Everything Is Becoming an Electricity Problem

Why climate policy, artificial intelligence, industrial strategy and energy security are increasingly converging on the same physical system.

An electric car looks like a transport technology. A heat pump looks like a heating technology. A data centre belongs to computing. An electric arc furnace belongs to heavy industry. An electrolyser producing hydrogen seems to belong to yet another technological system.

Increasingly, however, they are all asking the same question of the economy: can you supply enough electricity?

Not merely enough over the course of a year. Enough in the right place, at the right time, at a price that makes the activity viable, through a grid capable of delivering it, without making the wider system unreliable.

This is one reason debates about climate, artificial intelligence, industrial policy and national security are becoming harder to separate from debates about energy. Electricity is moving from being one important energy carrier among several towards something closer to a general-purpose infrastructure for the modern economy. Not everything can or should be electrified directly, but even some sectors that continue to depend on fuels or molecules may increasingly require electricity somewhere upstream.

The International Energy Agency now describes this shift as the beginning of an “Age of Electricity”. Since 2019, electricity demand has grown more than twice as fast as overall energy use, driven not only by electric vehicles and heat pumps but by air conditioning, industrial electrification, data centres and artificial intelligence. Electricity currently accounts for around 23% of global final energy consumption, but efficient electric technologies mean its share of useful energy services is larger; the IEA also estimates that electricity powers economic activities accounting for nearly half of global GDP.

The deeper energy transition is therefore not simply a movement from fossil fuels to renewables, or from one preferred generating technology to another. It is a movement towards an economy in which more of the work previously performed by burning fuels is performed through electricity, while new electricity-intensive activities are expanding at the same time.

Infographic generated with AI by the author.

More Electricity Can Mean Less Final Energy

There is an apparent contradiction at the centre of electrification. A heavily electrified economy can require substantially more electricity while using less final energy to provide many of the same services.

The reason is efficiency. Combustion throws away a large proportion of the energy contained in fuel as heat. An internal-combustion car converts only part of the chemical energy in petrol or diesel into motion, whereas an electric motor converts a much larger fraction of incoming electricity into useful work. The IEA estimates that an electric vehicle can be two to four times as efficient as its combustion-engine equivalent.

The difference can be even more striking in buildings. A boiler creates heat by burning fuel. A heat pump uses electricity to move heat from the surrounding environment into a building, typically delivering three to five units of heat for each unit of electricity consumed.

Electrification therefore does not require replacing every unit of fossil energy with an equivalent unit of electricity. It can reduce the amount of final energy required to provide transport, heating and many industrial services, while transferring more of that activity onto the power system.

Activity	Traditional input	Electrified pathway	What the power system must provide
Road transport	Petrol and diesel	Battery-electric vehicles	Charging capacity and additional generation
Building heat	Gas and oil	Heat pumps	Electricity during periods that can coincide with winter peaks
Industrial heat	Coal and gas	Electric furnaces, heat pumps and resistance heating	Large, often continuous industrial loads
Computing	Electricity	More computing, particularly AI	Concentrated and highly reliable power
Some fuels and chemical processes	Fossil fuels	Hydrogen, ammonia or synthetic fuels	Large quantities of low-carbon electricity upstream

This does not mean everything ultimately runs from a battery. Long-distance aviation, shipping, chemical production and some high-temperature industrial processes face physical constraints that make direct electrification difficult.

But that does not necessarily remove them from the electricity system. Hydrogen produced through electrolysis requires electricity. Synthetic hydrocarbons require electricity to produce hydrogen and capture or process carbon. Ammonia made from low-carbon hydrogen requires electricity upstream. Because each conversion step introduces losses, these indirect pathways can require considerably more electricity than direct electrification would.

Some of the hardest sectors to electrify may therefore become unusually demanding indirect electricity consumers.

That is one reason what net zero CO₂ actually requires is more complicated than replacing today's fossil electricity generation with low-carbon sources. If transport, heating and parts of industry are also to move away from fossil fuels, the electricity system must become cleaner and substantially larger at the same time.

Generating Electricity Is Only the First Problem

This makes arguments over which technology produces the cheapest electricity increasingly incomplete.

A solar farm may generate extraordinarily cheap power at noon. That does not automatically make the electricity useful to a factory hundreds of kilometres away at 7 p.m. in February. A nuclear reactor can provide large quantities of firm low-carbon electricity, but only after someone has financed it, built it and connected it to the transmission system. A battery can move electricity from one hour to another but cannot generate energy that was never produced. A transmission line can move electricity between regions but cannot solve a shortage affecting both ends simultaneously.

These technologies are not substitutes in the simple sense implied by arguments over whether solar, wind, nuclear or batteries will eventually “win”. They perform different functions inside the same machine.

That was the central argument of the energy transition is a system problem. Electricity systems combine generation, transmission, storage, flexible demand, interconnection and firm capacity. Their performance depends not only on the cost of individual technologies but on how those technologies interact.

The problem becomes more consequential as electricity takes over additional functions. Electrification can reduce dependence on continuously delivered fuels, but in exchange it places more transport, heating, industrial production and communication behind the same power infrastructure. The resilience of that infrastructure therefore matters more, not less, in an electrified economy.

Abundance and Scarcity Can Exist at the Same Time

One of the strangest features of modern electricity markets is that a system can simultaneously have too much electricity and not enough electricity.

Solar generation can exceed local demand during a sunny afternoon, pushing wholesale prices towards zero or below it. Wind farms may be curtailed because transmission networks cannot carry their output. A few hours later, demand may remain high after solar production has disappeared. Somewhere else, a proposed factory or data centre may be told that the grid connection it needs will take years.

Nothing about this is contradictory once electricity is treated as a system rather than a commodity measured only in annual terawatt-hours.

“Enough electricity” has several dimensions:

• Energy: was enough electricity generated over the year?
• Power: was enough capacity available at the moment of peak demand?
• Time: was electricity available when consumers needed it?
• Place: was it available where consumers needed it?
• Network: could the grid physically deliver it?
• Resilience: could the system keep operating when equipment, generators or transmission links failed?

Infographic generated with AI by the author.

A power system can possess enormous quantities of cheap generation while remaining deficient in one or more of these other dimensions.

This is why grids have moved from the background of energy policy towards its centre. The IEA's Electricity 2026 analysis of global grids estimated that more than 2,500 GW of renewable generation, storage and large new loads were stalled in connection queues worldwide. It estimated that annual grid investment would need to increase by roughly 50% from today's level of around $400 billion by 2030.

The timing mismatch is equally important. Large data centres can be developed in one to three years. Solar and wind projects can often be built considerably faster than major transmission infrastructure. New grid infrastructure may take five to fifteen years from planning through completion.

There is an appealing technological story in which rapidly falling solar and battery costs solve the energy transition through manufacturing scale. The less glamorous reality is that the transition also requires substations, transformers, distribution upgrades, transmission corridors, control systems, permits and construction crews.

The electrons may be cheap while the system required to move them is not.

The Grid Is Becoming a Different Machine

Electrification is also changing what exists on either side of the grid.

The traditional electricity system was dominated by large rotating generators producing alternating current and sending electricity outward towards comparatively passive consumers. The emerging system contains far more devices that naturally operate in direct current.

Solar panels produce DC. Batteries store DC. Electric vehicles contain large DC battery packs. Computers and servers ultimately consume DC internally. Modern power electronics constantly convert between different voltages and between AC and DC.

As explored in why the future grid is becoming increasingly bilingual, this does not mean alternating current disappears. It means the boundaries of the electricity system become more complicated.

The change is not merely electrical. Consumers themselves are becoming more active components of the system. A battery can consume power when electricity is abundant and return it when electricity is scarce. An electric vehicle can alter when it charges. Heating systems can sometimes preheat buildings or water. Industrial consumers may move some processes away from periods of extreme scarcity. Data centres may have limited ability to shift some computing loads in time or geography.

This creates flexibility without abolishing physical constraints. A household water heater can wait; a hospital cannot. A battery can move afternoon solar output into the evening but cannot necessarily bridge a prolonged period of weak wind and low winter solar production. A factory designed around continuous processes cannot casually shut down whenever electricity becomes expensive.

There is no single flexibility technology that removes the need for all the others.

AI Makes the Constraint Easier to See

Artificial intelligence did not create the electricity-system problem. It has made it unusually visible.

The IEA's 2026 updated assessment of energy and AI projects global data-centre electricity consumption rising from roughly 485 TWh in 2025 to around 950 TWh in 2030. That would still represent only about 3% of global electricity demand, but AI-focused data centres are expected to grow much faster than data centres as a whole.

The global percentage can obscure the local problem. A large data-centre campus may appear within a few years and demand hundreds of megawatts in one particular location. Several facilities clustered around the same transmission network can therefore become a substantial grid constraint long before data centres dominate national or global electricity consumption.

This is why AI's electricity use is only part of the infrastructure question. Asking how many watt-hours a particular AI query consumes tells us something, but not enough. Where is the data centre built? What generates its electricity? Is sufficient transmission available? Can any part of its demand shift? Who pays for the infrastructure required to connect it?

AI also illustrates a broader principle. If increasingly capable machines make some forms of cognitive work cheaper, other constraints become more important. As explored in when intelligence stops being the bottleneck, more abundant intelligence does not make electricity, semiconductor fabs, laboratories or physical construction disappear; it may increase demand for them.

A research system capable of generating vastly more useful computational work may discover that compute is scarce. More compute makes chips scarce. More chips make data-centre capacity scarce. More data centres make transformers, grid connections and electricity scarce.

Technological progress does not necessarily eliminate scarcity. It can move it downward until it collides with infrastructure.

Industrial Policy Is Becoming Energy Policy

The same logic extends beyond AI.

Governments increasingly speak about semiconductor sovereignty, battery manufacturing, green steel, hydrogen production and reshoring strategically important industries. Each ambition eventually encounters physical inputs.

A semiconductor fab requires specialised machinery, ultrapure materials, water and sophisticated technical expertise, but it also requires reliable electricity. Battery factories need energy as well as mineral supply chains. Electrified steelmaking requires sufficiently cheap power to compete with conventional production. Low-carbon hydrogen is effectively a claim on electricity generation before it becomes a fuel.

A government can subsidise the construction of a factory. It cannot permanently subsidise away an electricity system that makes operating the factory uneconomic.

Industrial strategy therefore increasingly depends on questions that once belonged mainly to utility planners. How quickly can a new plant obtain a grid connection? What will industrial electricity cost? Can the company secure a predictable long-term supply? Is there enough firm or flexible capacity? Can the regional network absorb several large new consumers at once?

The competition for industrial investment is consequently not only a subsidy race. It is also a competition between physical systems. Countries capable of supplying large quantities of reliable, reasonably priced and increasingly low-carbon electricity acquire an advantage that cannot easily be reproduced through tax credits alone.

Energy Independence Changes Its Meaning

Electrification also changes geopolitics.

Fossil-fuel systems depend on continuous flows of physical commodities. Oil crosses oceans in tankers. Gas moves through pipelines and LNG terminals. Coal arrives by ship and rail. These systems create obvious geopolitical vulnerabilities because energy supply can be interrupted through war, embargoes, pipeline failures or maritime chokepoints.

Europe experienced this directly after Russia's invasion of Ukraine and has continued to confront the wider vulnerability of internationally traded fuels. As examined in Europe's changing exposure to energy supply shocks, replacing one external supplier does not automatically remove the underlying dependence.

Electrification can alter that structure. A wind turbine does not require another delivery of wind. A solar panel does not need daily shipments of sunlight. Nuclear plants require fuel, but very large quantities of energy can be stored in relatively small volumes of material. The IEA's recent electrification analysis therefore emphasises that long-lived domestic electric technologies can reduce countries' continuous need for imported fossil fuels.

That does not amount to complete energy independence. Solar modules, batteries, transformers, power electronics and critical minerals have supply chains of their own. Nuclear programmes depend on fuel-cycle services and specialised industrial competence. Interconnected grids create mutual dependencies between neighbouring countries. Cybersecurity and physical protection become more important as additional parts of the economy depend on electricity.

Electrification does not eliminate dependency; it changes its structure. The strategic problem shifts partly from securing continuous flows of fuel towards ensuring that the infrastructure, equipment and industrial capacity needed to produce and move electricity remain available over decades.

Prices Cannot Build What Does Not Exist

Markets can coordinate much of this transition, but electricity exposes an awkward boundary between price signals and physical infrastructure.

High electricity prices can encourage generators to produce more where possible, consumers to reduce demand and investors to build additional capacity. Carbon pricing can change the relative economics of fossil and low-carbon energy by attaching a cost to emissions that would otherwise remain external to the transaction. That is the useful economic function discussed in why carbon taxes are prices rather than punishments.

Prices, however, operate inside physical possibility. A company cannot respond to attractive electricity prices by connecting a factory to a transmission network that has not been built. A renewable developer cannot sell electricity from a project waiting years for a grid connection. A battery in one region cannot resolve congestion hundreds of kilometres away simply because national generation is sufficient on average.

Infrastructure therefore sits uneasily between markets and planning. Too much central planning can protect favoured technologies, suppress useful price signals and commit capital to expensive mistakes. Too little planning can produce a different failure: individually sensible projects waiting for shared infrastructure that nobody had sufficient incentive or authority to build in advance.

Transmission lines often have to be planned before every future user exists. Spare capacity can look wasteful until demand arrives. Redundancy looks inefficient until something fails.

Markets remain indispensable, but some of the most valuable characteristics of an electricity system—capacity, resilience and optionality—must exist before their value becomes obvious.

What Electricity Abundance Actually Means

Calling for abundant electricity can sound like calling for unlimited consumption. That is not the useful definition.

Electricity abundance does not mean that every kilowatt-hour should be free, that efficiency no longer matters, or that countries should build arbitrary amounts of generation regardless of cost, land use and environmental effects. Nor can every country sensibly pursue the same mix of technologies.

Abundance means constructing a system in which access to electricity stops being a binding constraint on activities that society otherwise wants to electrify.

That requires more than abundant generation. It requires enough low-carbon production; transmission and distribution capacity; storage and other forms of flexibility; firm resources for periods when variable generation is weak; responsive demand where shifting consumption is practical; interconnection where geography can diversify supply; and enough redundancy to survive failures.

The cheapest generating technology is therefore not automatically the cheapest electricity system. Nor is the most technologically elegant system necessarily the easiest one for a particular country to build.

France, Sweden, Germany, Spain, Poland and the United Kingdom do not begin with identical grids, geography, hydro resources, nuclear expertise, industrial structures or political constraints. Their electricity mixes will differ.

The strategic objective is becoming more similar: create enough reliable electrical capacity that expanding electricity demand becomes an opportunity rather than a bottleneck.

The Energy Transition Is Larger Than the Energy Sector

Climate policy helped accelerate this process. Reducing carbon emissions encouraged cleaner electricity generation and the electrification of transport, buildings and industry.

The logic has since escaped the boundaries of climate policy. Artificial intelligence wants more electricity because computation is expanding. Industrial strategy wants more electricity because governments want semiconductor fabs, battery plants, hydrogen and cleaner heavy industry. Energy security favours technologies that reduce repeated exposure to imported fuels. Transport wants electricity because electric motors are efficient and batteries have become practical across much of road transport. Buildings want electricity because heat pumps can provide heating with far less final energy than direct combustion.

Even some sectors that remain dependent on fuels may demand electricity upstream if those fuels are eventually manufactured rather than extracted.

These interests do not always align politically. An AI company may prefer whichever electricity source can connect a new data centre fastest. Industrial consumers care about price and reliability as well as emissions. Local opposition can delay renewable generation, nuclear projects and the transmission infrastructure required by both. Electrification can reduce fuel imports while creating new dependencies on equipment and supply chains.

The convergence is therefore not ideological but physical. More and more ambitions are being placed upon the same infrastructure. That is why the central energy question of the coming decades may not be whether solar defeats nuclear, whether batteries defeat gas, or whether one favoured technology eventually becomes so cheap that every other option disappears.

Those arguments mistake components for the system. The more consequential question is whether societies can build electricity systems large enough, clean enough, flexible enough and reliable enough to carry functions that previously depended upon entirely different energy infrastructures.

If they can, electricity becomes one of the principal routes through which economies decarbonise, automate, industrialise and reduce their exposure to imported fuels. If they cannot, the electricity system becomes the bottleneck through which all of those ambitions must compete.

Everything is not literally becoming electric. But enough of the future is becoming dependent on electricity that the central constraint may increasingly be the same: whether the power system can grow fast enough to carry everything being placed upon it.