTE-M, PSM, eDRX, BWP, PDCCH, PDSCH, PUSCH, FFT, PRB — are defined in our technical glossary. For a deeper dive into 5G NR architecture and signaling mechanisms, see our 5G SA articles.