Energy debates still tend to begin with fuels, generation capacity and emissions. That framing made sense when the central economic problem was securing enough primary energy, then replacing high-carbon sources with cleaner ones. By mid-2026, however, a subtler constraint is moving to the foreground. Modern economies are electrifying transport, heating and parts of industry at the same time as they build out more data-intensive computing and confront hotter summers. In each case, electricity is only the first half of the story. The second half is heat: the heat produced by processors, motors, transformers, buildings, pipes and people, and the cost of moving it somewhere else.
That may sound like a technical detail. It is not. The ability to reject, recycle, store or live with heat increasingly shapes where industrial activity goes, how much grids must invest, what power prices look like at peak times and which cities remain economically competitive. Heat, in other words, is becoming an organising variable in energy economics.
Thermodynamics returns to the centre
Every economy runs on energy conversion, and every conversion creates losses. The elegant fiction of the digital age was that value could dematerialise even as physical throughput stayed manageable. Yet servers, cooling equipment, telecoms hardware, batteries, electrolysers, heat pumps and industrial electrification all remain bounded by thermodynamics. Even where efficiency improves, total service demand often rises faster. The result is not a simple shortage of electrons, but a more local and more awkward problem: too much low-grade heat concentrated in the wrong places.
Electricity is unusually versatile because it can do work before degrading. But once used, almost all of it ends as heat. A train motor, a graphics processor and an air conditioner all differ in function; in thermodynamic terms they converge. This matters economically because heat is much harder to trade than electricity or fuel. It must be managed where it is produced, or very near by. A market can ship molecules and route electrons across borders. It cannot easily export an urban heat island or a stressed river basin.
Every kilowatt-hour ends as heat, and modern economies are concentrating that fact in space and time.
Why this matters more in 2026 than it did a decade ago
Three trends are colliding. First, electrification is raising the share of final energy delivered as electricity. Second, computing infrastructure is becoming denser as demand for machine learning, cloud services and edge processing expands. Third, climate change is increasing cooling demand in both buildings and industry. The International Energy Agency has long warned that cooling is a major and growing source of electricity demand, while recent work on artificial intelligence and data centres has underlined the power implications of concentrated computing loads. Put together, these trends create a self-reinforcing cycle: more electricity use produces more heat, and hotter ambient conditions make that heat more expensive to remove.
This is not simply a question of total annual consumption. Power systems are built around peaks, local bottlenecks and reliability margins. A district with new data halls, electric vehicle charging, electrified warehouses and widespread air-conditioning can look manageable on a national balance sheet while proving costly on a substation map. The economics become intensely local. Cooling water, land-use rules, building design and proximity to heat networks can matter almost as much as wholesale power prices.
From generation economics to heat-sink economics
Every kilowatt-hour ends as heat, and modern economies are concentrating that fact in space and time.
Traditional energy strategy asks who can generate cheaply and reliably. The next layer asks who can absorb the consequences of that generation and use. This is where heat-sink economics enters. A heat sink is any medium or system able to take unwanted thermal energy away: ambient air, rivers, seas, aquifers, industrial loops, district heating networks, thermal storage or purpose-built cooling systems. These sinks are not free. They have seasonal limits, regulatory constraints and ecological side-effects.
Consider a familiar case. A data-processing facility may secure long-term electricity contracts and grid connections, yet still face binding limits from water availability, local planning or the inability to discharge heat without raising nearby temperatures. The same applies to many industrial processes. In a warmer climate, the margin between acceptable and unacceptable operating conditions narrows. What used to be a site-engineering consideration becomes an input to sovereign industrial policy.
The real competition, then, is not only for cheap power. It is for places with robust thermal infrastructure: cooler climates, resilient water systems, district energy networks, permissive but credible regulation, and the public capacity to coordinate land, grid and heat planning. Regions that ignore this may discover that they have procured generation yet failed to build a usable energy system.
AI sharpens the problem, but does not create it
Artificial intelligence has become a convenient symbol for energy stress because advanced computing clusters can draw large amounts of power in a compact footprint. But focusing solely on AI risks missing the wider structural shift. Dense computing simply makes visible a phenomenon already spreading through logistics, healthcare, telecoms, advanced manufacturing and consumer cooling. AI is best understood as an accelerant. It compresses demand spatially and temporally, exposing whether a region has invested in substations, backup systems, efficient cooling and pathways to re-use waste heat.
The economics are peculiar. A high-value computational workload can tolerate relatively expensive electricity if reliability is exceptional and curtailment rare. Yet its waste heat can impose costs on surrounding infrastructure and communities unless captured or dissipated intelligently. This means private profitability and system efficiency may diverge. Public authorities therefore face a planning task closer to transport network management than to old-style utility regulation.
The real competition is not only for cheap power, but for places that can absorb entropy at acceptable cost.
Waste heat is becoming an asset class of sorts
For decades, policy treated waste heat as a nuisance or a modest efficiency opportunity. That is changing. In northern and some central European systems, district heating networks can use low- and medium-temperature waste heat from industry, sewage systems and digital infrastructure, especially when paired with heat pumps. The European Commission has repeatedly highlighted district heating and cooling as a key efficiency lever because it turns a disposal problem into a supply resource.
Still, the economics are not automatic. Heat is low-value if there is no nearby demand, no pipe network, no seasonal storage and no temperature match between source and use. Retrofitting cities is expensive and institutionally difficult. The value of recovered heat therefore depends on density, planning coherence and the ability to socialise long-lived infrastructure costs. In some urban regions, the prize is substantial: lower gas dependence, improved system efficiency and reduced peak electricity stress in winter heating. In others, recovery schemes will remain marginal because the geography does not work.
The real competition is not only for cheap power, but for places that can absorb entropy at acceptable cost.
This distinction matters politically. Governments may be tempted to speak of circular energy systems as though all waste heat were recoverable. It is not. The practical question is where heat re-use beats the alternatives once network losses, pumping loads, seasonal variation and capital costs are included. Serious energy economics begins with those frictions, not with schematic diagrams.
Cooling demand is turning cities into power-system actors
Urban economics is also being rewritten. Air-conditioning, once treated in many temperate countries as a peripheral comfort load, is becoming a resilience technology for labour productivity, public health and basic urban functioning. The IEA and UNEP have both documented the likely surge in cooling demand as incomes rise and temperatures climb. For city governments, that creates a difficult triangle: protect residents from heat, contain electricity peaks and avoid deepening inequality between buildings that can cool safely and those that cannot.
The distributive implications are serious. Cooling poverty is becoming the warm-climate counterpart to fuel poverty. Households that cannot afford efficient cooling often live in buildings that trap heat and in districts with less tree cover and more asphalt. Their demand is most likely to surge during system peaks, precisely when prices are highest and grids most stressed. That makes heat not only a technical issue but a social one. An energy transition that neglects thermal comfort risks producing a cleaner system that remains economically punitive at the household level.
Water is the forgotten input
Discussions of sovereignty often emphasise fuels, rare minerals and semiconductors. Water deserves a place on that list. Many cooling systems rely on it directly, and heatwaves often coincide with water stress. Thermal power stations, industrial facilities and some large computing sites can find themselves competing with agriculture, ecosystems and municipal demand. In such circumstances, headline generation capacity means less than dry-year operability.
This reinforces the case for integrated planning. River temperatures, drought risk and wastewater availability increasingly belong in the same conversation as transmission build-out and industrial clustering. In policy terms, the separation between energy ministries, water regulators and urban planners is becoming harder to sustain. A state that wants resilient electrification must know not only how much power it can generate, but how much heat and water its economy can safely circulate.
Price formation will increasingly reflect thermal constraints
Energy pricing is often discussed as though markets clear on marginal generation cost alone. In practice, network congestion, balancing reserves, capacity mechanisms and weather-driven peaks already shape final prices. Thermal constraints add another layer. As cooling loads spike, as data infrastructure concentrates and as higher ambient temperatures reduce the efficiency of equipment, the effective cost of reliable power rises. Consumers may not see a line item labelled heat management, but they will pay for it in network charges, capacity payments, curtailment risk premiums and building retrofits.
This helps explain why some regions with ample renewable potential still face difficult power economics. Variable generation can lower energy costs over the year while local thermal and network constraints keep system costs elevated at critical moments. The challenge is not merely to produce clean electricity cheaply, but to deliver it with acceptable thermal overhead in the places and seasons that matter most. That is a sterner requirement.
Heat is no longer a by-product to ignore; it is an economic variable to govern.
Industrial policy will follow the map of thermal advantage
Just as past industrial eras clustered around coalfields, ports or gas pipelines, the next phase may cluster around thermal advantage. Cooler climates, access to seawater, strong district heating systems, abundant non-potable water, subterranean storage, and planning regimes capable of coordinating heat and electricity could all become location premiums. Conversely, regions facing chronic heat stress may need to specialise differently or bear higher infrastructure costs to stay competitive.
This will not produce a simple northward migration of industry. Labour markets, geopolitics, tax policy and proximity to customers still matter. But thermal advantage will increasingly influence which places host dense computation, pharmaceutical production, advanced materials processing or temperature-sensitive logistics. The industrial map of the 2030s may owe as much to heat rejection as the industrial map of the 20th century owed to fuel access.
Efficiency remains necessary, but it is not enough
There is a familiar policy instinct to treat efficiency as the universal answer. Better chips, better compressors, better building envelopes and smarter controls are all indispensable. Yet efficiency does not abolish thermodynamic reality. It reduces the amount of waste heat per unit of service; it does not eliminate waste heat, and it can be offset by scale. The history of energy use offers many examples where efficiency gains lower costs and thereby expand demand.
This does not make efficiency futile. It means its value must be judged alongside spatial planning, thermal storage, building codes, district energy, flexible tariffs and public-health adaptation. The systems that cope best will be those that treat heat as a shared infrastructure problem rather than a collection of private appliance choices.
Grid sovereignty now means thermal sovereignty
There is a tendency to define sovereignty in the energy sector as control over supply chains, generation assets and cyber resilience. Those remain essential, and institutions such as NIST have rightly stressed the security of critical infrastructure. But sovereignty in a hotter, more electrified world also includes the capacity to govern heat: to site loads prudently, protect water resources, recover waste heat where sensible, and ensure that the power system can survive prolonged thermal stress.
That is not glamorous statecraft. It consists of zoning rules, pipes, substations, weatherisation, public-health planning and dull but consequential engineering standards. Yet these are precisely the instruments through which advanced economies will decide whether electrification remains affordable and politically legitimate. If households experience cleaner energy as unstable summer bills and recurrent cooling failures, strategic narratives about transition will ring hollow.
Heat is no longer a by-product to ignore; it is an economic variable to govern. The countries and cities that grasp this will not escape thermodynamic limits. They will simply organise around them sooner, and at lower cost, than those that continue to treat energy as if generation were the whole game.
