Heat as a Failure Mode: The Derated Grid
Iron does not fear the hammer; it fears the forge. The blow it can answer, but the heat unmakes it from within, softening the very strength that was meant to bear the blow — and the smith who reckons only the hammer has misjudged which of his tools does the breaking.
Heat Is an Operating Condition, Not an Event
Most climate risk is filed under disasters — the flood, the storm, the fire, discrete events an asset either survives or does not. Heat belongs in a different category, and mispricing it starts with putting it in the wrong one. Heat is not an event that happens to infrastructure. It is a condition infrastructure operates in, and its effect is not to destroy the asset but to quietly reduce what the asset can deliver.
Every piece of thermal-limited equipment — a turbine, a transformer, a transmission line, a solar panel — carries a rating set against an assumed temperature. As the ambient rises, the physics that rating depends on degrades: air gets thinner, cooling gets weaker, resistance climbs, insulation ages faster. The asset does not fail. It derates — it goes on running, and delivers less. The whole grid was designed to a temperature envelope that the climate is now leaving, which is why the industry itself has started describing heat as moving “from a tail risk to a design baseline.”
Heat doesn’t break infrastructure. It shrinks it — and it shrinks every thermal-limited asset at the same time, in the same heat, exactly when demand peaks.
That single sentence contains the whole risk. The loss is a capacity derating rather than a discrete failure; it is correlated across assets because they all feel the same weather; and it is anti-correlated with supply margin, because the heat that shrinks supply is the heat that spikes demand. A risk that is quiet, correlated, and worst at the worst moment is precisely the kind markets price badly.
The Failure Mode Is Derating, Not Breaking
Walk through the fleet and the same pattern repeats: heat takes a slice of capacity off the top of nearly everything that makes or moves power. None of it is dramatic. All of it is simultaneous.
| Asset | How heat takes capacity | Rough loss at extreme heat |
|---|---|---|
| Gas turbine (simple cycle) | Warmer, thinner intake air means less mass flow through the machine | ~10% above roughly 32°C / 90°F |
| Combined-cycle (CCGT) | Same intake effect plus reduced condenser performance | ~5% |
| Thermal & nuclear (steam) | Warmer cooling water lowers efficiency; thermal-discharge limits force curtailment | ~0.3%/°C, plus curtailment risk |
| Transmission lines | Weaker convective cooling cuts the safe current rating; conductors sag, forcing clearance de-rating | Line rating cut, often ~5–15% |
| Transformers | Windings run hotter; output is derated to protect insulation, whose life halves per 10°C over rating | ~0.4%/°C above a 30°C average |
| Solar PV | Cell efficiency falls with temperature; inverters derate in the heat | ~0.3–0.5%/°C above 25°C |
The important word in that chart is “everything.” This is not a nuclear problem, or a gas problem, or a renewables problem — a misconception the trade press keeps correcting. Thermal generation curtails on cooling limits; gas turbines lose intake density; solar loses cell efficiency; the wires that carry all of it lose ampacity; the transformers that step it all lose rated output. The derating is a property of the physics, not of any one technology, so diversifying across technologies does not diversify it away.
The Correlation Is the Danger
Here is what separates heat from every other climate hazard, and it is the crux of the whole piece. A storm reduces supply but does not raise demand. A flood suppresses both. Heat pushes supply down and demand up at the same time, from the same cause, and holds the pressure for days. The air-conditioning load that spikes the demand curve is created by the identical weather that is derating the supply behind it. Margin is squeezed from both ends at once.
This is why heatwaves, not storms, produce the tightest grid conditions and the ugliest price spikes. And the correlation is not only physical but financial: the loss is multiplicative rather than additive, because the derated megawatt is worth most in the exact hour it goes missing. The same logic reaches the repair crew — the people restoring a heat-stressed network are working in the heat that caused the failure, which stretches restoration and compounds the outage.
Nameplate Is a Fair-Weather Number
Follow the derating to its financial conclusion and it changes what a megawatt of capacity is worth. The nameplate rating — the number on the asset, in the model, in the resource-adequacy study — is the capacity available on a mild day. But the system is tested on the hot day, when that same asset delivers less. The capacity you can actually rely on is the hot-day, derated capacity, not the nameplate. Every planning process that credits nameplate is overstating firm supply by exactly the derating, and doing so most where it matters most.
Firm capacity is the capacity that shows up on the worst day. Heat is what defines the worst day — so heat, not the nameplate, sets the number that counts.
This is why capacity-accreditation reform — how much credit a resource receives toward reliability — is quietly one of the most consequential fights in power markets. Move from nameplate toward performance-based accreditation and the value of a resource is repriced by how well it holds up in heat. Assets that keep delivering in extreme heat gain; assets whose ratings evaporate on the hot afternoon lose. The reform does not create the risk — it reveals a mispricing that was always there.
Two plants with identical nameplates are not identical assets if one holds capacity at 45°C and the other sheds 10%. In a market that increasingly pays for firm, hot-day delivery, that gap is a valuation gap the nameplate hides. The discipline is to underwrite generation and network assets on their derated summer-peak capacity, treat the nameplate as marketing, and price the difference as either a discount (for heat-fragile assets) or a premium (for heat-robust ones).
Designed Against a Climate That Moved
The reason this is a growing exposure rather than a stable one is that the design envelope was fixed and the climate was not. Ratings, clearances, cooling assumptions and thermal limits were all set against a historical temperature distribution — the hottest days on record at the time of design. As that distribution shifts warmer, the days that breach the envelope, once rare tail events the system could ride out, become a routine summer condition. The margin that used to absorb the occasional hot day is being spent as a regular operating cost.
The market is starting to register this in the one place that prices risk directly: insurance. Coverage for heat- and drought-exposed infrastructure is being repriced or withdrawn across the most exposed regions, and less than 15% of losses from these “climatological” events are insured — a protection gap that leaves the derating and its consequences sitting on asset owners’ balance sheets. Add stretched replacement timelines — distribution-transformer lead times ran to 80–120 weeks into 2026 — and a heat-driven failure is not only more likely but slower and costlier to put right.
Positioning: Price the Derated Megawatt
The mispricing runs in one direction: the market tends to credit nameplate and underweight the correlated, hot-day derating, which means heat-fragile capacity is systematically overvalued and heat-robust capacity systematically cheap. Position against that gap.
Buy the capacity that shows up in the heat; sell the capacity that only shows up in the brochure. And own the tools that buy the margin back.
Three places to stand. First, heat-robust firm capacity — storage, demand response, and generation that holds output in extreme heat — which is worth a premium a nameplate-based market underpays. Second, the margin-recovery toolkit: dynamic line rating, advanced conductors, transformer-cooling upgrades, and the grid-enhancing technologies that reclaim derated capacity for a fraction of the cost of new build. Third, the mispricing itself — discount the heat-fragile merchant asset whose “firm” summer capacity is a fair-weather number, and pay up for the asset whose worst-day delivery is genuinely firm.
Reading It Through the Frameworks
Where is the physical risk mispriced? Squarely here. Markets are fluent in financial risk and clumsy with physical risk, and heat derating is physical risk in its purest form — a quiet, weather-driven capacity loss that never appears as a discrete event on a loss run. Because it is correlated across the whole thermal-limited fleet and worst at peak, it is both larger and harder to diversify than a nameplate-based model implies. The gap between the modelled capacity and the deliverable capacity on the hot day is the mispricing.
Structural moat or temporary bottleneck? Neither, exactly — it is a permanent, worsening operating condition, which makes the toolkit that manages it a structural growth market rather than a one-off fix. The discipline is to separate the asset that is cheap because it is heat-fragile (a value trap dressed as a bargain) from the asset that is cheap because the market has not yet paid for its heat-robustness (a genuine mispricing), and to treat the margin-recovery vendors as sellers of a capability the grid will need every summer from here on.
Heat is not a disaster infrastructure survives; it is an operating condition infrastructure runs inside, and its signature is not failure but derating — a slice of capacity taken off the top of nearly every thermal-limited asset at once. The reason it matters more than its quietness suggests is the correlation: the same heat that shrinks supply spikes the cooling demand, so the reserve margin collapses from both ends in the same hours, and the missing megawatt goes missing exactly when it is worth most.
Nameplate is a fair-weather number. The capacity that counts is the one that shows up on the hot day, and a market that credits the rating rather than the hot-day delivery is systematically overpaying for heat-fragile capacity and underpaying for heat-robust. Underwrite the derated megawatt, buy the resources and the tools that hold or reclaim capacity in the heat, and treat the cheap heat-fragile asset as the value trap it is. The grid was built for the hottest day anyone had known — and the climate keeps a hotter one in reserve.
They built for the hottest day they had known, and called it strength; but the years kept a hotter one in reserve, and strength measured against a memory is only weakness that has not yet been asked the question.
Leave a Reply