6G: The AI-Native, Sensing-Integrated Network
6G is not yet a standard — it is an active global research and pre-standardisation effort targeting commercial launch around 2030. Where each previous generation evolved radio technology, 6G aims to embed artificial intelligence directly into the network architecture, merge communication with environmental sensing, and extend connectivity beyond Earth's surface into a true 3D network fabric.
Why 6G — What 5G Cannot Do
- Terahertz spectrum — the sub-THz and THz bands (100 GHz–10 THz) offer hundreds of GHz of available spectrum that 5G never touches; unlocking them requires entirely new transceiver technologies
- AI-in-the-loop is too slow — 5G uses AI as a management overlay; 6G envisions AI as a native component of the air interface itself, making real-time decisions at every layer without human intervention
- Sensing and communication are still separate — 5G radio signals carry data only; 6G targets dual-function waveforms that simultaneously communicate and sense the physical environment (localization, imaging, mapping)
- Coverage gaps remain — despite satellite add-ons in 5G, large portions of Earth have no connectivity; 6G targets native 3D architecture integrating ground, aerial, and satellite layers from day one
- Energy consumption is growing — mobile networks consume ~2–3% of global electricity; 6G research targets 100× better energy efficiency per bit delivered compared to 5G
Terahertz and Sub-THz Spectrum
The 100 GHz–10 THz range contains hundreds of gigahertz of spectrum — enough for individual channel bandwidths of 10–100 GHz, theoretically enabling air-interface data rates of 1 Tbps at short range. Sub-THz (100–300 GHz) is the most near-term target, with the first 6G deployments likely using 100–300 GHz before higher frequencies become practical.
THz signals suffer from extremely high free-space path loss and atmospheric absorption (especially water vapour at 60 GHz and 183 GHz windows). Useful range is typically 10–100 metres at 300 GHz. Overcoming this requires ultra-dense deployments, graphene and InP-based transceivers, and intelligent reconfigurable surfaces to bounce signals around obstacles — all active research areas.
AI-Native Air Interface and Architecture
In 5G, AI is applied on top of a standard interface — for beam management, network optimisation, and anomaly detection. 6G research targets AI as an integral part of the interface design itself:
- AI-based channel estimation and equalization — neural networks replace traditional signal processing algorithms; the network learns optimal decoding strategies per environment and UE type
- Semantic communication — instead of transmitting every bit of a message, semantic systems transmit the meaning; an AI encoder at the transmitter and decoder at the receiver agree on a shared vocabulary, reducing required bandwidth by orders of magnitude for structured data and media
- Goal-oriented communication — the network optimises for task completion (e.g., “allow the robot to pick the correct part”) rather than raw bit-error rate; only information relevant to the goal is transmitted
- Distributed AI inference at the edge — 6G base stations are envisioned as AI compute nodes; model inference runs close to the UE, reducing cloud round-trip latency for AI applications to sub-millisecond levels
Integrated Sensing and Communication (ISAC)
ISAC is one of the defining features of 6G — the same waveforms used for communication simultaneously sense the physical environment:
The 6G base station transmits a standard communication waveform. Some fraction of that signal reflects off objects in the environment and returns to the base station (or a separate receiver). By analysing the reflected signal — Doppler shift, time-of-arrival, angle-of-arrival — the network can determine object location, velocity, and shape, without any additional radar hardware.
High-precision indoor positioning (<10 cm accuracy); gesture and activity recognition for smart spaces; autonomous vehicle environment mapping; drone detection and tracking; weather monitoring; and crowd density estimation for urban planning. ISAC turns every 6G base station into a distributed sensing platform — creating a real-time digital twin of the physical environment.
Reconfigurable Intelligent Surfaces (RIS)
RIS (also called Intelligent Reflecting Surfaces or IRS) are large panels of programmable meta-material elements that can dynamically control how radio waves reflect off them:
- What RIS is — a passive or semi-passive panel with hundreds or thousands of low-cost tunable elements (e.g., PIN diodes, liquid crystals, varactors) that shift the phase of incident radio waves; no RF transmitter or amplifier required
- What RIS does — by programming element phases in real time, an RIS can steer reflected signals toward a specific UE, cancel reflections that cause interference, or create virtual line-of-sight paths around obstacles
- Coverage without base stations — RIS panels deployed on buildings, ceilings, and bridges extend 6G coverage into deep indoor areas, tunnels, and non-line-of-sight zones at very low power cost
- Channel control — rather than adapting to the channel (as all previous systems did), 6G with RIS can actively shape the propagation environment itself — a fundamental philosophical shift in wireless network design
3D Connectivity — Ground, Aerial, and Space
Ultra-dense indoor/urban THz small cells for capacity; wide-area sub-6 GHz macro cells for coverage. AI-driven dynamic spectrum sharing across all layers.
UAV base stations provide on-demand coverage for events, disasters, and rural areas. HAPs (stratospheric platforms at 20 km altitude) serve regional coverage in a single hop, bridging terrestrial and satellite layers.
Mega-constellations of LEO satellites (200–1200 km) provide global broadband with latencies of 20–40 ms — complementing terrestrial 6G for maritime, aviation, polar, and deep rural coverage.
6G research targets a single unified air interface that adapts dynamically to the access layer — the UE seamlessly handovers between terrestrial, aerial, and satellite segments with no application-layer disruption.
Timeline and Standardisation
- 2020–2023 — Research phase — IMT-2030 vision defined by ITU-R; major 6G research programs launched (Hexa-X in Europe, NextG Alliance in North America, B5G in Japan, IOWN in Japan, 6G flagship in Finland)
- 2024–2025 — IMT-2030 requirements — ITU-R finalises the IMT-2030 (6G) framework, defining minimum requirements for peak data rate (1 Tbps), latency (0.1 ms), reliability, sensing accuracy, and AI/ML capability
- 2025–2028 — 3GPP pre-standardisation and Rel-20+ — 3GPP begins 6G feasibility studies in Rel-20 (starting ~2025); core radio and architecture specifications expected in Rel-21/22
- 2029–2030 — Commercial launch — first 6G deployments expected in leading markets (South Korea, Japan, China, US, Finland); initial deployments will likely target enterprise/industrial use cases before broad consumer rollout
- 6G will coexist with 5G — as with every generation, 6G will overlay rather than immediately replace 5G; LTE and 5G networks will remain active well into the 2040s