The Economics of Infrastructure, Not Device Sales

When a new technical system appears, it is almost automatically read through the familiar lens of product economics: what the device costs, how many units can be sold, what margin remains on the hardware, and how fast sales volume can grow. For LSMAD, that framing is too narrow from the start.

Physical devices exist within this architecture, of course. Without them there is no observation of the environment, no formation of telemetry, no local analysis, no fixation of events, no interaction with the wider operational system. But the device here is not the final economic product — it is a physical element of a larger construction. The core economic unit is a long-lived infrastructure of observation and operational coverage.

It is not built for a single insurance year or a single sales cycle. It is deployed gradually, serviced, upgraded, and it operates for years. It includes hardware, software, connectivity, service teams, operator infrastructure, specialist training, identification and maintenance of technical elements, and the mechanisms that form a structured history of observation.

The first question, therefore, is not "how many devices can be sold?" It is a different one: what infrastructure is rational to build, how many assets can it cover, and what recurring economic consequences will it interact with over its operating lifetime?

The distinction looks purely terminological until the actual figures are examined.

In England, fire and rescue services recorded around 558,000 incidents for the period ending March 2020, around 577,000 in 2022, and more than 600,000 in 2024. The number of fires in those same reference periods was approximately 154,000, 153,000, and 139,000. In Poland, the State Fire Service registered roughly 583,000 interventions in 2020, more than 608,000 in 2022, and more than 550,000 in 2024. In Singapore, the number of fires stayed in the range of roughly 1,800–2,000 events per year. In Germany, where national statistics are considerably more fragmented, the number of rescue operations alone, according to data available for 2022, stood at around 2.8 million.

These figures say nothing about how many events LSMAD could actually prevent. That is precisely their value: they show a more fundamental fact — the risk environment exists continuously and reproduces itself year after year.

In the United Kingdom, property insurance payouts in 2024 totalled around £5.7 billion. In Germany, preliminary data for residential building insurance for the same year show approximately €9.7 billion in claims paid. In Poland, total insurance payouts across the market grew from roughly PLN 39.6 billion in 2020 to PLN 50.3 billion in 2024.

These figures belong to different insurance classes and cannot be mechanically matched against fire counts. UK property claims are broader than fire losses; the Polish figure covers the entire insurance market; the German figure applies to one specific building-insurance segment. It would be wrong to claim that LSMAD can "save one percent" of any of these money flows.

For the economic argument of this article, that claim is not required.

One percent of £5.7 billion is £57 million. One percent of €9.7 billion is €97 million. One percent of PLN 50.3 billion is roughly PLN 503 million. This is not a forecast of potential savings and not a promise of efficiency — it is simply a way of seeing the scale of already-existing annual flows.

New technology is usually justified by a forecast: it should cut losses by 10%, 20%, or 30%. Before years of operating statistics exist, such figures remain assumptions. For infrastructure, the more useful question runs the other way: how small can the real effect be before the cost of building and operating the infrastructure already becomes rational?

This is probably one of the most important economic questions surrounding LSMAD.

If the estimated cost of building national infrastructure coverage is only a small fraction of the annual burden the infrastructure operates inside, the required efficiency threshold may not sit in the tens of percent. It may sit at single-digit percentages, or fractions of a percentage point.

The key figure in that case is no longer the maximum potential saving one could present to an investor, but the minimum effect sufficient to justify the infrastructure's existence. This shift in reference point matters — all the more because LSMAD interacts with more than one line of cost.

There is direct physical damage from events. There are insurance payouts. There is the operational load on emergency services. There are false and unnecessary call-outs, which still consume equipment, personnel, and time. There is asset restoration, cause investigation, expert procedures, claims settlement, subrogation, recovery actions, and fraud investigation.

These categories cannot simply be added together — they overlap, and doing so would quickly produce double counting. But their sheer number shows why the economic logic of such a system cannot be described by one damage figure or by the margin on one piece of equipment.

There is a further layer that conventional statistics barely see: LSMAD forms a structured history of the environment's state before the incident itself occurs.

After a fire, an accident, or another event, the circumstances usually have to be reconstructed from whatever remains available: equipment logs, witness statements, service documents, individual sensors, photographs, video, and expert findings. If a structured telemetric history already existed beforehand, the starting position is different — there is a sequence of observations before the event, a change in the environment, a temporal structure, and a provenance of the data.

This does not mean liability is established automatically, and it does not remove the need for professional expert assessment. But a dispute or an investigation no longer starts from a largely empty field. The economic effect of this shift cannot yet be honestly expressed as a specific percentage. It may show up as a smaller zone of uncertainty, less expert work required, a more precise reconstruction of the sequence of events, faster settlement of part of the claims, and less room for conflicting reconstructions. This effect sits in a completely different plane than the value of burned property, but excluding it from the overall infrastructure assessment would be as strange as valuing a telecom network purely by the cost of its antennas.

There is one more factor capable of changing the denominator of the calculation itself.

A conventional built environment is usually assembled from many separate technical systems. Fire safety has one set of equipment. Security functions have another. Monitoring of individual environmental parameters has a third. Structural condition monitoring has a fourth. Each system may have its own sensors, controllers, cabling, software, maintenance, and lifecycle.

LSMAD applies a different principle: a single structured observation infrastructure can support several different applied functions. This does not mean it automatically replaces existing specialised systems or systems mandated by regulation, but the economic weight of the principle is no smaller for that. The capital base does not need to be rebuilt separately for every single function — one physical observation layer can participate in several applied circuits at once, while its maintenance and lifecycle remain shared.

The result is an infrastructure multiplier: one capital base, several applications, one maintained infrastructure, one long operating life. This is no longer a characteristic of a "smart device" — it is a property of infrastructure.

The same logic changes how scale itself is read.

In the working operator model, a single local operating cell is treated as capable of serving on the order of 30,000 connected households. This is not a hard technical ceiling but a planning reference point, and it illustrates the underlying economic logic well: the infrastructure unit is not built around one house or one device — it serves a large, standing base of assets.

This changes how capital cost is read. The investment applies not to an individual device but to an operating structure able to work with tens of thousands of connections over many years. As coverage grows, new local units are added as elements of national infrastructure, not as independent commercial products.

The subscription model in LSMAD should also be read more broadly than ordinary SaaS logic.

Here, a subscription is not only a way to generate recurring revenue. It ties financing to the ongoing existence of the infrastructure itself. After hardware is installed, software support, connectivity, diagnostics, service, operator work, updates, specialist training, and maintenance all continue to be required.

A connected base of assets can fund these processes regardless of how much new hardware was sold in any given year. The economic motive shifts, over time, from constantly growing unit sales to sustaining the quality and scale of coverage — for an infrastructure model, this shift matters.

It also creates room for a different distribution of cost. If safety exists only as an individual device purchase, access depends directly on the owner's purchasing power. An infrastructure system allows insurance mechanisms, municipal programmes, social housing, programmes for elderly residents, and other groups to participate, with costs distributed differently. Quantifying that direction will require separate statistical work of its own, but the possibility exists precisely because of the infrastructure model, not the product model.

Comparing the United Kingdom, Singapore, Germany, and Poland in this context is not meant as a ranking of countries.

These are four substantially different environments: compact, high-density Singapore; the United Kingdom with mature public statistics and a large insurance market; Poland with a strong, consistent series of fire and insurance data; Germany with a more decentralised structure and a more complex statistical picture. The differences here are more useful than the similarities. National infrastructure has to make economic sense not only in one perfectly chosen country — it has to be examined across different conditions of population density, asset structure, organisation of emergency services, labour cost, and insurance environment.

The data gathered so far does not prove LSMAD's economic effectiveness in any of these jurisdictions. It shows something else: in all four, there is a recurring environment of events, operational load, and monetary consequences — that is, the actual subject matter for an infrastructure-level economic analysis exists.

One could artificially produce a headline ROI figure: add up different categories of damage, insurance payouts, response costs, asset restoration, apply an assumed reduction percentage, and present an impressive total. That calculation would not survive serious scrutiny.

For some countries there are monetary figures with no matching event count. For others there are events with no confirmed cost per case. Some categories overlap. For false call-outs there is still no reliable, comparable cost per response. The absence of one large headline figure is therefore not a gap — it is a methodological constraint: the real capital and operating costs need to be established first, strictly comparable economic categories need to be identified, double counting needs to be excluded, and only then can a full economic-return model be built.

But a question is already visible that is hard to ignore: if the annual money flows tied to risk are measured in billions, and building the infrastructure requires only a small fraction of that scale, how large does the real effect actually need to be? If the answer, within individual, strictly comparable categories, turns out to be less than one percent, the conversation stops being a conversation about selling a new technical product — it becomes a conversation about infrastructure policy, insurance economics, risk management, and the rationality of long-term investment.

By this point, LSMAD is no longer an abstract idea for a single device. The project is developed at the level of a Master Concept, covering the technical, operational, service, operator, and economic architectures. This does not mean the national system has already been built, or that its actual effectiveness has been confirmed by years of operating statistics. But the architecture available for professional evaluation is already substantially further along than an initial idea.

The next question, therefore, is a practical one — not "can such a system be built?" and not "how many devices can be sold?" — but how much it is rational to spend on building infrastructure capable of operating for years inside a recurring risk environment, serving a large standing base of assets, supporting several applied functions, and forming an evidentiary context before an event ever occurs.

And if the minimum required effect of such infrastructure genuinely turns out to be very small relative to the annual economic flows involved, the next question follows: why are models of this kind still so rarely examined from precisely this point of view?