Integrating Adaptation and Mitigation

Executive Summary

The standard question — how should climate finance be split between cutting emissions and coping with the consequences? — assumes the two are alternatives. In economies building both capital stocks at once, they are largely inputs to each other.

The accounts show how the conventional framing has resolved in practice. Of the US$1.9 trillion of climate finance the Climate Policy Initiative tracked in 2023, US$1,780 billion went to mitigation and US$65 billion to adaptation. The category that captures integration — dual-benefit finance, pursuing both objectives at once — reached US$58 billion, about three per cent of the total. On the other side of the ledger, developing countries' adaptation needs are put at US$310–365 billion a year by 2035 against international public flows of US$26 billion.

That is not simply an allocation failure. It is what happens when a financial system asks each investment which of two boxes it belongs in. An asset that does both fits neither box comfortably, and the smallest line in the accounts is the one describing the assets this report argues are the most valuable.

The report proceeds as follows. Section 1 sets out the two-budget-line framing and what the finance data show. Section 2 establishes that adaptation is electricity-intensive, and Section 3 that clean energy is climate-exposed. Section 4 organises the interaction into three regimes — joint, enabling and rival — with a screening test for each. Section 5 explains why the integration is cheapest now, Section 6 examines three cases, and Sections 7 and 8 give recommendations and conclusions.

1. Two Budget Lines, One Balance Sheet

Adaptation and mitigation entered climate policy as separate negotiating tracks, and the accounting followed. The result is a financial architecture that struggles to fund the assets that do both.

1.1 What the accounts say

The Global Landscape of Climate Finance 2025 records global climate finance at an all-time high of US$1.9 trillion in 2023. Mitigation took US$1,780 billion of it. Adaptation reached US$65 billion — a figure the report itself calls a likely underestimate, because adaptation is harder to identify in a project's accounts than a megawatt is. Dual-benefit finance reached US$58 billion.

The demand side is documented with equal clarity. The UNEP Adaptation Gap Report 2025 estimates developing countries' adaptation finance needs at US$310 billion a year by 2035 on modelled costs, or US$365 billion when extrapolated from countries' own national plans, against international public adaptation flows of US$26 billion in 2023 — down from US$28 billion the year before. Needs are twelve to fourteen times current flows, and the Glasgow goal of doubling adaptation finance by 2025 will be missed.

1.2 Why the split is the wrong question

Faced with those numbers, the instinctive response is to argue for reallocation: adaptation deserves a larger share. That argument is sound as far as it goes, and this report does not dispute it. But it accepts the premise that the two agendas are competing uses of one pot, and for economies building their energy systems and their protective infrastructure at the same moment, that premise is mostly false.

The relationship is one of production, not allocation. Adaptation services consume electricity, so the character of the power system determines whether protection is bought at the price of more emissions. Clean energy assets are physical objects in a hazardous climate, so their resilience determines whether the emissions saving persists. Each agenda is an input to the other's output. Dual-benefit finance, the smallest line in Figure 1, is the only one that describes investments where this is recognised.

2. Protection Runs on Power

Almost everything that protects people from a hotter, more volatile climate is an electrical load. That makes clean generation a precondition for adaptation at scale, not a competing use of the money.

2.1 Cooling

Cooling is the clearest case, because it is simultaneously the largest single adaptation technology in the world and one of the fastest-growing electricity loads. On the IEA's The Future of Cooling, air conditioners and electric fans already account for nearly 20 per cent of all electricity used in buildings and about 10 per cent of global electricity consumption. Without efficiency action, energy demand for space cooling more than triples by 2050; around two-thirds of the world's households could have an air conditioner, with China, India and Indonesia accounting for half the global stock. Cooling will drive peak demand, especially in hot countries.

Read from the adaptation side, that is protection from heat — which kills, closes schools and cuts labour productivity. Read from the mitigation side, it is one of the largest incremental demands on the grid of the next quarter-century. Those are not two facts. They are the same fact, and it can resolve in either direction: met by coal, cooling converts heat exposure into emissions; met by solar, whose output peaks broadly with the cooling day, it converts heat exposure into demand for clean capacity.

2.2 Water, health and food

The same logic holds across the protective portfolio, and the numbers are substantial:

• Irrigation. Pumping water is how farmers adapt to erratic rainfall, and it runs on energy. India has installed roughly 900,000 standalone solar irrigation pumps under Component B of its PM-KUSUM scheme, against a target of 1.4 million — adaptation delivered as distributed generation.
• Health facilities. The WHO estimates that close to a billion people in low- and lower-middle-income countries are served by health facilities with unreliable electricity or none at all (Energizing Health, 2023); in sub-Saharan Africa about 15 per cent of facilities have no electricity and only half of hospitals have a reliable supply. A clinic that cannot refrigerate, sterilise or run a fan is not a climate-resilient clinic.
• Cold chains. Vaccines, insulin, blood and perishable food all require continuous refrigeration, and heat makes the requirement harder exactly when power is least reliable.

3. Power Runs on Protection

The dependency runs in the other direction with equal force. Clean energy assets are physical infrastructure exposed to the hazards they are meant to limit, and unhedged exposure shows up in the cost of capital.

3.1 Hydropower and drought

Hydropower is the largest source of low-carbon electricity in much of the developing world and the most climate-exposed. In 2024 the Zambezi drought took Lake Kariba's live storage — the water actually available for generation — down to eight per cent, and Zambia halted generation at the 1,080 MW Kariba North Bank station for the first time in its history, with only one of six turbines running for part of the year. Households faced outages of up to twenty hours a day; Zimbabwe, sharing the lake, faced up to twelve.

The exposure is structural rather than incidental: hydropower is about 86 per cent of Zambia's installed generating capacity. A power system that is overwhelmingly low-carbon was also, in that year, overwhelmingly unavailable. No emissions inventory records this as a problem, and every household did.

3.2 Heat, hail and wind

Even where water is not the constraint, heat erodes supply at the moment demand peaks. Analysis by Electric Insights puts the derating of a gas station at 40 °C at roughly 13 per cent of capacity and 7 per cent of efficiency against operation at 20 °C. Transmission lines lose thermal rating in the same conditions, and a study of heat-wave impacts on grid operation in Applied Energy finds heat waves simultaneously raising peak load and reducing generation and transmission capacity — the two curves moving apart precisely when cooling is most needed.

Variable renewables carry their own exposures, and the insurance market has begun to price them. Hail losses on utility-scale solar in Texas alone have exceeded US$600 million since 2018 on industry estimates, and the March 2024 hailstorm at the 350 MW Fighting Jays project in Fort Bend County reached its hail coverage sublimit, with insurers anticipating a US$50 million payout (VDE Americas; Renewable Energy World). The engineering response — stowing trackers at steep angles ahead of a storm, tougher glass, revised site selection — is adaptation spending inside a mitigation project.

3.3 Why this is a financing problem, not only an engineering one

Physical risk that is not designed out is priced in. The IEA's Cost of Capital Observatory, as reported in 2025, puts the cost of capital for solar at 5 to 6.5 per cent in advanced economies against a median 9.4 per cent in Indonesia, 9 per cent in Vietnam and 8 per cent in the Philippines. For capital-intensive technologies whose cost is almost entirely up front, that spread is a large tax on deployment — and every unmitigated physical risk, from drought correlation to hail exposure, is one of the inputs to it.

This turns resilience into part of the financing case rather than a charitable addition to it. A hydropower scheme with drought hedging, a solar plant with hail-rated glass and stow logic, a grid with fire-resistant corridors and flood-proofed substations: each is a cheaper project to finance than the same asset with the risk left in the discount rate.

4. Three Regimes: Joint, Enabling, Rival

Once the dependencies are visible, the useful question is not whether an investment is adaptation or mitigation, but which of three relations it occupies — and only one of them justifies the word co-benefit.

The interaction is not uniform, and treating it as uniformly positive is its own error. Three regimes cover the cases.

4.1 Joint: one asset, both duties

Joint assets discharge both duties in a single investment. Distributed solar with storage is the canonical instance: it displaces fossil generation and keeps power flowing when the central grid fails. When Hurricane Fiona took down Puerto Rico's grid in September 2022, solar-and-storage systems continued to supply tens of thousands of households, and a community microgrid in Adjuntas kept Casa Pueblo powered for nine days while the surrounding area went dark. Efficient cooling is another: it protects people from heat and suppresses the very load that protection creates. Solar irrigation pumps are a third.

4.2 Enabling: a precondition, not a substitute

Enabling assets make the other agenda possible without themselves delivering it. An electrified cold chain does not cut emissions on its own, but nothing else protects vaccines and food through a heatwave. Grid strengthening does not generate a clean kilowatt-hour, but without it neither the renewables nor the cooling can be delivered. The test for an enabling asset is whether the agenda it serves is achievable without it: where the answer is no, it should be appraised as part of that agenda's cost, not as an unrelated line item.

4.3 Rival: where they genuinely compete

Rival cases are real, and naming them is what makes the framework usable. Four recur:

• Water. A reservoir cannot simultaneously maximise generation and hold water for irrigation and drinking through a drought. Kariba in 2024 was a mitigation asset and a water-security asset in direct competition, and the scarcity resolved against both.
• Land. Utility-scale generation, food production and flood-buffering land uses compete for the same hectares near demand centres, most sharply where landholding is fragmented.
• Emissions. Air conditioning on a fossil grid protects the household and raises the emissions that drive the next decade's heat.
• Lock-in. Diesel generators bought as resilience are among the most common adaptation purchases in unreliable-grid economies, and they entrench combustion — the clearest form of maladaptation in the energy sector, and a reason distributed clean backup should be the default alternative.

Table 1 — The three regimes, with a screening test

The typology and its screening questions are the author's analytical framework, constructed with the cited cases already known; the fit is therefore not evidence for the framework. Its value is as a checklist applicable to a project not yet appraised.

5. The Build Window in Converging Economies

Integration is a design decision, and design decisions are only available before construction. That is why this argument is specific to economies building both stocks now.

In advanced economies, most of the energy system and most of the protective infrastructure already exist. Integrating them means retrofit: re-rating a substation, re-siting a plant, replacing an installed air conditioner. In converging economies, a large share of both stocks is still to be built — which means the integration can be specified in the original design, at a fraction of the later cost. A substation raised above the flood line during construction costs a small premium; raising it afterwards means building it twice.

The window is not open indefinitely, and it is narrower than the investment figures imply. The IEA's World Energy Investment 2025 finds that just one-fifth of global clean energy investment reaches emerging and developing economies excluding China, and that Africa accounts for around 2 per cent of clean energy investment while holding 20 per cent of the world's population. International public finance supplies only about 7 per cent — some US$32 billion a year — of clean energy investment in those economies.

Cost of capital compounds the problem in the same places, as Section 3.3 set out: 5 to 6.5 per cent in advanced economies against 8 to 9.4 per cent across several fast-growing Asian markets. A high discount rate is especially punishing for integrated assets, whose benefits are spread over a long life and across two objectives, and it is one reason the cheap, well-designed version of a project loses to the cheap-to-build one.

There is a further reason the window matters, and it concerns sequence rather than money. The protective demand and the clean supply have to arrive together. Cooling deployed a decade before clean capacity locks in fossil generation to serve it; clean capacity built without the resilience to survive a drought or a hailstorm will be unavailable in the years it is most needed. An earlier report in this series examined how climate hazard slows convergence through growth channels. The claim here is narrower and more practical: the assets that hold the hazard off are the same assets that decarbonise, if they are specified that way while they are still on the drawing board.

6. Case Evidence

Three cases, one for each regime: a rivalry that resolved against everyone, an enabling investment delivered at scale, and a joint asset tested by a hurricane.

6.1 Zambia and Kariba: the rival case, unmanaged

Zambia entered 2024 with about 86 per cent of its generating capacity in hydropower — on paper, one of the cleanest power systems in the world. The Zambezi drought then took Lake Kariba's live storage to eight per cent, and in September the country halted generation at Kariba North Bank for the first time in its history, having exhausted its annual water allocation. Outages reached twenty hours a day for Zambian households and twelve for Zimbabwean ones.

What makes this a rivalry rather than simply a drought is that the same water was needed for generation, for irrigation and for drinking, in a year when the agricultural economy was also failing. The allocation rule — a shared annual water quota between two countries — was the only instrument, and it distributed the shortage rather than resolving it. The lesson for the framework is that rival cases require an allocation rule agreed before the stress event, and a supply portfolio diversified against the hazard that the dominant asset shares.

6.2 India's solar irrigation pumps: the enabling case, at scale

Under Component B of PM-KUSUM, India has installed roughly 900,000 standalone solar irrigation pumps against a target of 1.4 million, in areas with limited or unreliable grid access. The pump is an adaptation asset — it buffers erratic monsoon rainfall — and a mitigation asset, because it displaces diesel pumping. It is also enabling in the strict sense of Section 4.2: without it, neither irrigation reliability nor the removal of diesel from the field is achievable for those farms.

The programme also illustrates the rival regime inside a joint asset, and the report would be dishonest to omit it. A pump with no marginal running cost invites groundwater extraction beyond the aquifer's recharge, which is precisely the kind of maladaptation that looks like success in the year it is installed. The instrument that resolves it is not a different pump but a water-allocation rule and metering — the shared scarce input, again, requiring a rule.

6.3 Puerto Rico's distributed solar: the joint case, tested

When Hurricane Fiona made landfall in September 2022 and Puerto Rico's central grid failed, distributed solar-and-storage systems kept running. A community microgrid at Casa Pueblo in Adjuntas held power for nine days; a solar microgrid in Castañer, commissioned only months earlier with 41 kW of panels and 74 kWh of storage across five sites, supplied the community while the surrounding island was dark.

The generation was low-carbon on an ordinary day and it was resilience infrastructure on an extraordinary one — the same hardware, no retrofit, no second budget line. It is the clearest demonstration available that the joint regime is not a rhetorical device: the asset was specified to island itself from the grid, and that specification is what made it both.

7. Recommendations

What follows for the actors who appraise, finance and build these assets.

7.1 National governments and planning ministries

• Plan the cooling build-out and the clean build-out as one programme. Pair every cooling access or heat action plan with the generation, storage and efficiency standards required to serve it, so that protection is not delivered as new fossil load.
• Write allocation rules for shared scarce inputs before the stress event. Water between generation, irrigation and drinking supply; land between generation, food and flood buffering. A rule negotiated in a drought is a crisis, not a policy.
• Diversify against the hazard the dominant asset shares. A system that is 86 per cent hydropower is exposed to one weather variable; solar, wind and storage are hedges against drought as much as against carbon.
• Make resilience a design requirement in clean-energy procurement — hail rating and stow logic, flood levels for substations, drought scenarios in hydropower planning — rather than an optional resilience annex.

7.2 Development finance institutions and climate funds

• Stop making projects choose a box. Dual-benefit finance is three per cent of tracked flows partly because instruments are labelled adaptation or mitigation; a joint asset should be fundable from either window without discounting the benefit the window does not name.
• Count both benefits in economic appraisal, and record the avoided damages of resilience features alongside abated tonnes, so the integrated design is not penalised for delivering two things.
• Price resilience into the cost of capital explicitly. Where designed-in protection lowers a project's physical risk, it should lower its financing cost; concessional instruments and guarantees should reward the specification, not just the megawatts.
• Close the adaptation gap on its own terms. Integration does not substitute for the US$284–339 billion annual shortfall UNEP identifies by 2035; much protective investment has no mitigation component and still has to be paid for.

7.3 Utilities, regulators and project developers

• Treat distributed clean backup as the default alternative to diesel. Diesel generators are the most common resilience purchase in weak-grid economies and the clearest energy-sector maladaptation.
• Plan for coincident stress. Heat raises demand while derating plant and lines; system adequacy studies should use the joint distribution rather than treating them separately.
• Electrify the protective load first where it is easiest to serve cleanly — health facilities, cold chains, water pumping — because these are the loads whose failure is measured in lives rather than in revenue.

7.4 Researchers and statistical agencies

• Measure dual-benefit investment properly. The category is acknowledged to be undercounted, and what cannot be counted cannot be targeted.
• Publish the rival cases. The co-benefit literature is rich on synergies and thin on the documented instances where protection and decarbonization competed and one lost. Those are the cases practitioners most need.

8. Conclusions

The integration question is not how to divide the money. It is which assets discharge both duties, which are preconditions for the other, and which genuinely compete.

Adaptation and mitigation arrived in climate policy as separate tracks, and the finance system inherited the separation: US$1,780 billion to mitigation in 2023, US$65 billion to adaptation, and US$58 billion — three per cent of the total — to investment recognised as doing both. In economies whose energy systems and protective infrastructure are being built simultaneously, that third category describes the most valuable assets available and the smallest line in the accounts.

The reason is that the two agendas are inputs to each other. Cooling, water pumping, cold chains and powered clinics are how people survive a hotter climate, and all of them are electrical load — around a tenth of world electricity goes to cooling already, and that demand more than triples by 2050. Clean supply, in turn, is exposed to the hazards it exists to limit: Kariba at eight per cent live storage, a gas plant losing 13 per cent of capacity at 40 °C, US$600 million of hail losses on Texas solar. Neither agenda is complete without the other, and in high-cost-of-capital markets the missing half is priced into every project.

None of this dissolves the genuine conflicts, and a framework that pretended otherwise would be useless. Water cannot simultaneously generate and irrigate; land cannot simultaneously host panels and crops; air conditioning on a coal grid protects a family and warms the decade; diesel bought as resilience locks in combustion. What the three regimes offer is a way to tell these cases apart before the money is committed — to count both benefits where an asset delivers both, to appraise an enabling investment as part of what it enables, and to write an allocation rule for a shared scarce input before the drought rather than during it.

The practical urgency is that integration is a design decision, and design decisions expire. The substation not raised, the plant sited without a drought scenario, the cooling programme built ahead of the clean capacity to serve it: each becomes a retrofit, at several times the cost, in systems that can least afford to build twice. The window is open now, in exactly the economies where both stocks are under construction, and the appraisal systems that would have to see it are the ones still asking each project to pick a box.

References

Every quantitative claim above is attributed inline. The principal sources are collected here.

• Climate Policy InitiativeGlobal Landscape of Climate Finance 2025 — US$1.9 trillion total in 2023; US$1,780bn mitigation, US$65bn adaptation, US$58bn dual-benefit. See also the press release.
• UNEPAdaptation Gap Report 2025: Running on Empty — needs of US$310–365bn a year by 2035 against US$26bn of international public flows in 2023; a gap of 12–14 times.
• IEAThe Future of Cooling — cooling at ~20% of building electricity and ~10% of global electricity; demand more than tripling by 2050; two-thirds of households with an AC by 2050; China, India and Indonesia half the global stock.
• WHOEnergizing Health: Accelerating Electricity Access in Health-Care Facilities (2023) — close to a billion people served by facilities with unreliable power or none; sub-Saharan Africa figures. See also the WHO fact sheet.
• Ministry of New and Renewable Energy, IndiaPM-KUSUM — Component B standalone solar pumps, roughly 900,000 installed against a 1.4 million target.
• Energy Regulation Board, ZambiaEnergy Sector Report 2023 — hydropower about 86 per cent of installed generation capacity.
• Reporting on the Kariba shutdownbne IntelliNews on the first-ever halt at Kariba North Bank at 8 per cent live storage; African Arguments and The Watchers on outage hours and turbines in service.
• Electric InsightsHow heat waves will change the power system — gas station capacity down about 13 per cent and efficiency 7 per cent at 40 °C against 20 °C.
• Applied EnergyQuantifying impacts of heat waves on power grid operation — coincident peak load increase with generation and transmission derating.
• VDE AmericasReevaluating hailstorm damage at the Fighting Jays solar project; Renewable Energy World on the March 2024 event; industry loss estimates above US$600 million on Texas solar since 2018 via Insurance Thought Leadership.
• IEACost of Capital Observatory, with 2024 values as reported by pv magazine — 5–6.5% in advanced economies against 9.4% Indonesia, 9% Vietnam, 8% Philippines.
• IEAWorld Energy Investment 2025 — one-fifth of clean energy investment to EMDEs excluding China; Africa 2 per cent of investment with 20 per cent of population; international public finance about 7 per cent (US$32bn).
• IREC and EESISolar microgrid keeps the lights on in Castañer and Microgrids in Puerto Rico keep rural communities connected — performance during Hurricane Fiona, September 2022.
• IPCCAR6 Working Group III, Summary for Policymakers — the assessed synergies and trade-offs between mitigation options and sustainable development, including where competition for scarce resources depends on the scale of implementation.
• H HeuristicsRelated reports in this series: the clean transition as mitigation infrastructure; the capital and delivery gates on adaptation; adaptation and convergence; and air quality as a co-benefit of clean energy in megacities.