6G: Building a Network That Can Sense, Learn and Extend Its Reach

Every generational transition in mobile networks brings familiar questions: What will it deliver? Who will pay for it? How disruptive will it be?

Much of today’s 6G debate circles around those same themes, often narrowing the conversation to whether the industry needs another cycle of higher peak rates or simply a more expensive version of what already exists. But that framing misses the real opportunity. The most meaningful advances in 6G will not come from more speed. They will come from sensing, coverage and intelligence — and from a Radio Access Network (RAN) that works very differently from the one we deploy today.

As early research and standardisation signals become clearer, three developments stand out: Integrated Sensing and Communication (ISAC); a new mid–high-band spectrum layer (FR3); and continuity of physical-layer waveforms between 5G and 6G. Together, these point to a powerful evolution that plays directly to more agile silicon, more capable Radio Units (RUs) and more intelligent, AI-native RANs.

1. ISAC: Turning the Network Into a Sensor

ISAC is widely expected to form a core pillar of 6G. ISAC enables the same transmissions to be used for both communication and sensing — mapping the environment, detecting motion, improving localisation and providing real-time spatial awareness.

ISAC matters not simply as a new capability, but because of where it leads.

A perceptive network can enhance safety-critical systems through better environmental awareness, support industrial automation with more precise localisation and tracking, improve energy and spectrum efficiency through context-aware operation, and provide world-model insights (internal representations of RF, environmental and positional context) that help the network interpret its surroundings and make more informed decisions.

In short, ISAC elevates the RAN from a passive transport mechanism to an active instrument capable of interpreting and tracking its surroundings.

2. The Missing ‘Killer App’ — and Why the 6G Narrative Feels Flat

Another reason the 6G conversation feels fragmented today is the search for a familiar pattern: a single, consumer-facing “killer app” that will justify investment. Because no obvious equivalent to mobile broadband or low-latency mobility has yet emerged, some industry experts question whether 6G has a clear rationale.

But this framing is increasingly unhelpful.

6G is unlikely to be defined by a single breakthrough application. Its value will come from foundational capabilities that enable many new services, not one: sensing and positioning embedded into the RAN; broader geographical coverage across land, sea, air and space; intelligent, agentic and AI-native base stations that adapt and coordinate in real time; a spectrum layer (FR3) that supports both advanced sensing and efficient mobility; and an evolutionary air interface that accelerates innovation rather than resets it.

The absence of a killer app signals not a lack of purpose, but a shift in what generational transitions mean.

6G is infrastructure-level transformation — the creation of a perceptive, ubiquitous and intelligent network foundation on which new applications can emerge organically.

3. FR3: The Spectrum Layer Where ISAC Becomes Meaningful

While ISAC can operate across many bands, FR3 (7–24 GHz) is emerging as a practical enabler. These frequencies offer significantly more throughput than FR1 while avoiding many of the deployment challenges associated with FR2 (mmWave).

Crucially, FR3 unlocks the resolution needed for meaningful sensing.

Shorter wavelengths, wider channels and more agile beamforming enable higher-precision range and angle estimation, richer environmental signatures, more accurate positioning and clearer sensing of structures and objects.

This combination shifts the focus of the RU’s design challenges.

In FR3, the RU becomes not only a transmitter and receiver but also a sensing device — one that requires higher dynamic range, tighter timing precision, more flexible radio front ends and intelligent beam control.

This aligns closely with RANsemi’s vision for the future: a software-defined, adaptable and AI-native radio architecture powered by programmable silicon and designed for sensing-rich, context-aware operation.

4. Ubiquitous Coverage, Not Just More Capacity

There is an important distinction between expanding coverage to more people and expanding coverage across a wider geographical area. 6G’s societal impact will come from the latter.

Extending the network’s physical footprint — across rural regions, industrial corridors, maritime environments and non-terrestrial domains — creates the foundation for sensing at scale. A network cannot perceive what it cannot reach. FR3 and ISAC together support a vision of ubiquitous, persistent coverage, enabling the network to function as an information-gathering layer for automated systems, logistics, defence and public safety.

This is not an argument for bigger towers or denser grids.

It is an argument for more efficient, more reliable and more intelligent RAN architectures that can operate autonomously, manage themselves in sparse or infrastructure-light areas and make local decisions in real time.

5. Continuity of Waveforms: An Evolution That Accelerates Innovation

Early signs from research groups and standards bodies indicate that 6G is likely to retain the foundational waveform family used in 5G NR, with enhancements rather than a wholesale redesign.

This continuity has wide-ranging implications: silicon does not need to be reinvented; PHY pipelines can evolve rather than be replaced; operators can adopt 6G capabilities incrementally; and innovation shifts from the air interface to intelligence, sensing and control — particularly within the baseband.

Most importantly, early ISAC features — even with limited capability — may be deployable on today’s silicon through software upgrades, long before complete 6G hardware cycles arrive. This accelerates learning, lowers risk and supports a phased transition aligned with operator investment cycles.

For a semiconductor baseband vendor, this evolution is far more valuable than a disruptive reset.

6. Agentic Base Stations: Where ISAC and Coverage Converge

If the network is to see more, understand more and cover more, the RAN must become more intelligent. Here, the concept of the agentic, AI-native base station becomes fundamental.

Agentic Base Stations (the per-cell processing functions within the RAN stack) learn from the cell’s local environment, build and refine world models (internal representations formed from RF conditions, sensing inputs and local context) and use local learning on local data to make real-time decisions. This is self-learning in the true sense: not models pre-trained elsewhere and pushed to the edge, but intelligence developed in situ and shaped by the cell’s own environment. These capabilities allow the base station to adapt continuously, interpret conditions more accurately and coordinate more effectively with neighbouring cells.

In this model, the base station becomes an intelligent agent rather than a passive endpoint, and much of 6G’s value shifts toward edge-based autonomy.

Conclusion: A Wider View of What 6G Really Represents

6G should not be judged by whether it produces a single killer app. Its value lies in the capabilities it adds to the network: the ability to sense, understand and cover more of the physical world with greater intelligence, reliability and efficiency.

It is still early in the 6G journey, but the direction of travel is increasingly clear. ISAC, FR3 and the continuity of waveforms all point toward a generation defined less by peak throughput and more by pervasive awareness, broader geographical reach and autonomous operation. Much of this capability may emerge on existing silicon through software, accelerating time-to-adoption. But this is only the first phase. Realising the full potential of ISAC, world-model reasoning and multi-agent coordination will require the next wave of innovation in the baseband and RU, with AI-native architectures that support learning, adaptation and edge autonomy.

For RANsemi, this trajectory aligns directly with our vision and roadmap for an intelligent AI-native RAN built on Agentic Base Stations — systems that learn locally on local data, refine their world models and adapt continuously to their RF and environmental context. All this marks a decisive shift in what radio infrastructure is capable of: a move from reactive operation to proactive, self-optimising behaviour at the edge.

A network that models its world will consistently outperform one that merely connects it — and the foundations for those networks are already being laid.

If you’re exploring what 6G will demand from RAN base station design, get in touch with our team.

Author: Oliver Davies, VP Marketing, RANsemi

Related content: The Case for Agentic Base Stations in 6G RAN