Can rail connectivity survive a demand curve it was never designed to handle?

The rail industry has been having the same connectivity conversation for a decade. How do we close coverage gaps? How do we get faster download speeds? How do we improve the passenger experience? At this year’s RIA Innovation Conference, Mark Barrett of Blu Wireless and Tim Wood of Amey Consulting reframed the discussion. Not how to improve connectivity, but whether the current modelling behind it can endure what tomorrow’s railway demands.

The wrong problem, restated

Connectivity still matters for passengers, but it is no longer solely a passenger problem. The railway now depends on it to operate safely and in real time. CCTV systems now run multiple 4K cameras per carriage while remote condition monitoring, predictive maintenance sensors and onboard analytics generate continuous data flows. The operational model has moved from periodic to real-time.

This isn’t exclusively about capacity; it’s about whether the railway can operate safely, respond to incidents in real time and maintain visibility across the network.

A demand’s curve the current model cannot absorb

Ofcom estimated in 2018 that the railway needs approximately 2 to 2.5 Gbps per train, based on roughly a thousand passengers each requiring around 2 Mbps. Most trains ran on 4G connections delivering 30-40 Mbps. The gap was already significant then and it has widened considerably since..

When operational requirements are added, the picture becomes even more demanding.  Security cameras can push uplink demand beyond 1.5 Gbps per train, while total data consumption on a busy service has been modelled at as much as 68 TB per train, per month.

Low Earth Orbit (LEO) satellite can help improve coverage, particularly on rural and low-density routes, but estimated LEO bandwidth is only around 200-300 Mbps per train, which only scratches the surface of the required 2 to 2.5 Gbps. Shared bandwidth, line-of-sight dependency in tunnels and throughput limitations mean that satellite alone cannot meet the full capacity requirement. It’s certainly complementary, but far from a comprehensive solution.

What happens when capacity falls short

If connectivity cannot support real-time CCTV and operational systems, the impact is reduced visibility, slower incident response and increased operational risk. When those critical decisions get deferred, they compound risk.

Where the cost model breaks

Today, train operators pay mobile network providers on a per-gigabyte basis for connectivity. Across the UK rail network, that adds up to around £90 million each year for a service that still falls short of acceptable.

Now factor in the demand curve. As Mark Barrett put it at RIA, “Today’s solutions solve yesterday’s problem.” If the industry were to move to gigabit-level performance using the same per-gigabyte model, costs could rise towards £1 billion per year. At that point, the model becomes unsustainable.

In a publicly supported railway, those costs don’t sit with a single operator. They feed into the long-term funding model and, ultimately, fall to the taxpayer. As Tim Wood said at the same event, “Today we want to talk about protecting taxpayers’ money.”

The result is growing pressure on funding, weaker business cases and delayed decision making. The rail industry’s digital ambitions cannot be delivered through a cost model that becomes more expensive with every gigabyte consumed.

A different architecture is required

The approach developed by Blu Wireless and Amey Consulting deploys millimetre wave (mmWave) trackside infrastructure as a high-capacity backbone, with radios placed at regular intervals along the route and connected to a resilient fibre backhaul. Operating in licence-exempt spectrum between 57 and 71 GHz, the system delivers consistent, deterministic throughput at multi-gigabit speeds. It empowers existing mobile network operators rather than replacing them, absorbing the heavy-capacity demand through the trackside backbone while MNOs remain part of the connectivity mix.

Because mmWave operates in licence-exempt spectrum, it carries no per-gigabyte data charges. Capacity increases do not drive cost escalation, and connectivity behaves as a fixed asset, not a variable cost.

The ongoing challenge has not been identifying the right technology, but finding a model that can be approved, funded and defended over time. For decision-makers working within existing rail funding structures, an infrastructure-led model offers predictable spend, reduced financial exposure and alignment with how rail investment is typically approved. It reduces both the financial and reputational risk for those making the case.

Proof in a live environment

South Western Railway (SWR) is now the first railway in the world to operate all three connectivity methodologies – 4G/5G, mmWave and LEO satellite – across a single network. Rather than backing a single technology, SWR has adopted a hybrid approach, using mmWave on the busiest, most passenger-heavy sections and other methodologies where bandwidth requirements are lower.

The early results speak for themselves. Since deployment, SWR has recorded a 10 per cent increase in overall customer satisfaction and a 65 per cent increase in onboard Wi-Fi usage. Across the network, SWR generates c300TB of passenger data per month. Of that, 100TB flows through the mmWave backbone alone across just 70 km of the 998 km network.

The question the industry must answer

The rail sector does not lack ambition around digital operations, and it is well-versed in the business cases for real-time monitoring, onboard analytics and predictive maintenance. However, the industry lacks a connectivity architecture that can deliver the capacity required without creating open-ended cost exposure.

Every year the industry delays this decision, the cost curve steepens and the gap between what the railway needs and what it can deliver continues to widen. The model exists and the proof is already in operation. The question is whether the industry acts before the numbers force its hand, or after.