Skip to main content
Radio: Bands, RF & NR Evolution · 3GPP Quarterly December 2025

6G-era MIMO locks onto one-shot channel signals, but overhead cuts stay contested

Aperiodic tracking signals and explicit SCell activation won broad backing; a 1/8 reference-signal density and who schedules the uplink report are still fought over.

36 decisions46 work areas1,799 contributions

Plenary cycle

SA#110 · RAN#110 · CT#110

Baltimore · 8 Dec 2025 – 12 Dec 2025

Source meetings for this article (10)

Meetings in this article’s source material. Inclusion does not mean that every meeting approved every decision.

Samsung, Huawei and Ericsson want the network to hand a phone its first uplink grant for channel reporting directly inside the MSG4 connection message; NTT DOCOMO, Qualcomm, vivo, Spreadtrum, NEC and Lenovo want it scheduled afterward with a separate command. Nineteen companies back the first approach, six the second, and after a full quarter of discussion RAN1 still could not close the gap — the proposal was noted, not adopted, leaving chipset vendors designing against an unresolved uplink-grant procedure.

Companies converged on aperiodic-only signaling for the moments that matter most: right after a phone connects, and right after a secondary cell wakes up. Tracking reference signals will arrive in up to N bursts triggered by a MAC-CE in MSG4, with the exact burst count still open; twenty-one companies backed this over periodic alternatives, with only Futurewei and Panasonic dissenting. For secondary cells, the activation command itself will now explicitly name which SRS and CSI resources to use — a scheme backed by 21 companies against a single holdout, Spreadtrum, which preferred a simpler pre-configured approach.

For massive MIMO arrays of 64 or 128 antenna ports, companies agreed to allow a channel-reference-signal density as low as one resource element per eight resource blocks per port — half the overhead of the previous 1/4 floor. Apple and Qualcomm pushed back on accuracy grounds, so the compromise restricts the thinnest setting to wideband reports using the Release 19 Type-I codebook and makes it a separate, optional phone capability. That means not every 6G-era phone will support the lowest-overhead mode, and networks planning to exploit it will need to check device capability first.

Two related overhead-saving ideas failed outright. Mixing different densities within one aggregated CSI-RS resource set was rejected 18-to-7, with Huawei and InterDigital on the losing side against Samsung, Apple, Nokia and 15 others; cycling the frequency position of reference signals across transmissions lost 13-to-7, with Huawei and Nokia again backing it against a majority led by Samsung, Apple and Ericsson. Both would have squeezed more accuracy from sparse signals but added receiver complexity that the larger camp wasn't willing to accept.

The uplink-scheduling split between the Samsung/Huawei and DOCOMO/Qualcomm camps is the item to watch — RAN1's next meeting needs to pick MSG4-embedded scheduling or a post-MSG4 grant, because chipset teams can't finalize the connection procedure until that's settled. Separately, whether the 1/8 density extends beyond wideband reporting to subband and Type-2 codebooks remains for further study, which will decide how much overhead savings networks actually get in practice.

The decisions behind this

Deferred
NR MIMO Phase 6
R1-2508520

No consensus on mixing CSI-RS densities within a single resource set

A group of companies proposed allowing different frequency-domain densities to be configured for the individual CSI-RS resources (K NZP CSI-RS resources) that are aggregated within a single CSI-RS resource set for 48, 64, or 128 ports. The moderator's summary states: 'There is no RAN1 consensus to support different frequency-domain densities configured to the K NZP CSI-RS resources in the same CSI-RS resource set for 48/64/128 CSI-RS ports aggregation.' The list of companies supporting the proposal (Kyocera, InterDigital, vivo, LGE, NEC, Fujitsu, Huawei) and those opposing it (MediaTek, Nokia, HONOR, OPPO, Sharp, Xiaomi, Samsung, Apple, AT&T, Lenovo, NTT DOCOMO, ZTE, Ericsson, ETRI, CATT, TCL, Sony, Qualcomm) is provided. The document notes that company positions were unchanged and the conclusion reflects the current situation. The reason for the lack of consensus is not recorded in the provided materials.

Support mixing densities in a setOppose mixing densities in a set
Kyocera · InterDigital · vivo · LGE · NEC · Fujitsu · Huawei7 companiesMediaTek · Nokia · HONOR · OPPO · Sharp · Xiaomi · Samsung · Apple · AT&T · Lenovo · NTT DOCOMO · ZTE · Ericsson · ETRI · CATT · TCL · Sony · Qualcomm18 companies
Why it matters

The standard will not, at this time, allow a network to configure one CSI-RS resource with density 1/4 and another with density 1/8 within the same aggregated set. This keeps implementation simpler for devices but reduces network flexibility to potentially optimize overhead per individual resource. The issue returns to the next meeting without a decision.

Deferred
NR MIMO Phase 6
R1-2508518

Aperiodic TRS for initial access gets broad support, but no final decision

When a phone transitions from idle/inactive to connected mode, it needs reference signals for synchronization and channel estimation. A group of companies led by Samsung, Huawei, and Nokia proposed supporting aperiodic tracking reference signals (TRS) triggered by the network's MSG4 response in the random-access procedure. The proposal, labeled 1.1.1A, suggests allowing up to N bursts of these signals. A large coalition of companies including Samsung, NTT DOCOMO, Huawei, vivo, Nokia, Xiaomi, LG, OPPO, Spreadtrum, Qualcomm, ZTE, Sharp, CATT, Fujitsu, TCL, Ofinno, Sony, Apple, Lenovo, Ericsson, and Fainity expressed support. Only Futurewei and Panasonic did not support it. The meeting moderator noted that 'there is no RAN1 consensus' on supporting periodic or semi-persistent signals in this scenario, and the proposal for aperiodic TRS was not adopted as a final decision at this meeting.

The RAN1#123 meeting record records a partial agreement on this point: "For early TRS reception when UE transition from IDLE/INACTIVE to CONNECTED mode, at least support aperiodic TRS with up to Nmax TRS burst(s) triggered by a MAC-CE in MSG4 of 4-Step RACH."

Support aperiodic TRS triggered by MSG4Prefer other methods (e.g., periodic TRS)
Samsung · NTT DOCOMO · Huawei · vivo · Nokia · Xiaomi · LGE · OPPO · Spreadtrum · Qualcomm · ZTE · Sharp · CATT · Fujitsu · TCL · Ofinno · Sony · Apple · Lenovo · Ericsson · Fainity21 companiesFuturewei · Panasonic2 companies
Why it matters

If this proposal is eventually standardized, phones could receive more dynamic, on-demand synchronization signals immediately after connecting to the network, potentially speeding up the process of achieving high-quality MIMO channel estimates compared to waiting for periodic signals.

  • DeferredA group of companies proposed a method for early CSI reporting when a phone leaves idle modeR1-2508519
  • DeferredA coalition of companies failed to agree on supporting varying CSI-RS densities within a single resource setR1-2508519
  • DeferredLowest density CSI-RS (1/8) faces opposition, tied to conditionsR1-2508518
  • DeferredCompanies disagreed on a method for frequency-domain cycling of CSI-RSR1-2508519
  • DeferredNo agreement to link SRS and CSI reporting during initial accessR1-2508518
  • DecidedExplicit triggering favored for secondary cell activationR1-2508518
Quotations are verbatim from moderator summaries in the 3GPP archive. Document numbers link to the original files. Back to the full report.