Every piece of infrastructure I have ever worked on carries an invisible assumption. Somewhere in the design calculations, buried in a drainage coefficient, a thermal expansion tolerance or a load combination factor, is an implicit claim about the climate that the asset will operate in for the next thirty to fifty years. That claim is almost always derived from historical weather data, and it is increasingly wrong.
I am a structural engineer working on bridges, transit facilities and utility structures. I also study how climate risk intersects with infrastructure finance. There is a growing information asymmetry between what engineers know about infrastructure performance under changing climate conditions and what financial models assume about future costs and revenues.
This asymmetry leads to mispricing. But AI-driven climate analytics are starting to close the gap, translating forward-looking physical risk into inputs that affect cash flow projections, insurance costs and required returns. Investors who pair these analytics with asset-level engineering assessments can identify where value is at risk before the broader market adjusts and target resilience investments to preserve returns that would otherwise erode. The tools to do this exist today and are improving rapidly, and most infrastructure capital still does not use them.
What engineers see that markets do not
The engineering profession designs for return periods. A 100-year flood. A 50-year wind event. These are probabilistic benchmarks derived from historical frequency distributions. According to the National Oceanic and Atmospheric Administration, the heaviest one percent of rainfall events have intensified in most US regions since 1958. The Fifth National Climate Assessment identifies increasing risks to transportation, energy and water infrastructure from extreme precipitation, wildfires, sea-level rise and heat stress. The historical benchmarks that underpin both engineering design and financial underwriting are shifting.
When investors think about climate risk to infrastructure, they typically picture catastrophic events: a hurricane, a wildfire, a flood. Those matter, but the more pervasive financial exposure is accelerated degradation, the slowly compounding erosion of asset performance that shows up not in a single insurance claim but in rising maintenance budgets, shortened replacement cycles and declining service reliability.
Drainage systems designed for historical storm intensities are increasingly stressed by higher rainfall rates. Steel expansion joints cycle through thermal ranges more frequently as extreme heat days increase. Concrete elements in coastal and riverine environments face exposure conditions that challenge their original durability assumptions. None of these is a headline failure, but all of them shift capital expenditure timelines forward, often by years.
Financial models typically treat maintenance and capital replacement as stable, predictable cost lines. But engineers are already seeing those lines move. Research published through the National Bureau of Economic Research on sea-level rise and municipal bond yields and on heat stress pricing confirms that capital markets have begun to price some of this exposure, particularly in long-duration municipal securities. But the pricing remains uneven and largely limited to observable event risk. The slow-moving degradation embedded in maintenance logs and inspection reports has not yet reached the investment committee. That is where the asymmetry lives.
The tools changing the conversation
A new generation of climate analytics platforms is making it possible to model forward-looking physical risk at the asset level, something that was not commercially available even five years ago. Firms such as Jupiter Intelligence, ClimateAI and Moody’s Climate Risk Solutions now offer probabilistic models that project flood, wildfire, heat and wind risk under multiple emissions pathways, with resolution down to individual assets and time horizons extending through 2100. Jupiter’s platform is used by roughly 20 percent of the world’s largest companies and half of the largest US lenders. Moody’s integrates physical risk scoring into its existing credit analytics, making climate data accessible alongside conventional financial metrics.
What these platforms share is a shift from backward-looking historical averages to scenario-based probability distributions, enabling structured sensitivity analysis across disruption frequency, maintenance cost trajectories, insurance pricing and revenue stability. At the portfolio level, they allow managers to simulate correlated stress events across geographies, testing whether diversification assumptions hold under simultaneous extreme weather.
The tools are also evolving. Jupiter’s new Adaptation Hub quantifies avoided losses and return on investment for specific resilience measures. Earlier platforms answered the question, “Where is risk increasing?” The current generation is beginning to answer, “What should we do about it, and does it pencil out?”
But these tools have a limitation. Most model hazard exposure: the probability that a climate event will affect a location. They do not yet fully model asset vulnerability: how a specific structure, with its design, materials and maintenance history, will respond. Two bridges in the same floodplain can have very different vulnerabilities depending on their foundation type, drainage capacity, age and protective coatings. Closing that gap by connecting climate projections to asset-specific engineering condition data is the frontier, and it is where the most decision-relevant insights for infrastructure investors will come from.
Practical steps for investors
Upgrade due diligence with forward-looking climate screens. Any infrastructure investment with a holding period beyond ten years should be evaluated against forward-looking climate projections, not just historical loss data. If a thirty-year water utility concession is underwritten with a historical disruption probability of one percent annually, but forward modeling projects that probability rising to 2.5 percent by mid-century, the required return adjustment can be on the order of tens of basis points. That is the difference between a comfortably financeable asset and a marginally viable one. Platforms like Jupiter, ClimateAI and Moody’s make this analysis accessible without requiring in-house climate science expertise.
Evaluate resilience investments as return-preserving capital, not as a sustainability cost center. Targeted upgrades such as upsizing drainage infrastructure, installing corrosion-resistant reinforcement or elevating critical systems above revised flood elevations can measurably reduce disruption probability and extend useful asset life. When quantified against forward-looking risk scenarios, these upgrades reflect defensible capital preservation: resilience investments that maintain the internal rate of return by avoiding the revenue losses and accelerated capital expenditures that a changing climate would otherwise impose. Engineers routinely perform this cost-benefit analysis when evaluating rehabilitation versus replacement decisions. The missing step is connecting that engineering analysis to the financial models that drive investment decisions.
Look for the information asymmetry as a source of alpha. The assets most likely to be mispriced are those where the gap between engineering observations and financial assumptions is widest: aging infrastructure in regions experiencing measurable climate shifts, where maintenance costs are rising but market valuations still reflect historical performance. This is an extension of value-investing logic that seeks assets where the market has not yet incorporated available information. Emerging structures like resilience bonds, which link risk-reduction investments to insurance and capital markets, signal the direction: Climate resilience is moving from a reporting category into a priced attribute of infrastructure assets.
Leading the repricing
The investors who will navigate this repricing most effectively are not those waiting for better disclosure frameworks or more dramatic headline events. They are the ones already integrating forward-looking analytics into underwriting, commissioning engineering assessments that connect climate projections to asset-specific vulnerability, and evaluating resilience as a return-preserving investment rather than a compliance cost.
Infrastructure will be repriced. Not through a single shock, but through the steady accumulation of better information. The question for impact investors is whether they will lead that repricing and capture the value that comes from seeing the risk clearly.
Prateek Srivastava is a structural engineer at Stantec. He co-leads the ASCE Infrastructure 2050 Bridge Group and serves on the national ASCE SEI Advisory Council.
Guest posts on ImpactAlpha represent the opinions of their authors and do not necessarily reflect the views of ImpactAlpha.
