Release 19 Frozen, Release 20 Underway: What It Means on the Ground

It is 2 AM in a NOC somewhere in Western Europe. The main dashboard flags a degradation in attach success rate across a cluster of NTN cells. The on-call engineer pulls up the RRC logs, scrolls through RRCSetup and RRCReconfiguration messages, and spots IEs that were never there before in production. The explanation: the operator just activated the first Release 19 parameters on the core network. Welcome to the new RAN reality.

In December 2025, at plenary TSG#106, 3GPP officially froze Release 19. Three months later, Release 20 — the final 5G-Advanced iteration before 6G study items begin — is in full development with 126 active Work Items. For engineers who live between protocol traces and field KPIs, now is the time to sort out what is deployable today versus what to watch for tomorrow.

Release 19 — What Is Frozen and Deployable

Freezing a 3GPP Release means Stage 3 specifications — protocols, message formats, procedures — are stable. For R19 specifically, three pillars will impact field deployments starting in 2026.

5G Broadcast / MBS (Multicast-Broadcast Services). R19 finalizes the replacement of LTE eMBMS with a native NR mechanism. SFN (Single Frequency Network) mode enables broadcast over a wide geographic area without consuming unicast resources. Immediate use cases: enhanced public warning systems (PWS), event streaming in stadiums, OTA firmware updates for IoT fleets. On the protocol side, this translates to new IEs in MBSBroadcast and modifications to SystemInformationBlockType20 (TS 38.331).

RedCap Phase 2 (NR-Light enhancements). Introduced in R17, RedCap targeted mid-tier IoT devices — too complex for LTE-M but not requiring full NR capabilities. R19 pushes the envelope further (details in the next section).

NTN Phase 2 (Non-Terrestrial Networks). R19 introduces regenerative payloads — the satellite no longer just relays; it processes signals on board. Store-and-forward enables asynchronous communication in areas with no terrestrial coverage. For engineers, this means new timers in NAS procedures (TS 24.501) and ground-to-satellite handover mechanisms to validate.

This timeline is not just a history lesson. It shows an accelerating pace. Between R15 and R17, foundations were laid. Since R18, each release stacks operational capabilities. R19 is the first to deliver simultaneous advances in broadcast, IoT, and non-terrestrial — three domains that historically evolved in silos.

RedCap Phase 2 — Massive IoT Goes NR-Native

RedCap Phase 2 is arguably the R19 feature that will generate the most field activity. Why? Because it repositions NR as a credible platform for massive IoT — a market until now dominated by LTE-M and NB-IoT.

The numbers are straightforward. In R17 (Phase 1), RedCap operated on a maximum bandwidth of 20 MHz in FR1 — already a reduction from the 100 MHz of standard NR. R19 drops this to 5 MHz. That is a 20x reduction from full-capability NR, bringing RedCap into LTE-M territory in terms of modem complexity.

Beyond bandwidth, R19 delivers several critical optimizations for battery longevity:

• Relaxed MIMO — from 2 mandatory Rx antennas down to 1, reducing the device’s RF power consumption.
• Extended DRX (eDRX) — longer sleep cycles, allowing sensors to wake up only every few seconds or even minutes depending on the profile.
• Reduced transmit power — target consumption below 10 mW in connected mode for the most power-efficient profiles.

For field engineers, the immediate challenge is validating these new RedCap profiles on existing networks. SIB1 parameters change — the ue-CategoryNR-RedCap field and associated capabilities in UECapabilityInformation must be correctly decoded. A reliable Layer 3 decoder becomes essential for tracing these exchanges, especially when the first RedCap Phase 2 devices hit the field with sometimes incomplete implementations.

Key specifications to track: TS 38.306 (UE radio access capabilities) for new categories, and TS 38.331 (RRC) for procedure modifications.

Release 20 — The Final Stretch of 5G-Advanced

While R19 stabilizes, R20 is moving at full speed. With a Stage 3 freeze expected around March 2027, roughly one year of specification work remains. But the major directions are already clear — and some are transformative.

AI/ML for the RAN is the densest topic. Three use cases are being standardized: beam management — predicting the optimal beam without exhaustive measurement — , CSI feedback — ML-based compression and prediction of Channel State Information — and positioning enhancement — sub-meter localization through inference. These mechanisms introduce new signaling between UE and gNB (prediction reports, models exchanged via RRC). TR 38.843 and TR 38.844 document the feasibility studies; the normative specifications land in R20.

Ambient IoT may be the most radical disruption. The concept: battery-less devices powered by RF energy harvesting or backscatter. 3GPP is studying communications at a few tens of bits per second over distances of a few meters to a few tens of meters. The potential impact is enormous — smart labels, sensors embedded in building materials, maintenance-free logistics tracking. But standardization is still in the requirements definition phase (TR 38.848).

ISAC (Integrated Sensing and Communication) is the third transformative pillar. The principle: use NR signals not only for communication but also to sense the environment — motion, presence, velocity, distance. Think radar integrated into the cellular network. Use cases range from vehicle detection to industrial zone monitoring. The studies are documented in TR 22.837 (service requirements) and TR 38.857 (channel and signal study).

What This Actually Changes on the Ground

For a network engineer, the question is never “which Release is exciting?” but “what do I need to test now and what is going to break?”. Here is a pragmatic breakdown.

What R19 changes right now:

• New RRC IEs to decode. RRCReconfiguration and RRCSetup messages contain R19 extensions for MBS, RedCap Phase 2, and NTN. If your decoding tool does not support R19 ASN.1 schemas, you are flying blind. Verify that your ASN.1 decoder is up to date with the December 2025 schemas.
• Measurable RedCap KPIs. Attach success rate for RedCap devices, battery life in eDRX mode, average throughput at reduced bandwidth. These metrics need to be integrated into your dashboards now.
• NTN procedures to validate. Ground-to-satellite handovers and new NAS timers (extended T3XXX for satellite propagation delays) require specific end-to-end testing. The first commercial NTN networks are activating these R19 features throughout 2026.

What R20 requires you to prepare for:

• AI/ML traces. Model exchanges and prediction reports between UE and gNB will generate a new type of signaling traffic. Understanding these messages early — even from R20 early drops — will be a competitive advantage.
• ISAC test environment. Early ISAC implementations will require correlating radar-type measurements (range, velocity) with traditional network performance metrics (RSRP, SINR). No legacy tool is designed for this.
• Ambient IoT — passive monitoring. Ambient IoT devices do not generate traditional signaling. Detecting their presence and validating their operation will require entirely new monitoring approaches.

The R19-to-R20 transition is not a simple version increment. It is a paradigm shift in what the RAN is supposed to do — and therefore in what engineers must measure.

Conclusion

Release 19 is now a done deal. Specifications are stable, first deployments are arriving, and engineers who master the new IEs, RedCap Phase 2 profiles, and NTN procedures will have a clear head start. Release 20, meanwhile, redefines the boundaries of the RAN — artificial intelligence in the radio loop, battery-less devices, a network that “sees” its environment.

The question for every field team is straightforward: are your decoding and analysis tools ready for R19 message structures? And does your skills roadmap already include the AI/ML and ISAC fundamentals of R20?

The next freeze is coming faster than you think. The time for watching from the sidelines is over — it is time to prepare.