6G downlink waveform study narrows after Nokia and Apple beat back DFT-s-OFDM push

Nokia, Ericsson, Samsung, InterDigital, OPPO and Apple pushed to drop DFT-s-OFDM as a general 6G downlink waveform, arguing its lower power-amplifier stress isn't worth the added scheduling complexity; LG Electronics, CATT, Lenovo, Thales, Google and ZTE wanted to keep studying it, especially for satellite links. The moderator sided with the drop camp for phone-to-network transmissions like PDSCH and PDCCH, narrowing the debate to whether DFT-s-OFDM survives at all for non-terrestrial networks — a much smaller prize than the original proposal.

6G's timing structure is inheriting 5G almost wholesale: the 10 ms radio frame, 14-symbol slot and normal cyclic prefix all carry over from 5G NR, and 3GPP confirmed slot-based scheduling as the 6G baseline. The one crack in that continuity is a proposal, backed by the meeting moderator, to drop the extended cyclic prefix for general use entirely — pushing it into niche corners like satellite and sensing. For chipmakers, this means most of the 5G timing hardware and logic can be reused, with new engineering effort concentrated on genuinely new features rather than rebuilding the clock.

Control-channel coding shows the same pattern at a different layer: for messages within 5G's existing size limits (140 bits for DCI, 1706 bits for UCI), 3GPP agreed to reuse the 5G NR Polar code outright — a call backed by 16 of 20 and 16 of 21 contributing companies respectively. That leaves the real fight over messages that exceed those limits, where options on the table include Polarization-Adjusted Convolutional codes, 16QAM modulation, or a 1024-length Polar sequence, but RAN1 first wants confirmation from other work items that such oversized messages are even needed.

The sharpest open fight is architectural: how should a 6G phone handle a 400 MHz channel near 7 GHz, four times wider than typical 5G mid-band channels? Qualcomm, CMCC, Nokia, Ericsson and eight others want to study a single-RF-chain design using one 16k-point FFT; MediaTek, DOCOMO, Samsung and OPPO prefer two RF chains covering 200 MHz each; Apple, Xiaomi, vivo and Spreadtrum/UNISOC argue plain 200 MHz carrier aggregation, as used in 5G today, is good enough and a dedicated 400 MHz radio is overkill. The outcome will decide whether 6G devices need new, more expensive RF hardware just to use the widest 6G channels.

RAN4 is due to settle the smallest maximum bandwidth a low-tier 6G device must support — a choice among 5, 10 or 20 MHz — by June 2026, a decision that will set the floor for how cheap and power-efficient a 6G IoT sensor can be. Separately, watch for RAN#113 in September 2026, when the plenary resumes deciding whether 6G needs formal 'device types' at all, a choice that determines whether 6G repeats LTE's rigid UE-category system or moves to a more flexible, capability-based model.