6G Has a Date Now. Your RF Front End Has a Deadline.

A base station radio, fixed wireless terminal, or private-network small cell that enters architecture definition in 2027 will likely still be in the field in 2035, well after the first 6G networks are expected to launch. In June 2026, 3GPP approved the detailed Release 21 freeze schedule, which has the first 6G specifications at final freeze in March 2029 and gives teams firm dates to plan against.

In existing bands, some infrastructure may reach 6G through software upgrades, depending on its baseband and radio capabilities. The RF front end for new spectrum is the exception: the candidate bands around 7 GHz call for band, array, and material decisions that come early in a design. 

Key Takeaways

• 3GPP has set the Release 21 freeze schedule: The first 6G specifications freeze between late 2028 and March 2029, and commercial rollout forecasts cluster around 2029 to 2030. 
• Much of the 5G air interface carries over into 6G: Baseline choices for waveform, coding, modulation, and frame structure are in place, so some existing-band infrastructure may migrate through software upgrades.
• New spectrum around 7 GHz requires new hardware: Channels up to 400 MHz and much larger antenna arrays call for a new RF front end.
• Any radio platform with a service life past 2030 now has a firm planning horizon: For products exposed to the upper mid-band spectrum, candidate band and subband selection should be early architectural decisions.

The 6G Calendar Is Set

3GPP set the schedule at its June 2026 Radio Access Network (RAN) plenary in Singapore. Release 21, the first release expected to define 6G, moves from radio work item approval in March 2027 to a physical-layer freeze in September 2028, a protocol-layer freeze in December, and the final ASN.1 freeze in March 2029. Ericsson expects the first commercial systems around 2030, while Qualcomm places commercial availability in 2029 to 2030.

What 6G Retains From 5G

The decisions made during the 2025 study phase and confirmed at the June plenary show deliberate continuity. CP-OFDM remains the downlink baseline, while the uplink supports both CP-OFDM and DFT-s-OFDM. Low-density parity-check (LDPC) codes continue for data, with a third base graph aimed at improving decoder area efficiency, while Polar codes remain the control-channel baseline. Uniform QAM remains the modulation foundation, with higher-order modulation and constellation shaping only partly settled after the September plenary.

The frame structure also remains closely aligned with 5G NR, supporting dynamic spectrum sharing across generations. The architecture is standalone from the start, avoiding the 4G-core dependency of early 5G non-standalone deployments. Multi-RAT spectrum sharing is the agreed baseline for running 5G NR and 6G in the same carrier. Whether 3GPP adds a further migration option, such as 6G-anchored dual connectivity with existing 5G carriers, was deferred at the September plenary to a December decision.

Uplink Gets the Real Upgrade

Ericsson’s June 2026 Mobility Report found uplink traffic growing faster than downlink for 43 of 55 measured operators, with 17 seeing uplink grow at more than 1.5 times the downlink rate. Video calls, user-generated content, cloud backups, and increasingly AI applications all push more data off the device. Ericsson’s urban-area scenario modeling suggests that AI adoption could bring uplink traffic to three times its 2025 level by 2031 under medium-adoption assumptions, or five times under high-adoption assumptions.

The uplink trend helps explain one of Release 21’s few waveform changes. 6G extends DFT-s-OFDM from 5G’s single-layer uplink use to MIMO operation with up to two layers, increasing potential uplink throughput while retaining its power-efficiency advantages.

Why 7 GHz Means New Hardware

Qualcomm has been specific about what carries over and what needs new hardware. Deployments in existing frequency bands may be possible through software upgrades, provided the installed infrastructure supports the required capabilities. New spectrum around 7 GHz requires new RF front-end designs and supporting hardware.

The industry commonly uses Frequency Range 3 (FR3) as shorthand for the upper mid-band region between today’s FR1 and FR2, roughly 7 to 24 GHz. Within that span, ITU-R is studying specific ranges for possible International Mobile Telecommunications (IMT) use ahead of the 2027 World Radiocommunication Conference (WRC-27), including 7.125 to 8.4 GHz and 14.8 to 15.35 GHz, and near-term 6G attention is focused on the lower end.

3GPP has aligned on network bandwidths up to 400 MHz, with devices supporting up to 400 MHz downlink and 200 MHz uplink. Typical 5G mid-band deployments use 100 MHz channels, so the fourfold jump affects filter design, converter rates, PA linearity budgets, and validation requirements.

Propagation gets harder at these frequencies, but the shorter wavelength also supports much denser arrays. Over the same link distance and with equal antenna gains, free-space path loss is about 6 dB higher at 7 GHz than at 3.5 GHz, and about 9 dB higher at 10 GHz. Qualcomm’s 6 to 8 GHz Giga-MIMO work uses that antenna density and narrower beams to recover part of the propagation penalty while adding spatial capacity. 

Three Hardware Challenges Created by 7 GHz

Each of the three forces a decision that is cheap during architecture definition and expensive after component qualification starts.

1. FR3 Front Ends Are Designed Band by Band

An RF front end targets a defined subband or a small set of adjacent bands, because PA efficiency, matching networks, filters, antennas, and packaging all change with frequency, and regulatory allocations differ by region. A design tuned for 7.125 to 8.4 GHz and one tuned for 14.8 to 15.35 GHz are different front-end projects. Band selection, therefore, comes early in architecture definition, ahead of component qualification and layout. 

The device technology for these bands is not settled either. Imec has demonstrated a GaN-on-silicon E-mode transistor achieving 66 percent power-added efficiency at 13 GHz and 5 V, targeting future FR3 mobile applications. RFIC 2026 featured a full-day workshop on FR3 power amplifiers, covering silicon, GaAs, GaN, broadband architectures, and digital predistortion. 

2. Antenna Count Scales Up

Qualcomm’s 7 GHz Giga-MIMO prototype base station uses 1,024 antenna elements and 256 digital ports, with the companion prototype device supporting up to eight downlink layers and four uplink layers. 3GPP doesn’t prescribe an antenna count, so the prototype is a reference point for scale.

This scale creates new work in calibration, RF distribution, signal processing, thermal design, power delivery, and validation. Test plans built around conducted measurements and modest arrays will need over-the-air (OTA) capability, which should be in your project plan before bring-up.

3. Materials and Integration Move Up the Critical Path

As operating frequency and bandwidth rise, dielectric loss, conductor roughness, dielectric-constant stability, stackup tolerances, and interconnect length consume more of the RF loss and phase budget. At that point, a drawing note that reads “low-loss laminate” no longer specifies a material.

Released designs should document the validated laminate system, copper profile, prepreg, and critical stackup parameters, with any substitutions verified against the same budget. The same pressure pushes more RF integration closer to the antenna-module boundary, where every millimeter of interconnect adds loss.

What to Do Before March 2027

Four tasks come before March 2027, when 3GPP approves the Release 21 radio work items.

• Map band exposure across the portfolio. Products confined to existing bands are candidates for software migration. Products headed for upper mid-band spectrum belong on a hardware roadmap now.
• Pick a candidate band for the target market early. It sets the PA, filter, antenna, and material stack, so plan a checkpoint for spectrum availability after WRC-27 and national allocations. Changing it after qualification starts can force redesign and requalification of all four.
• Check which front-end parts are already rated above 7 GHz. Look up the PAs, filters, switches, and LNAs on the current BOM in the Octopart BOM Tool and compare their rated frequency ranges with the target band. Any part rated below the band needs a replacement, which you should schedule now.
• Tighten materials documentation and budget OTA validation before prototype bring-up.

Modulation refinements and 5G-to-6G migration are still open in 3GPP, with the migration decision now due at the December plenary, and the 7 GHz ranges remain candidates until WRC-27 and national regulators act. Frequency, bandwidth, and array size are settled enough to plan against now, with checkpoints as the standard and the spectrum decisions firm up.

The Supplier Question

While 3GPP was setting the 6G schedule, the supplier base for RF power devices was getting smaller. The next article in this series, Who Still Makes RF Power? The 2027 Supplier Map, identifies the suppliers still positioned to build the front ends that upper mid-band 6G will require. The third, The Gallium Problem in Your Radio: Export Controls and Single-Source Front Ends, covers the export-control and sourcing risks for the gallium-based devices in those front ends.