Infrastructure performance is increasingly shaped by forces that do not respect regulatory cycles or capex plans. Rainfall patterns are changing, groundwater behaves differently across catchments, and population pressure keeps rising. In that world, the question is not who had the worst headlines last year. The question is who is structurally less sensitive to stress, and who is one bad season away from persistent underperformance.
Vallorii’s work on asset health and resilience starts from a gap in the market. The most important variable is often the one we do not directly observe: the underlying condition of the assets and the system and how it reacts under environmental pressure. If you can estimate that hidden state credibly, you can do two things that matter commercially. You can compare operators on a like for like basis, and you can price resilience risk rather than treating it as an anecdote or an unknown.
That is what we did in UK wastewater.
Turning “unknown health” into a measurable signal
Our approach is deliberately pragmatic. We combine what external factors (population, rainfall, groundwater levels) do to a network with what the network does in response. On one side sit external conditions such as rainfall, groundwater, seasonality, and population effects. On the other side sit observed “failure signals”, such as storm overflows, alongside measurement intensity. From those inputs we infer an underlying resilience score using a probabilistic model, designed to cope with patchy and imperfect real-world data.
The output is a single, interpretable index: VARI (Vallorii Asset Resilience Index). A high score means assets behave as if they are relatively insensitive to external factors stress. A low score means performance drops off quickly as stress increases, i.e. the asset resilience is low.
This distinction matters because it reframes “asset health” away from a static label and towards a behaviour under pressure. That is closer to how investors experience risk and how operators experience operational pain.
The spread is real, and it is investable
What stands out in the initial results is not the ranking itself. It is the dispersion. Some networks appear meaningfully more resilient than others, even before you try to explain why.
Vallorii asset resilience index, (normalized)
- Anglian96.7%
- Severn Trent91.0%
- Southern89.4%
- Wessex85.6%
- Yorkshire65.4%
- South West64.5%
- Thames54.1%
Show the data as a table
| Index, 100 = high asset resilience | |
|---|---|
| Anglian | 96.7% |
| Severn Trent | 91.0% |
| Southern | 89.4% |
| Wessex | 85.6% |
| Yorkshire | 65.4% |
| South West | 64.5% |
| Thames | 54.1% |
100: High asset resilience. Assets are insensitive to environmental stress.
0: Low asset resilience. Assets are highly sensitive and fail quickly under environmental stress.
Remaining wastewater companies excluded due to absence of data
A simple way to read the chart is this: it separates “systems that bend” from “systems that break”. Anglian and Severn Trent sit at the higher-resilience end in this first dataset, with Southern and Wessex also scoring strongly, while Thames, South West, and Yorkshire sit lower.
For investors, dispersion is where pricing power comes from. If resilience were uniform, there would be little to learn. When resilience differs, you can start to ask sharper questions about operating risk, future penalty exposure, capex efficiency, and the credibility of performance improvement plans or the likely terminal value of your investment. VARI is a way to make those questions quantitative, data driven, and therefore repeatable.
One nuance is worth highlighting because it comes up often in boardrooms. Resilience is not the same as the headline penalty narrative. A company can show high resilience even while facing ODI penalties. The point is not to argue with penalties. The point is that penalties capture outcomes, while VARI focuses on external factor sensitivity and underlying system behaviour. Those two can diverge, and when they do, there is insight.
The outliers are the point, not the noise
The second visual makes the core idea tangible: resilience is related to asset age, but it is not explained by age.
Average sewer age vs Health
Show the data as a table
| Average sewer age (years) | Health | |
|---|---|---|
| Anglian | 73 | 0.97 |
| SevernTrent | 66 | 0.91 |
| Southern | 67 | 0.89 |
| Wessex | 62 | 0.86 |
| Yorkshire | 72 | 0.65 |
| SouthWest | 55 | 0.64 |
| Thames | 85 | 0.54 |
Health is shown as a percentage (0–100%).
There is an intuitive relationship. Older systems often embed legacy design standards and constraints that can show up as higher sensitivity to stress. But the interesting cases are the deviations. In the slides, Thames’ weaker resilience is consistent with older assets, while South West looks weak despite having comparatively younger sewers.
That is not a curiosity. It is an investable and operationally actionable signal. If a network underperforms relative to its asset vintage, age is not your explanation. You are pushed toward more useful drivers such as capacity relative to environment, network complexity, local exposure, and operational performance.
For utilities, this is where VARI becomes more than a league table. It becomes a prioritisation tool. If you can separate “old but coping” from “young but fragile”, you can allocate attention and capital toward what actually reduces sensitivity, not what merely increases spend.
From measurement to pricing
Estimating unknown asset health is only half the value. The other half is what happens once the estimate exists.
Once resilience is measurable, it can be treated like any other risk factor. It can be compared across operators, tracked over time, stress-tested against environmental scenarios, and mapped into valuation. This is where Vallorii’s broader platform connects. VARI provides an interpretable resilience signal. Vallorii’s work on pricing risk, including VAPRI, is designed to translate such signals into implications for future returns and capital allocation.
In practical terms, that means moving the conversation away from vague statements like “we need to invest in resilience” and toward decisions that withstand scrutiny, such as “this intervention reduces sensitivity by X, which reduces downside risk by Y, which supports Z in value”.
Why this matters now
Wastewater is simply a clear example because the environmental link is strong and the consequences are visible. The same logic applies across infrastructure: when volatility rises, sensitivity becomes destiny.
The same logic applies across infrastructure: when volatility rises, sensitivity becomes destiny.
This first analysis is explicitly presented as preliminary and constrained by data availability. Some companies could not be included due to absence of usable data. That limitation is also a signal. Better data makes better markets. And better markets reward the operators that actually build resilience, rather than the ones that are best at explaining it.
If you are an investor, the question is straightforward. Which assets are quietly robust, and which are exposed to stress in ways current valuations do not reflect? If you are a utility, the question is equally commercial. Where does the next pound of spend reduce sensitivity the most, and how do you prove it in a way that holds up to scrutiny?
That is the work VARI is designed to support: estimating the unknown, and making it priceable.
More research

UK regulation in motion: December-January updates reshaping utility finance, and what to price next
UK utility regulation has shifted sharply from policy headlines to a governance-and-finance reset. Across water, energy, telecoms and transport, regulators are tightening incentives, hardening enforcement, and demanding credible delivery as a condition of “financeability.” Ofwat’s blind-year reconciliation makes performance real cash, while RIIO3 pairs major investment with tougher bill politics and consumer protection. Payments and billing are emerging as the next regulatory frontier.

Cassandra Etter-Wenzel
7 min read

The Broken Link between AI Growth and Electricity Demand
The world will use far more AI. That does not mean it will use proportionally more electricity. As AI systems become more efficient, the relationship between adoption and power demand weakens. This creates a forecasting problem with real consequences for grid planning, investment timing, and policy—where flexibility matters more than precision.

Jorge Cárdenas
8 min read
