Top 3 things that happened on NTN and satellite direct-to-device at RAN1 #126, RAN2 #135 and RAN4 #120
The verdict
Under the NR-NTN GNSS-resilience work (agenda item 9.4.1), RAN1 agreed which inputs a UE uses to pre-compensate uplink timing and frequency on the service link for initial access: a reference location plus satellite ephemeris in (quasi) Earth-fixed cells, and common TA and Doppler information in Earth-moving cells, cell-specific with one SSB beam per cell and SSB-beam-specific with several. (R1-2606759, chair notes, paragraph 57. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip; R1-2606759, chair notes, paragraph 65. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip)
RAN1 agreed what to consider, not values, for Msg2/RAR in initial access: frequency adjustment command range, granularity and sign, and an extended timing advance command range. Separately, RAN2 agreed to define a new downlink MAC CE for the frequency adjustment command, with an expiry timer; the timer design and whether the MAC CE also carries timing advance are FFS. (R1-2606759, chair notes, paragraph 73. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip; R1-2606759, chair notes, paragraph 82. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip; R2-2605832, meeting or session report, paragraph 798. R2-2605832, 3GPP TSG RAN WG2 #135. https://www.3gpp.org/ftp/tsg_ran/WG2_RL2/TSGR2_135/Docs/R2-2605832.zip)
Uplink pre-compensation for initial access (NR-NTN GNSS resilience). Reference location with ephemeris for (quasi) Earth-fixed cells; common TA and Doppler for Earth-moving cells. Service link only; whether a cell-specific value can be used with multiple SSB beams, multiple values per beam, and signalling are still open.
AgreementFor open-loop UL time/frequency pre-compensation on the service link in (quasi) Earth-fixed cells, for initial access support a reference-location-based approach together with satellite ephemeris where [...]
Timing and frequency commands (RAN1 for Msg2/RAR, RAN2 for a new MAC CE). RAN1 agreed what to consider for Msg2/RAR: a frequency adjustment command range, granularity and sign, and an extended timing advance command range with today's granularity as baseline. RAN2's agreement is a new downlink MAC CE for the frequency command; whether it also carries timing advance is FFS.
AgreementRAN2 to define a new DL MAC CE for FAC command, and a timer will be used to control its expiry. [...]
R2-2605832, meeting or session report, paragraph 798. R2-2605832, 3GPP TSG RAN WG2 #135. https://www.3gpp.org/ftp/tsg_ran/WG2_RL2/TSGR2_135/Docs/R2-2605832.zip
RAN4: 3 MHz channel and higher power classes (endorsed, not yet agreed). Endorsed documents add the 3 MHz channel for PC3 UEs in n250, n251 and n253 (Viasat), PC1 and PC2 at 3 MHz in the same bands (Viasat with Qualcomm), and the 3 MHz channel and PC2 in n252 (Apple with Terrestar and Skyworks).
RAN4: PC2 for band n254 (endorsed, not yet agreed). A CR adding the PC2 power class for n254 was endorsed (Apple with Amazon, Globalstar and Skyworks).
6G NTN link-budget scope (study). RAN1 agreed at least a set of evaluation scenarios by band and orbit. A note attached to the link-budget working assumptions says studying a scenario does not commit RAN1 to closing any coverage gap it finds.
AgreementFor the link budget template of 6G NTN, at least the following scenarios are evaluated NOTE: APC = “Adverse propagation conditions” [...]
Working assumptionR1-2606765, chair notes, paragraph 103. R1-2606765, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606765.zip
6G NTN satellite parameters (working assumptions). Working assumptions set MEO-7000 S-band parameters, including a maximum bandwidth per beam of 5 MHz and a 160 km reference beam, and LEO-1200 Ku-band G/T and EIRP density.
Working assumptionR1-2606765, chair notes, paragraph 128. R1-2606765, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606765.zip
Working assumptionR1-2606765, chair notes, paragraph 133. R1-2606765, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606765.zip
6G NTN study boundaries. Co-channel inter-satellite interference is studied at least for same-orbit LEO, assuming one satellite connection at a time. The GNSS-less, GNSS-free and degraded PRACH study runs in the common 6GR agenda item, with NTN-specific adjustments not precluded if the common design misses NTN requirements.
Should Rel-20 GNSS-resilience work serve as a baseline for 6G NTN?
MediaTek proposes Rel-20 as the baseline; LG Electronics calls it a useful baseline for uplink synchronization but flags wider 6G conditions; ST Engineering iDirect argues for GNSS-free initial access unconstrained by the Rel-20 study's design constraints. All were submitted to RAN1 and not treated.
MediaTekMediaTek proposes Rel-20 enhancements as the baseline for 6G GNSS-degraded and GNSS-less operation.
Proposal 8: Adopt Rel-20 NR NTN GNSS resilience enhancements as baseline for 6G NTN GNSS-degraded and GNSS-less/GNSS-free resilient operation.
LG ElectronicsLG sees a useful baseline for uplink synchronization, but flags UEs without GNSS and with degraded or outdated location.
Rel-20 GNSS-resilient NR-NTN provides a useful baseline for 6G NTN uplink synchronization, but 6G NTN may need to support broader conditions, including UEs without GNSS [...]
ST Engineering iDirectST Engineering iDirect proposes designing 6GR GNSS-free initial access without the Rel-20 study's design constraints.
Proposal 1 Design procedures and signals for GNSS-free initial access in 6GR NTN without the design constraints that were present during the Rel-20 study for [...]
How should measurements and mobility work without reliable UE location?
MediaTek points to a valid last-known location, while Sharp proposes studying time-based and beam-RSRP measurement conditions. ETRI identifies a need for non-location alternatives in inter-orbit mobility. These are company positions, not mobility agreements.
MediaTekMediaTek observes that a valid last-known GNSS location can sustain location-based mobility.
Location-based mobility like that in NR-NTN can works in GNSS-free/degraded scenario at least when valid last-known GNSS-based location is available.
SharpSharp proposes studying time and beam-RSRP conditions for measurement initiation and relaxation.
For GNSS-free operation, RAN2 should study time-based and beam-level RSRP-based conditions together with S-measure for NTN-specific measurement initiation and relaxation in connected and idle/inactive states.
ETRIETRI observes that alternatives not based on UE location are needed for basic inter-orbit mobility.
Observation 3: In multi-orbit NTN, UE GNSS location is used for at least three mobility-related functions: SMTC offset calculation, location-based measurement triggering, and inter-orbit coverage [...]
How should NTN-to-terrestrial downlink coexistence be modelled?
vivo addresses the simulation platform and multi-satellite aggregation, not the victim model. AST SpaceMobile, and Verizon with AST as co-submitter (AST is on both documents), argue that the existing interference-limited victim model hides adjacent-channel interference; Verizon adds a rural rationale. All were noted; no coexistence decision is in the evidence.
vivovivo proposes TR 38.863 Scenario 3 unchanged as the baseline, scaling the single-satellite adjacent-channel interference term for aggregation.
To adopt TR 38.863 clause 6.4.3 (Scenario 3) as the baseline simulation platform for the NTN-TN DL coexistence study, without structural changes, and to model [...]
AST SpaceMobileAST SpaceMobile argues for a noise-limited terrestrial victim model.
The Release-17 interference-limited model returns a 12 dB requirement and near-zero loss at every constellation size and therefore does not capture this interference path; [...]
Verizon (with AST)Verizon, with AST, says the TR 38.863 co-channel-limited victim model hides satellite adjacent-channel interference, giving too low an ACLR requirement for NTN.
Victim TN model in TR 38.863 is Co-channel Interference Limited, which does not represent the reality in sparsely populated Rural America. [...]
For both pre-compensation approaches, cell-specific values with multiple SSB beams and multiple values per beam are FFS, and signalling will come in a separate proposal. (R1-2606759, chair notes, paragraph 57. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip; R1-2606759, chair notes, paragraph 65. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip)
RAN1 left the random access occasion arrangement for two PRACH sequences FFS; RAN2's separate-resource working assumption does not resolve that question, and RAN2 still has to consider the listed MAC impacts. (R1-2606759, chair notes, paragraph 93. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip; R2-2605832, meeting or session report, paragraph 744. R2-2605832, 3GPP TSG RAN WG2 #135. https://www.3gpp.org/ftp/tsg_ran/WG2_RL2/TSGR2_135/Docs/R2-2605832.zip; R2-2605832, meeting or session report, paragraph 760. R2-2605832, 3GPP TSG RAN WG2 #135. https://www.3gpp.org/ftp/tsg_ran/WG2_RL2/TSGR2_135/Docs/R2-2605832.zip)
For S-band UEs in adverse propagation conditions, other losses has a bracketed 6 to 15 dB value for downlink broadcast evaluation, to be decided at the next meeting. (R1-2606765, chair notes, paragraph 141. R1-2606765, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606765.zip)
A noted offline discussion report, not a confirmed agreement, says time-based measurement and mobility can be supported at least in quasi-Earth-fixed 6G NTN regardless of GNSS status; applicability to all UEs or only Idle/Inactive UEs remains FFS. (R2-2605881, meeting or session report, paragraph 74. R2-2605881, 3GPP TSG RAN WG2 #135. https://www.3gpp.org/ftp/tsg_ran/WG2_RL2/TSGR2_135/Docs/R2-2605881.zip)
What to watch next meeting
Frequency adjustment command granularity: RAN1 plans to down-select between two ways forward at the next meeting, with combining them not precluded. (R1-2606759, chair notes, paragraph 95. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip)
RAN2 has a working assumption that separate random access resources can be configured for two PRACH sequences, while RAN1 keeps FFS whether they use the same occasion or separate occasions in time; watch whether the two groups converge. (R1-2606759, chair notes, paragraph 93. R1-2606759, 3GPP TSG RAN WG1 #126. https://www.3gpp.org/ftp/tsg_ran/WG1_RL1/TSGR1_126/Docs/R1-2606759.zip; R2-2605832, meeting or session report, paragraph 744. R2-2605832, 3GPP TSG RAN WG2 #135. https://www.3gpp.org/ftp/tsg_ran/WG2_RL2/TSGR2_135/Docs/R2-2605832.zip)
R1-2606759 Session Notes of AI 9.4. Session record; register lists Ericsson as editor. Chair notes. Register status: endorsed at R1-126.
R2-2605832 Report from session on MIMO, AIPHY, CovEnh, NTN and other topics. Session record; register lists CATT as editor. Meeting report. Register status: approved at R2-135.
R4-2610031 Introduction of the 3 MHz channel bandwidth for n250, n251, n253 PC3 UE. Viasat. Draft CR. Register status: endorsed at R4-120.
R4-2612717 Introduction of PC2 and PC1 for n250, n251, n253 for channel BW of 3 MHz. Viasat (with Qualcomm). Draft CR. Register status: endorsed at R4-120.
R4-2612716 Introduction of the 3MHz channel and the PC2 power class for the NTN band n252. Apple (with Terrestar, Skyworks Solutions). CR. Register status: endorsed at R4-120.
R4-2612715 Introduction of the PC2 power class for the NTN band n254. Apple (with Amazon, Globalstar, Skyworks Solutions). CR. Register status: endorsed at R4-120.
R1-2606765 Session Notes of AI 10.7 (NTN specific requirements and design). Session record; register lists Ericsson as editor. Chair notes. Register status: endorsed at R1-126.
R1-2606137 Discussion on 6G NTN specific requirements and design. MediaTek. Contribution. Register status: not treated at R1-126. No decision recorded in the register.
R1-2606591 Discussion on NTN specific requirements and design. LG Electronics. Contribution. Register status: not treated at R1-126. No decision recorded in the register.
R1-2605279 Proposals for GNSS-free 6GR NTN operation. ST Engineering iDirect. Contribution. Register status: not treated at R1-126. No decision recorded in the register.
R2-2604817 Discussion on 6G NTN. MediaTek. Contribution. Register status: noted at R2-135.
R2-2604790 GNSS-free mobility and NTN-specific measurement handling for 6GR NTN. Sharp. Contribution. Register status: available at R2-135. No decision recorded in the register.
R2-2604887 Multi-Orbit Mobility in NTN. ETRI. Contribution. Register status: available at R2-135. No decision recorded in the register.
R4-2611099 Discussion on NTN-TN co-existence aspects. vivo. Contribution. Register status: noted at R4-120.
R4-2610548 Initial results on aggregate NTN-NTN and NTN-TN DL-DL adjacent-channel coexistence. AST SpaceMobile. Contribution. Register status: noted at R4-120.
R4-2611983 On Victim TN Model for NTN-TN Coexistence in Adjacent DL Channels. Verizon (with AST). Contribution. Register status: noted at R4-120.
R2-2605881 Report of NTN GNSS free/resilience offline discussion. 3GPP RAN2. Meeting report. Register status: noted at R2-135.
Coverage: RAN1 #126 (Maastricht, 2026-08-24 to 2026-08-28), agenda items 10.7.1, 9.4.1; RAN2 #135 (Maastricht, 2026-08-24 to 2026-08-28), agenda items 9.6, 8.11; RAN4 #120 (Maastricht, 2026-08-24 to 2026-08-28), agenda items 5.5.2, 5.5.3, 5.5.4, 5.5.5, 8.4.3.2, 8.4.4. Scout read 48 recorded decisions and 561 company positions or contribution passages on these agenda items (catalog synced 2026-09-23) and cites 21 decisions and 9 positions from 17 TDocs. Every quoted passage was matched verbatim against the TDoc text; the summaries around the quotes are Scout's reading. Register statuses are copied from the 3GPP TDoc register: not treated means a document was not handled on its own, though its content may have fed feature-lead summaries or merged revisions. Not covered: other agenda items and meetings, plenary TDocs and plenary decisions after the meeting, email discussions after the meeting, and anything not yet in the chair notes or register.
Chair notes and session reports record group outcomes, not the views of their editors or moderators.
The frequency-command granularity formula in the chair notes did not survive text extraction, so the k values are not explained here.
The coexistence contributions shown were noted; no coexistence decision is in the evidence. Plenary outcomes after these meetings were not checked.
RAN4 register titles do not give numerical power limits or frequency ranges for the bands.