Model-neutral spectrum: beyond MNOs

A huge potential diversity of new licensees and stakeholders means both opportunity & challenge for spectrum policymakers

This post was originally published on my Substack newsletter on August 15 2025. It is being syndicated here. Select future posts of this type will be Substack-first, or Substack-only, so please subscribe!

Overview

Regulators should structure future spectrum releases to stimulate interest and investment from multiple stakeholder groups. As well as existing MNOs and new entrants, they should look for ways to attract:

• Infracos and neutral hosts

• Fixed telcos, ISPs and cablecos

• Hyperscalers and AI leaders

• Critical infrastructure operators and their suppliers

• Private equity and infrastructure funds

• Satellite companies

• Enterprises and consortia

• Various government-related bodies, agencies and functions

Note: many of the ideas in this post are believed to be original and unique, and unavailable from GenAI sources at the time of publication, based on various test prompts and queries.

The historic logic for MNO-centric spectrum releases

Over the last 20-30 years, most countries have had a consistent model for most spectrum releases: identify and clear out bands occupied by incumbent radio users, package spectrum into appropriate-size chunks, and then auction or award them to MNOs (mobile network operators) for 3G / 4G / 5G and soon 6G cellular networks.

There has been a consistent vision here - national (or in some places regional) MNOs build out coverage in the newly-awarded frequency ranges, upgrade their existing capacity or move up to a new G generation, and hopefully improve services for consumers and businesses. Sometimes there’s an aim to get a new entrant, with rules designed to smooth the path, such as MVNO arrangements with incumbents.

The underlying idea is not just enhancing connectivity, but also generating direct revenue for the national treasury coffers from auction/license fees, contribute to the economy during construction and upgrade, and then yield tax income from MNOs’ profits. Meanwhile consumer surplus increases from greater network capability, and hopefully there are additional GDP benefits from transforming business and public sector via greater use of mobile solutions. Some countries also benefit from having local equipment vendors and software suppliers for both domestic and export business.

Competition between MNOs has been viewed as positive, as each brings its own sources of capital and investment, and rivalry ensures continued upgrades and enough price-pressure and feature differentiation to continue the economic flywheel.

Typically this has meant shuffling around existing spectrum users - TV broadcasters, fixed-link microwave connections, satellite users, military communications and others have been squeezed out of existing bands. It’s been a one-way transfer from those historic radio applications to mobile.

(There has also been a parallel move toward repackaging spectrum for unlicensed use, primarily for technologies such as Wi-Fi, Bluetooth and LoRaWAN, as well as some proprietary options - and occasionally, unlicenced 3GPP cellular).

The spectrum landscape is changing rapidly

That world-view of spectrum is now changing. And while large-scale mobile networks and services remain hugely important, it is rapidly becoming clear that both 5G / 6G, and especially the legacy-model national MNOs, are not the only games in town, nor the ones which are growing fastest and show the greatest potential.

I’ve written extensively about the rise of localised private cellular networks for enterprises, in bands such as CBRS in the US (3.55-3.7GHz), Shared Access Licenses in the UK (3.8-4.2GHz, plus others), Campus Networks in Germany (3.7-3.8GHz) and about 40-50 other countries with a variety of spectrum ranges and rules. Some such as India and China are still debating the options.

I’ve also covered the different options for unlicensed or dynamic-shared 6GHz at length, both in my own newsletters and posts, and for various clients such as CRFS.

Both private 5G and Wi-Fi usage will continue to increase and create value for businesses, consumers and national economies. The same bands are also often used for FWA (fixed wireless access), especially where higher power is possible, for instance using the 6GHz AFC (automatic frequency coordination) model.

But I’m thinking beyond local or indoor-specific wireless networks. I’m wondering which other organisations might line up for larger-scale networks, with investments - and perhaps auction bids - comparable to the MNOs. The conventional operators have mostly failed to monetise their recent 5G investments and seem reticent about future expansion, especially with much-slowed growth in mobile data consumption. Enthusiam for a big investment in 6G is muted, especially in Europe.

It’s easy to imagine that alongside enterprise private wireless, public 5G and future 6G is where spectrum demand and growth lies, so that’s the obvious place to aim new spectrum. However, mobile data traffic growth has stalled, down to single-digit % in many places, and there’s no sign of an AI-led boom yet, let alone the original visions of AR / VR. Instead, network improvements in efficiency are growing faster than data. There could even be overcapacity risk, not a “spectrum crunch” that some suggest.

But there’s plenty of growing spectrum demand elsewhere - in-home connectivity fed by gigabit fibre, new satellite services, new utility and energy grids, vastly updated defence needs, aviation and maritime services, smart cities and regions and many more.

Many of these don’t just want commercial MNO services either - they want direct ownership and control, and perhaps the ability to use unusual wireless and radio systems rather than 3GPP or IEEE standardised ones. Look no further than drones to see the evidence of this. Yes, you can use 5G or Wi-Fi for control or video, but that may not be the best option. Satellite services are the same, irrespective of the “5G NTN” potential.

And for many applications, such as radar, or event audiovisual and wireless microphones, cellular isn’t an option at all. There’s a lot of discussion about AI variants of SDR (software-defined radio) too, as well as military-grade use of cellular, perhaps with special variants of Open RAN architecture.

In other words, spectrum demand is growing faster than supply. That ought to imply increased value, whether that’s directly from auctions or indirectly from deployment and usage.

There’s also no shortage of telecom and technology-sector funding around - look at the $trillions pouring into AI datacentres, semiconductor manufacturing and supporting infrastructure, or the 100s of $billions that have gone into FTTX and long-haul / subsea fibre infrastructure.

Yet the European Commission claims there’s an “investment gap” for telecoms, the new US OBBB Act aims to raise $80bn+ from spectrum auctions, and the UK Government hopes telecoms will drive economic growth through service innovation and industrial transformation.

How can spectrum regulators and governments capture more of the infrastructure cash which is sloshing around? And how can they align spectrum assignments with other government policy objectives, such as AI adoption and national security?

Fresh capital, fresh spectrum owners, fresh ideas?

Maybe the answer is for regulators to structure spectrum releases to appeal to new audiences, as well as MNOs?

Is it possible to construct auctions or other mechanisms to catch the eyes of international stakeholders - say Microsoft, Anthropic, Thales, NATO, Digital Bridge, Tesla or Apple?

It doesn’t just need to be multinational investors either. There’s also scope for national (or EU regional) peers, such as sovereign AI entities, fixed operators, national satcos, local consortia of utilities, defence agencies, municipalities or pension funds. Maybe Mistral in France, or SAP or BMW in Germany?

This isn’t completely unheard of in telecoms. There are plenty of unconventional asset owners, both for spectrum and fibre networks. Many FTTH assets are owned by energy companies (look at Denmark, for instance), municipal authorities (many countries), private equity (KKR in Italy, which bought TIM’s network), Internet companies (e.g. Google Fiber) and regional co-ops.

In mobile, there are conglomerates such as Reliance Industries that owns Jio, and cloud companies such as Rakuten which own spectrum and mobile networks. Malaysia has a government-run shared 5G network, while Vietnam’s Viettel is owned by the nation’s Ministry of Defence. Neutral hosts such as Dense Air have bid for spectrum in the past. TampNet operates offshore mobile networks, and Southern Linc is owned by a US utility. Germany has licensed 450MHz mobile spectrum to a group of utilities (450connect).

Satellite operators are increasingly blurring the lines between terrestrial and space spectrum usage too, with direct-to-device concepts and support from regulators such as the US SCS (supplementary coverage from space) and UK’s Ofcom.

There’s no reason why SpaceX, OneWeb/Eutelsat, Kuiper, SES or others might not want a terrestrial spectrum complement to their space services in future, rather than Vodafone, DT or BT seeking a space complement to terrestrial mobile. Maybe they’d bid more in an auction, or offer better aggregate coverage and service? Or perhaps they’d be willing to share with alternative and incumbent spectrum users, rather than demanding exclusivity?

Maybe a military that has been encouraged to give up some of its existing spectrum could choose to bid to buy it back again, perhaps with the idea of offering ultra-secure enterprise wireless alongside its own operational 5G/6G network? Maybe it would be done by a contractor like Lockheed Martin or Rheinmetall, rather than the defence department itself. In a way, that’s not that much different to the FirstNet / AT&T public safety model in the US.

In other words, new spectrum releases doesn’t have to be an MNO-only game. Policymakers should cast a wider net for direct and indirect value.

A wildcard: AI spectrum managers?

What about completely new concepts, such as an AI-powered Spectrum Management player, intending to offer wholesale-optimised “dark air” or spectrum-as-a-service, equivalent to “dark fibre” in the fixed world?

Imagine an AFC or SAS operator on steroids, with its own funding and spectrum assets, with the goal of maximising utility of the band for multiple purposes. Such a player could create a multi-technology, multi-tenant spectrum landscape, perhaps blending a digital twin with advanced spectrum-sensing and interference management.

It could carve up spectrum in three dimensions, with linear allocations for road, rail or power transmission, and aerial allocations for drone corridors. It could offer spectrum to enterprises for private networks, to MNOs for peaks and extra capacity - and manage allocations at different timespans from minutes/hours to years/decades.

That’s more akin to the sort of yield-management seen in the cloud world, or maybe property or aviation sectors. There’s lots of analogies to work with and learn from.

At the time of writing, AI vendor Perplexity was considering a $35bn acquisition of the Chrome browser from Google. Imagine if OpenAI decided to spend an equivalent sum, but on spectrum?

An AI-spectrum hybrid could also run its own private wireless at its datacentres and edge facilities for various purposes, or for uplink-heavy devices and applications on a wider basis. I’ve already started thinking about what “Private 5G in the AI/Cloud Sector” might look like.

None of this is easy

To be fair, the easiest path is to go with inertia. Just listen to the GSMA and various operators’ groups like CTIA or Connect Europe, or their tame consultants and political affiliates. Look at the ITU’s agenda for WRC-27 and its debates on “IMT” and simply assume that the old-school MNO way is the only response. National licenses, coverage obligations, maybe a few words about allowing MVNO access.

Otherwise, there’s a lot of hard work involved.

Regulators would need to work out different classes of spectrum license - national, local, corridor-type, 3D, wholesale vs retail, complex sharing of mobile with unlicensed or radar, and all sorts of other parameters. That’s certainly not easy, and probably not something that can be done in a single step, although existing efforts on dynamic/hybrid sharing are a good start.

There might need to be use-it-or-lease-it rules, or some other mechanisms to avoid spectrum “squatting” or underuse. They’d certainly need to codify and simplify methods for spectrum trading. Possibly there would have to be enforced standards on interoperability, such as cross-technology signalling.

Competition authorities find it easy to understand MNO vs. MNO discussions, either in terms of spectrum ownership or M&A transactions. How would they compare MNO vs. satellite operators? Or utility vs. military vs. cloudco vs. local council?

There are plenty of rules on foreign ownership and control to consider here as well, plus thorny issues around state aid, if government entities ended up competing with purely commercial services.

And then there’s the obligations that might come with spectrum licenses, such as coverage and build-out targets. These would have to be rethought, although in many cases that’s long overdue anyway. % population coverage is pretty irrelevant in an era where we expect farms, ports, offshore zones and mountains to get wireless signal, despite having a population of near-zero, or perhaps a few sheep.

Given that critical infrastructure may be involved, targets should look more at land/sea area coverage, resilience, security, backups, energy supply and other metrics. Yet another approach would look at specific sites to cover, from ports to railways stations and universities.

Outside of regulation, there might also be major hurdles in terms of financial metrics - how would analysts cope without a convenient ARPU measure? How could a wholesale-only play be valued?

Conclusion: time to widen the lens for spectrum

5G has proven that the mobile train runs on multiple tracks. I’ve said for some time that the mobile industry is becoming more heterogenous, with many new network owner types.

But I think the 6G version of the slide below will need to show multi-radio coexistence as well - not just public and private 5G, but also coexistence with radar, unlicensed, satellite networks and even new AI-generated waveforms for specific use-cases.

I think that policymakers need to think carefully about their assumptions, especially when it comes to investment in spectrum and wireless networks. The old model of MNO-focused auctions, that pre-suppose a conventional operator business model, is at or near its end.

The future involves many more spectrum stakeholders, model-neutral approaches to spectrum auctions and awards, and a much broader calculus for looking at value - including national security and resilient infrastructure, as well as basic mobile revenues and investments.

There should be openness to innovative consortia, and unique proposals for efficient use of spectrum in space, time, and application. And especially, there should be as much alignment (and regular updates) between spectrum policy and AI, infrastructure and security objectives.

As a side-benefit, diversity of demand will likely pull through diversity of supply, with new vendors needed not just for new forms of RAN and other systems such as satellites, but also a new segment of advanced spectrum management tools such as databases and sensors. There will be scope for new national champions.

It’s time for policymakers to look beyond the historical MNO model for new spectrum releases, and seek out new sources of value that align better with the AI era, rather than merely 5G/6G. We’ve had technology-neutrality. That should be extended to business-model neutrality in future.

______________________________________________________________________________________

Thanks for reading this newsletter. Please subscribe to my Substack (currently free) to receive new thought-leadership posts first and support my work.

#spectrum #5G #6G #policy #regulation #wifi #mobile #telecoms #spectrumsharing #satellite #innovation