Outcomes from the 3GPP RAN meeting in Fukuoka

Analysis of 3GPP TSG RAN #111, Fukuoka, Japan, 9–12 March 2026

An uncanny likeness, thanks Gemini…

The 3GPP plenary for the TSG RAN #111 in Fukuoka last month produced the usual high volume of technical content; a result of evaluating hundreds of stakeholder submission documents and the resulting discussions. It is always tricky to get a good read on progress, as these plenaries are extremely technical by nature and are part of an ongoing (never ending?) standards development process.

However, the meeting was more newsworthy than might first appear. There was no single landmark decision but the cumulative effect is significant. The TR 38.914 document is being steadily populated with endorsed requirements text across deployment scenarios, architecture, KPIs, spectrum, device types and services. The discussions that began a year ago are now converging into agreed positions, with moderators explicitly trying to close topics rather than keep them open.

This is my personal take on the material that I have seen; it is in no way an official account – I wasn’t there, I am only an impartial observer. But as I see it, three main themes emerged from the Fukuoka RAN meeting:

• the Rel-21 timeline is now taking shape, with real dates and real consequences;

• the architecture and migration debate has solidified around the politically-charged question of what role does 5G plays in a 6G world;

• and the NTN-on-terrestrial-bands debate rolls on (or should that be “completes another orbit?), still unresolved but now at least channelled into an approved study framework.

Let’s take them one at a time, and yes we will be getting into the weeds somewhat with this post! Please do bear in mind, any errors or assumptions are mine: there was a huge amount of documentation generated at the meeting and it was impossible to sift it all.

1. Rel-21 timeline

The joint RAN/SA/CT session (all three of the 3GPP TSGs) on the Release 21 timeline was perhaps the meeting’s most strategically important development. The three TSG Chairs presented a unified proposal with five headline dates:

• 5G-A/6G Package Approval: March 2027

• Stage 1 freeze: March 2027

• Stage 2 freeze: March 2028 (≥80%), June 2028 (100%)

• Stage 3 freeze: December 2028

• ASN.1/OpenAPI freeze: March 2029

While the full timeline won’t be formally approved until TSG #112 in June, the Fukuoka meeting endorsed the first of these milestones: Stage 1 freeze in March 2027.

March 2027 is also when Rel-20 Stage 3 is scheduled to freeze, meaning the same meeting will need to close out both the current release’s protocol work and approve the Rel-21 work packages.

The TSG Chairs are proposing that Rel-21 will span roughly two years of normative work, from package approval in March 2027 through ASN.1 freeze in March 2029. It’s tight, but so was the 5G work a decade ago. It was suggested that the load could be eased by splitting 5G and 6G package approvals, given that the 5G content in Rel-21 should be lighter. The RAN Chair acknowledged this but is planning for a five-day meeting and expects the 5G Rel-21 package to be substantially smaller than that for Rel-20.

More interesting is what happens between now and then. Some operators proposed that the quarter between March and June 2027 should be used specifically to evaluate 6G performance against 5G Rel-20, essentially a benchmarking window to demonstrate that the new radio is actually better than what it’s replacing. After all, if 6G can’t demonstrably outperform the latest 5G in measurable ways, the business case for deployment weakens considerably. After some debate, it appears likely that the reference release for comparison will be chosen on a case-by-case basis. Need to keep a watch on this one.

It is worth bringing attention to the ASN.1 freeze in March 2029. This caused some disagreement between vendor stakeholders – those who think this will be challenging and wanting to push it to June 2029, and those saying “we should not yet plan with a potential delay.” The ASN.1 freeze is the date that chipset and device vendors work backwards from, and any uncertainty here could spread through the whole ecosystem.

What this means:

The industry now has a complete proposed date set for the release that will contain normative 6G specifications. The process is: package approval and Stage 1 freeze in March 2027, architecture freeze across 2028, protocol freeze in December 2028, and chipset-ready ASN.1 encoding by March 2029. That’s roughly two years from normative work starting to implementation-ready specifications.

What to watch out for at RAN #112: will the ASN.1 date hold? If it doesn’t, what will be the knock-on effect for commercial rollout?

---

2. Architecture and migration

The architecture and migration proposal discussion went through four revisions before reaching an endorsed text, and the debate exposed the industry’s disagreement about how 6G will be deployed.

Ahead of the meeting, 18 operators submitted a joint document for consideration expressing their desired requirements. It was a global grouping of telco heavyweights, although it excluded any China operators. And it worked. The final moderator draft summary reveals that the operators’ proposed text was adopted into the TR essentially verbatim. Every requirement in the operator contribution appears in the agreed TR text.

This matters because these requirements encode specific operator priorities that will shape the 6G system for a decade or more. Looking at the highlights:

Standalone first: The very first requirement, “the 6G RAN architecture shall support standalone RAN architecture”, is listed before anything else. After the NSA experience with 5G, operators are making standalone the non-negotiable baseline.

Operator control of data: Three related issues establish that operators control what data is collected, have visibility into it, and can start or stop collection. This is a direct response to the AI/ML data collection debates that both the RAN and SA groups held during 2025. Operators want to write their data governance requirements into the architecture itself, not leaving them to be negotiated one feature at a time.

Lower cost, not just higher performance: Two requirements explicitly mandate lower CAPEX/OPEX and lower energy consumption “with respect to current networks.” This is unusual as 3GPP requirements typically focus on capability, not cost. The operators are stating what we all feel - that 6G won’t be funded by the same investment enthusiasm that launched 5G. The business case for 6G must be better than for existing 5G, not just the technology.

Open interfaces: “3GPP-defined interfaces for 6G RAN shall be open for multi-vendor interoperability” appears as a core requirement, not an optional feature or a later addition. This reflects the Open RAN debate of the 5G era recast as a first-principles requirement for 6G.

4G coexistence: The text includes idle-mode inter-RAT mobility with 4G and coexistence with NB-IoT and eMTC. This is a tricky one. Some operators genuinely believe that 4G will be gone by the tie 6G appears, but many more maintain that these legacy networks will persist far longer than some ambitious roadmaps suggest. Chalk one up for the pragmatists.

At its core, though, the discussion was about three migration options beyond the baseline MRSS approach: options for how 6G radio aggregates spectrum with the existing 5G network. The final moderator text defines them precisely as:

• Option 1: 6G RAN node as anchor, aggregated with NR cells (e.g. deploy 6G via MRSS on FR1/low bands, and aggregate with NR bands)

• Option 2: 5G RAN node as anchor, aggregated with 6G cells (e.g. deploy 6G as secondary node on new FR3 band, aggregating with NR FR1/low bands)

• Option 3: No RAN-level aggregation — 5G RAN connected to one core network, 6G RAN connected to another, with coordination above the RAN layer (a “dual stack” approach)

The framing is more precise than the “NSA vs SA” argument that caused friction with 5G. It’s about where the anchor sits and where the aggregation happens. Option 1 puts 6G in charge and lets it use 5G spectrum. Option 2 puts 5G in charge and adds 6G capacity on top. Option 3 avoids tight coupling altogether.

“A big mistake of 5G was NSA.” So said one of the leading US operators, and no doubt endorsed by many more. This is a reference to the Non-Standalone architecture that dominated early 5G deployments, where 5G New Radio was bolted onto 4G LTE core networks. NSA allowed faster time-to-market but created years of operational complexity, confused consumers about what “5G” actually meant, and left operators running dual stacks far longer than planned. Which is why a number of operators were vocal in their opposition to Option 3, “to avoid the 5G mistake.”

But not all operators think this way. Those in markets with older 5G infrastructure will deploy their first 6G sites on top of those legacy networks, making some form of Option 3 necessary for initial rollout. So in theory, standalone is better; but in practice, migration requires coexistence.

By the end of the meeting, Option 2 was off the table. It now comes down to Option 1 (6G-anchored dual connectivity with NR) and Option 3 (dual stack, no RAN-level aggregation), with a final decision made by September 2026. The RAN1 and RAN2 groups will now study 6G-6G spectrum aggregation for non-collocated deployment scenarios (i.e. can a 6G anchor efficiently aggregate carriers across sites?), providing the technical groundwork for a vote on Option 1.

Linked to this was the discussion on Multi-RAT Spectrum Sharing (MRSS), where 5G and 6G share the same frequencies. Every operator deploying 6G will need to share spectrum with their existing 5G network for many years. If the overhead of running both simultaneously is too high and inefficient, then the business case for 6G weakens. Some operators were adamant that 3GPP finds out by September just how much overhead MRSS will cost, because this information feeds directly into migration planning and investment decisions. The proposal was endorsed and RAN1 will deliver a preliminary MRSS performance evaluation by September. If the numbers don’t work, is there a “Plan B”? Er...

The question of coexistence with 4G, was partially endorsed. Most, but not all, legacy design choices will be accommodated, such as support for both idle mode and connected mode inter-RAT mobility with E-UTRAN, and coexistence with NB-IoT and eMTC.

What this means:

The architecture and migration debate is the indicator for how fast and how expensively 6G will be deployed. Standalone purists are trying to prevent a repeat of the 5G NSA era, whereas pragmatists recognise that real-world deployments will involve coexistence. At least 5G-anchored dual connectivity is off the table.

Watch out for a September 2026 deadline ( 3GPP RAN #113) for MRSS overhead numbers, and an explicit study mandate for non-collocated spectrum aggregation. If RAN1 can show that control channel sharing brings MRSS overhead down to acceptable levels, the standalone-with-MRSS path becomes the favourite. If it can’t, the dual connectivity and dual stack options will be in the running.

---

3. NTN on terrestrial bands

The NTN-on-terrestrial-bands debate went through four revisions of the moderator’s summary before reaching approval, illustrating just how tricky this aspect has already become. The RAN #110 meeting in Baltimore agreed in principle to scope a coexistence study but left the details unresolved.

The core issue hasn’t changed: satellite operators and some terrestrial operators want 3GPP to define NR-NTN bands using terrestrial IMT frequencies, citing national regulatory mandates and commercial urgency. However, other terrestrial operators worry about interference to their networks and argue that 3GPP should not act ahead of ITU-R processes. It is the ITU that handles global spectrum allocations.

At Baltimore, a liaison statement to ITU-R was agreed but deliberately left open on some key points. By Fukuoka, ITU-R WP5D had acknowledged the issue but not yet responded substantively, with nothing expected before October 2026 (apparently there are a lot of documents involved...).

The pro-NTN camp argued that waiting for ITU would cost another year of commercial delay. The opposing camp argued that LEO satellites crossing borders every six seconds create a fundamentally different interference scenario from terrestrial networks, and that 3GPP cannot enforce coexistence rules — that remains an ITU/regulatory function.

A phased approach was therefore prosed and, after some heavy revisions, finally approved, although it is somewhat narrow in scope.

It is not really the coexistence study that everyone wanted; it’s more of a regulatory survey of what national, regional and international rules exist for operating satellite services on three specific NR bands (n1, n7, n25). It doesn’t start coexistence simulations, doesn’t define interference thresholds and doesn’t reserve band numbers. But it does provide a structured basis for the June 2026 plenary to decide what comes next, informed by actual regulatory facts.

The cross-border interference problem is genuinely novel, with no pre-existing regulatory guidance in many regions, including Europe. Crossing countries every six seconds is unprecedented. Also of note, proprietary implementations (which would be harder to regulate) are likely to happen if 3GPP stakeholders can’t come to an agreement – these companies are not going to wait. But then aren’t we all ultimately waiting for the ITU?

What this means:

The NTN-terrestrial debate has moved from “should we study this” to “what exactly do we study?” The answer, for now, is not a technical study, but a regulatory map. The approved survey is deliberately modest: it gathers facts without making technical commitments. For the pro-NTN camp, this is slower than they wanted but keeps the work alive within 3GPP. For the more cautious camp, it avoids the risk of 3GPP replicating or pre-empting ITU-R work. The ITU WP4C April meeting and WRC-27 preparations will determine whether the June checkpoint leads to a genuine coexistence study or further delay.

---

What to Watch

Looking ahead to RAN #112 (June 2026), the key items are:

1. Rel-21 timeline formal approval. The June meeting will set binding dates. Expect the ASN.1 date to attract the most debate.

2. MRSS overhead data. The September 2026 deadline (RAN #113) for RAN1’s preliminary MRSS performance evaluation is now formally mandated. We might have some early evidence on this.

3. Migration option down-selection. Also due by September, the choice between 6G-anchored dual connectivity with NR (Option 1) and dual stack (Option 3) should be made. This might, perhaps, happen at the June meeting.

4. NTN regulatory survey completion check. There will be a June check on whether the survey can be declared completed. If it can, a coexistence study scope will likely be proposed.

5. Device type resolution. The RAN Chair wants this decided; it has now been unresolved across three consecutive plenaries.

6. TR 38.914 target completion. June is the original target for the 6G scenarios and requirements. Will studies be complete?