5G FWA (Fixed Wireless Access) Overview
Fixed Wireless Access (FWA) uses radio links rather than fiber or copper to connect Customer Premises Equipment (CPE) to an operator’s network, offering broadband speeds without trenching or cable deployment.
While 4G FWA provided up to 100 Mbps in ideal conditions, 5G FWA leverages new spectrum (mm waves), massive MIMO, and network architectures to deliver multi-gigabit throughput and sub-millisecond latency, positioning it as a competitive alternative (or complement) to fiber in both dense and underserved areas.
Contents
Driver factors
- Fiber Roll-out Constraints:
- mmWave licenses enabling localized gigabit deployments in dense urban hotspots
- Explosive residential broadband demand fueled early 5G FWA trials and POC
5G FWA Use cases
- Enterprise: Dedicated slices and private FWA networks deliver secure, high-performance connectivity to campuses, factories, and temporary sites.
- Urban/suburban: Mid-band deployments on rooftops and light poles serve dense neighborhoods, often with combined sub-6 GHz and mmWave sectors.
- Rural/low-density: Edge-cloud architectures powered by renewable energy and Open RAN reduce fiber trenching costs, using higher towers and directional antennas to cover kilometres economically.
Core Architecture of 5G FWA

A typical 5G FWA architecture includes:
- Customer Premises Equipment (CPE): Outdoor or indoor units with integrated antennas, often supporting both sub-6 GHz and mmWave bands.
- Radio Access Network (RAN): 5G gNB sites equipped with massive MIMO arrays and beamforming, dynamically steering beams toward each CPE.
- Transport/backhaul: Fiber or microwave links to connect gNBs to the 5G core, sometimes augmented with multi-hop wireless transport in rural deployments.
- 5G Core (5GC): A service-based architecture supporting control/user plane separation, network slicing orchestration, and edge computing for ultra-low-latency services.
* Control Plane (CP): AMF (access), SMF (session), PCF (policy)
* User Plane (UP): UPF edge nodes for local breakout and MEC
* Network Slicing: NSMF, NSSMF managing slice templates (RAN + Core + transport)
Spectrum Considerations
5G FWA deployments exploit a mix of bands:
- Low-band (<1 GHz) offers broad coverage but limited capacity.
- Mid-band (1–6 GHz): the sweet spot for FWA, balances range and throughput, enabling shared-spectrum access for new entrants.
- High-band/mmWave (>24 GHz) delivers fiber-like speeds (1–3 Gbps) but requires dense site placement due to limited propagation.
Key Enabling Technologies
- Massive MIMO & beamforming: Arrays of 64–128 antennas form narrow beams, boosting spectral reuse and signal robustness.
- Carrier aggregation (CA): Combines multiple carriers (sub-6 and mmWave) to scale capacity linearly with aggregated bandwidth.
- Network slicing creates isolated end-to-end virtual networks tailored for FWA, providing guaranteed throughput, prioritization, and SLAs. Operators can monetize slices via tiered residential, premium gaming, and enterprise plans.
- Mobile edge computing (MEC): Hosts applications at the network edge to minimize round-trip latency for gaming, AR/VR, and IoT.
- Virtualization & O-RAN: Disaggregated RAN functions split into CU/DU/Radio Units with open interfaces, enabling multi-vendor interoperability and cloud-native scaling.
Summary:
Parameter | Typical Value | Notes |
Spectrum Bands | Sub-6 GHz: 3.3–4.2 GHz | 50–100 MHz carriers; mmWave: 24–39 GHz with 200–400 MHz carriers |
Channel Bandwidth | 100 MHz (sub-6) | Up to 400 MHz (mmWave) via contiguous or CA |
Peak DL/UL Throughput | DL: 1–4 Gbps; UL: 200–500 Mbps | Depends on MIMO layers and CA |
Latency (air interface) | 0.5–1 ms TTI | Mini-slot scheduling (2-4 symbols) |
MIMO Layers | 64×64 (sub-6); 256×256 (mmWave) | Hybrid beamforming with dynamic panel selection |
Duplex Mode | FDD/TDD | TDD common in mid/mmWave for flexible DL/UL ratios |
Scheduler | Dynamic grant-free | Mix of Type 1 (scheduled) and Type 2 (configured grants) for URLLC |
Network Slice Bandwidth | SLA-based (e.g. 100 Mbps) | RAN slice scheduler enforces minimum and maximum throughput per slice |