Opinion: Infrastructure’s Biggest Risk Is the One Investors Cannot See

NEW YORK, March 31, 2026. For two decades, infrastructure has been one of global finance’s most reliable bets. Pension funds, insurers and sovereign wealth funds have poured capital into roads, utilities and energy networks, drawn by predictable cash flows and long-term contracts.

The sector is now dominated by a handful of platforms, including Macquarie Asset Management (MQG.AX), Brookfield Asset Management (BAM), and Blackstone (BX). Together, they oversee infrastructure portfolios spanning continents and supporting essential services.

From the outside, this ecosystem appears tightly controlled. Investors track performance through sophisticated financial systems and governance frameworks. Portfolio risks are modelled, stress-tested and monitored continuously. At the operating level, assets are maintained through mature systems designed to manage inspections, maintenance and lifecycle planning.

And yet, the industry has a blind spot. It is not financial. It is physical.

Modern infrastructure investors excel at measuring financial risk, but far less so at observing physical reality. They can model cash flows with precision yet, often lack a clear view of how the underlying assets are actually changing, until those changes begin to affect returns.

Infrastructure does not fail abruptly. It deteriorates gradually in pipes, in pavement, in spillways, often beyond the visibility of financial reporting. By the time the consequences appear in earnings, the underlying risk has already materialised.

The experience of Southern Water in the UK illustrates the point. The utility was fined £126mn in 2019 for serious operational failures, including illegal sewage discharges, according to the regulator Ofwat and Reuters. A further £90mn fine followed in 2021. In subsequent years, the company required substantial new equity, up to £1.2bn, to stabilise operations and fund environmental improvements. These were not sudden shocks but the financial expression of problems that had accumulated over time.

The same pattern is visible in the US. Even where assets do not fail physically, operational assumptions can break down. The Indiana Toll Road, leased in 2006 by a consortium including Macquarie, entered bankruptcy in 2014 after traffic forecasts and the asset’s lifespan proved overly optimistic, leaving roughly $5.8bn in debt to be restructured.

These cases are often treated as exceptions. They are not. They reveal a structural feature of infrastructure investing. 

Operators manage the assets. Engineers inspect them. Investors own the risk.

What is missing is a continuous translation layer between those worlds, one that turns raw engineering data into forward-looking insight at the portfolio level. Today, that translation remains largely manual, fragmented and episodic.

This gap is becoming harder to ignore. Infrastructure portfolios are expanding in scale and complexity, while climate volatility is introducing new stresses on physical assets. At the same time, investors are moving into more operationally sensitive sectors, from energy transition systems to data infrastructure.

In this environment, the ability to see physical risk early is no longer optional. It is fundamental. Infrastructure does not deteriorate in financial models; it deteriorates in the field, often quietly and unevenly. The signals are there, but they are dispersed across inspection reports, imagery and maintenance records, rarely forming a coherent picture at the investment level. By the time risk is formally reported, the opportunity for low-cost intervention has often passed.

A new generation of tools, increasingly built on artificial intelligence, is beginning to close that gap by analysing these fragmented signals at scale. The shift is not about better reporting but earlier understanding. Instead of documenting what has already gone wrong, these systems aim to identify where deterioration is accelerating, where assumptions no longer hold and where action is needed before risk becomes loss. For infrastructure investors, this marks a transition from periodic oversight to continuous awareness, and potentially from reactive capital deployment to informed, anticipatory control.

Because in infrastructure, the most expensive risks are the ones that remain invisible until they are not.

About the author

Saar Dickman is the CEO of Dynamic Infrastructure, an Engineering AI company focused on transforming how civil infrastructure is analyzed and managed. With experience working across government agencies, engineering teams and infrastructure investors, he focuses on bridging the gap between engineering data and decision-making, and on deploying AI technologies across the different layers of the infrastructure ecosystem, from field inspections to portfolio-level intelligence.

Media Contacts: 

Orly Ben-Eliyahu, COO, orly.b@diglobal.tech Phone: 2407318864