Energy economics is usually told as a story about fuels, turbines, grids and prices per kilowatt-hour. Yet the more revealing timeline runs through a humbler variable: heat. Not high-temperature flame alone, but the low-grade warmth pouring from furnaces, pipes, chillers, data centres, electrolysers and buildings. For most of the fossil era, advanced economies treated that heat as background noise. By mid-2026, they are learning that this neglect was not merely inefficient. It shaped urban design, industrial geography, power pricing and strategic dependence.
The distinctiveness of the present moment lies in the convergence of three pressures. First, electrification is making power systems tighter and more politically salient. Secondly, AI-related computing growth is increasing local electricity demand while also producing substantial recoverable heat. Thirdly, industrial policy in Europe and parts of Asia has revived attention to district heating, heat pumps, thermal storage and waste-heat recovery as instruments of competitiveness rather than environmental virtue alone. The result is a quiet reordering of what counts as economically valuable energy.
1824-1910: The age that normalised waste
Modern industrialisation began with a thermodynamic compromise. Since Sadi Carnot, engineers have understood that converting heat into useful work is constrained by temperature differences and entropy. Steam engines, blast furnaces and later thermal power stations were never going to use all the energy they consumed. But the abundance of coal made that limitation tolerable. Industrial capitalism therefore matured with a permissive attitude to losses. Chimneys, cooling towers and hot effluent were signs of activity, not policy failure.
The habit mattered. Once cheap primary energy underwrote growth, firms optimised around throughput rather than whole-system efficiency. Cities zoned industry and housing apart. Factories vented steam. Power plants dumped heat into rivers or the atmosphere. Economic accounting rewarded output; it rarely rewarded avoided thermal loss. The modern energy economy was built on a habit of discarding heat.
1910-1973: Centralisation without thermal intelligence
The twentieth century deepened that pattern. Electricity, because it was versatile and tradable across distance, became the prestige carrier. Central generation grew, grids expanded and appliances colonised everyday life. Combined heat and power systems existed, and some cities built district heating networks, especially where density, climate and municipal capacity aligned. But these remained exceptions within a system organised chiefly around central power and one-way delivery.
Oil and gas strengthened the bias. They offered high energy density, flexible transport and, for decades, prices low enough to make local heat recovery appear parochial. The economics of convenience dominated. In many countries, buildings were designed for abundant fuel rather than thermal thrift. When energy was cheap, low-grade heat was nearly worthless because institutions had not been built to capture, move and price it.
1973-1989: The oil shocks briefly changed the question
The oil crises did not create thermodynamics, but they forced governments to notice it. Security concerns reframed efficiency as a strategic asset. Industrial countries introduced building standards, better boilers and some support for cogeneration. Denmark, among others, used the crisis to accelerate district heating and integrated heat planning. The lesson was simple: a nation vulnerable to imported fuel had reasons to care about every lost joule.
Even so, the shift was incomplete. As hydrocarbon markets stabilised and liberalisation gained ground, much of the structural ambition faded elsewhere. Efficiency policy survived, but often as a lower-status annex to supply policy. Heat remained technically important yet politically secondary, rarely carrying the symbolic weight of oil fields, pipelines or nuclear plants.
The modern energy economy was built on a habit of discarding heat.
1990-2008: Climate policy measured carbon, not always temperature
The rise of climate policy changed the ledger but not immediately the imagination. Carbon accounting made visible the cost of wasted fuel. European policy in particular began to treat heating and cooling as central to decarbonisation, because buildings and industry consume large shares of final energy. But public debate still favoured electricity-side spectacles: wind farms, gas disputes, carbon markets, later solar modules and batteries.
This mattered because heat infrastructure is less glamorous and more local. Retrofitting pipes under streets, integrating municipal waste heat, or aligning industrial sites with residential demand requires patient governance. Returns depend on planning law, utility regulation and building stock, not just technological breakthroughs. Many states pursued decarbonisation while leaving the thermal architecture of cities largely untouched.
The modern energy economy was built on a habit of discarding heat.
2008-2019: The digital economy produced invisible furnaces
As cloud computing, streaming and mobile platforms expanded, the energy debate about digital infrastructure focused mainly on electricity use. Data centres became symbols of rising power demand and efficiency gains in chips and servers. Less discussed was the fact that computation is physically a heat process. Every calculation ends as heat, and every server hall is therefore also a thermal plant of a peculiar kind: highly reliable, often urban or peri-urban, and generally cooled to protect uptime.
For years, that heat remained difficult to monetise. Data centres were frequently sited where power, fibre and land aligned, not where heat demand existed. Moreover, low-grade waste heat often requires upgrading through heat pumps before it can feed district systems. Where gas remained cheap, little incentive existed to build the connective tissue. The result was a strange asymmetry: economies worried about digital power consumption while ignoring a potential thermal by-product.
2020-2022: Pandemic shocks and the return of physical realism
The pandemic era exposed brittle supply chains, then the European energy crisis following Russia's invasion of Ukraine exposed something deeper: energy security is a systems problem, not merely a commodity problem. Households discovered that heating is where geopolitics reaches the kitchen table. Policymakers rediscovered that thermal demand, especially in winter, can dominate national vulnerability.
In Europe, the response broadened from fuel substitution to structural efficiency. Heat pumps moved from niche policy to industrial strategy. Renovation, thermal storage and district heating acquired a sharper strategic rationale. The European Commission's work on heating and cooling reflected a long-standing truth newly made urgent: if heating remains tied to imported molecules, sovereignty remains partial. As electricity became more valuable, heat stopped looking like an afterthought.
2023-2024: AI changed the map of local demand
As electricity became more valuable, heat stopped looking like an afterthought.
By 2023 and 2024, the discussion about AI and energy had become more concrete. Major studies from the International Energy Agency stressed that data centres could become significant sources of load growth in some power systems, even if global shares remained manageable. What mattered economically was not only aggregate consumption, but concentration. Large computing facilities can stress local grids, accelerate network investment and reshape wholesale price patterns.
That same concentration revived interest in co-locating data infrastructure with heat demand or industrial clusters. In principle, a data centre attached to a district heating network is not merely a consumer of electricity. It is also a supplier of usable warmth, particularly when paired with heat pumps and thermal storage. In practice, feasibility depends on climate, density, regulation and timing. Heat needs pipes; pipes need planning; planning needs institutions that can think beyond the meter.
The novelty here is not that AI servers produce heat. It is that in tighter and more politicised power systems, every by-product starts to look strategic. A megawatt consumed is no longer judged only by the value of its digital output, but by whether its thermal exhaust can displace gas, reduce peak demand elsewhere or anchor local energy resilience.
2024-2025: District heat stopped looking old-fashioned
For decades, district heating carried a mixed reputation: efficient in some Nordic and eastern European contexts, cumbersome or dated elsewhere. That view is changing. Once heat pumps, industrial recovery, wastewater heat, geothermal sources and data-centre waste heat are seen as inputs to a network, district heating begins to resemble a platform rather than a relic. Its economic role is to aggregate dispersed thermal sources and match them with variable demand.
This does not make it universally superior. Sparse suburbs, mild climates and fragmented property rights can weaken the case. Upfront capital costs remain substantial, and badly governed systems can lock in inefficiency as easily as they can reduce it. Yet the analytical shift is important. Economies are moving from seeing heat networks as municipal utilities to seeing them as sovereignty infrastructure, akin to ports, grids or water systems.
The implications for price formation are subtle. Electricity markets are increasingly shaped by variability, congestion and flexibility needs. Thermal networks can absorb some of this pressure by shifting when and how heat is produced. Large hot-water storage, for example, can convert cheap or surplus electricity into usable heat, effectively arbitraging time in a way that households with standalone boilers cannot. The scarce commodity was not energy in the abstract, but organised temperature.
As electricity became more valuable, heat stopped looking like an afterthought.
2025-2026: Industrial policy discovered the economics of temperature
By mid-2026, the link between competitiveness and heat is clearer. Sectors such as food processing, chemicals, paper, refining, steel and cement have long grappled with thermal intensity. What is changing is the policy framing. Waste-heat recovery and electrified process heat are no longer treated solely as emissions measures. They are increasingly understood as ways to reduce exposure to volatile fuel imports, relieve pressure on power systems and preserve industrial output under tighter carbon and security constraints.
This is where grid sovereignty acquires a more precise meaning. A sovereign grid is not simply one with enough generation capacity. It is one whose demand profile is governable, whose thermal loads are partly flexible, and whose industrial clusters can circulate heat rather than constantly purchasing new primary energy. In such a system, efficiency is not a moral preference. It is a buffer against scarcity, volatility and coercion.
The scarce commodity was not energy in the abstract, but organised temperature.
There are hard limits. Not all waste heat is equal. Temperature levels matter. Distance matters. Seasonal mismatch matters. Summer heat from computation does not automatically solve winter heating demand. Upgrading low-grade heat consumes electricity, and poor economics can follow if power prices are badly designed. Thermodynamics still sets the outer boundary. But within that boundary, institutional design determines whether heat is useless entropy or economic value.
Why pricing systems are lagging behind physics
Most energy markets still price electrons and molecules more clearly than they price avoided thermal loss. That creates distortions. A factory may have an incentive to optimise its own fuel bill while lacking any mechanism to sell excess heat to neighbouring buildings. A data centre may face strong pressure to secure power but weak incentives to invest in heat export if tariffs, contracts or planning rules are misaligned. Municipal authorities may want integrated systems but lack financing structures or legal powers.
This mismatch helps explain why apparently obvious efficiency gains often remain unrealised. Heat is local, bulky and time-sensitive. Unlike oil, it cannot easily be shipped around the world. Unlike electricity, it cannot be moved far without loss unless embedded in another medium. Economically, this makes heat less like a commodity and more like an infrastructural relation. Its value depends on proximity, synchronisation and governance.
The geopolitical meaning of low-grade heat
Viewed this way, the politics of energy in the late 2020s looks slightly different. The contest is not only over lithium, LNG, uranium or transmission lines. It is also over whether states can re-engineer urban and industrial systems to use lower-quality energy more intelligently. Countries that can capture, upgrade and distribute low-grade heat reduce fuel imports, moderate peak electricity demand and improve industrial resilience. Those that cannot may find themselves buying premium energy to perform tasks that cheaper organised warmth could have met.
This is a quiet geopolitical divide because it rarely appears in dramatic headlines. Pipelines and blackouts are visible; heat mapping is not. Yet the economic consequences accumulate. A city that can combine wastewater heat, industrial exhaust, heat pumps and thermal storage changes its exposure to gas prices. An industrial cluster that shares heat internally can stay viable under conditions that would squeeze more wasteful competitors.
The next decade will be built around thermal literacy
The strongest thesis for the rest of the decade is not that electricity becomes less important. It is that electricity's rising strategic value forces economies to become more literate about heat. Once transport, buildings and industry electrify further, wasting thermal energy becomes more expensive in system terms. Recovering heat can defer grid investment, cut fuel imports and soften political resistance to new power demand from AI and industry alike.
The challenge is cultural as much as technical. Energy policy has long preferred glamorous supply additions to mundane efficiency systems. But mature economies are running into the realities of land use, permitting, capital cost and public tolerance. Under those conditions, the cheapest unit of strategic energy may be a unit never wasted, or a degree of temperature moved from the wrong place to the right one.
That is why the return of heat matters. It marks a shift from an energy economics of abundance assumptions to one of thermodynamic discipline. The old model could afford to ignore low-grade warmth because fuel was often cheap and externalities were discounted. The emerging model cannot. It must treat heat as a governed asset, not a disposable residue. In that sense, the next century's base layer may not be defined by spectacular new energy sources alone, but by whether advanced economies finally learn to value the warmth already in their midst.
The scarce commodity was not energy in the abstract, but organised temperature.
