Energy Transitions: The Pattern Nobody Expected

The Statistics Nobody Quotes

Biomass - firewood, charcoal, crop residue - is usually treated as the fuel of the pre-industrial world. Global consumption of it is higher now than at any point in the nineteenth century, because population grew faster than the share using other fuels fell. Roughly 2 billion people still cook over solid fuels.

Coal follows the same shape. Its share of world primary energy peaked around 1910 at roughly half, and it has fallen steadily since - yet absolute coal consumption reached its all-time high in the 2020s, driven by Asian electricity demand. A falling share and a rising quantity are entirely compatible, and conflating them is the most common error in energy commentary.

Oil did not replace coal either. It took over transport, where coal was hopeless, while coal kept electricity and steel. Gas then took heating and a growing share of power without displacing either. At every stage the new entrant occupied ground the incumbent could not hold, and the incumbent kept the rest.

So the honest historical generalisation is not that energy sources replace one another. It is that they accumulate, and that total consumption has risen enough to make room for all of them. Any projection that assumes substitution is assuming something history has not yet produced.

Why They Took So Long

From the first commercial oil well in 1859 to oil supplying a tenth of world energy took about sixty years. Natural gas took a comparable span. Nuclear went from first commercial reactor to its share peak in roughly thirty-five years, and was the fastest large-scale transition on record until recently.

The delay was rarely the conversion technology. Steam engines, internal combustion and turbines all worked well before they were widespread. What took decades was everything around them: pipelines, refineries, tankers, filling stations, grids, standards, trained workers, and capital willing to fund assets that pay back over thirty years.

Existing infrastructure also resists. A coal plant built in 2015 has a design life into the 2060s, and a boiler, a pipeline or a vehicle fleet all embody decisions that keep producing demand for their fuel long after the decision looks wrong. This is lock-in, and it is why the age of a country's capital stock predicts its emissions path better than its stated policy does.

There is a partial counter-example worth keeping in view. Solar went from a negligible share to a substantial fraction of new capacity additions in under two decades, faster than any prior source, because a solar farm is a manufactured product installed in months rather than a megaproject built over a decade. Manufacturing scales differently from construction, and that may be the single most important structural difference in the current case.

How New Sources Actually Won

In each historical case, the newcomer did not beat the incumbent at the incumbent's own job. It did something the incumbent could not do at all.

Coal beat wood not on price - wood was free to anyone near a forest - but on concentration and availability. British forests were exhausted near the towns that needed fuel, and coal was dense enough to be worth transporting. Coal also enabled iron smelting at a scale charcoal never could.

Oil beat coal on portability. A coal-fired car is not a serious proposition, and neither is a coal-fired aircraft. Oil's advantage was that it is liquid, energy-dense and pumpable, which made engines possible that had no coal-fired equivalent. Where coal was adequate - stationary power, steel - it survived for another century.

Gas won on cleanliness at the point of use and on capital cost. A gas turbine is cheap and quick to build compared with a coal plant, which mattered enormously once electricity markets were liberalised and investors wanted shorter payback.

The pattern suggests the useful question about any new source is not whether it is cheaper per kilowatt-hour than what exists, but whether it does something the incumbent cannot. A source that is merely a slightly better version of an existing one competes against depreciated assets and usually loses; a source with a genuinely distinct capability creates its own ground.

What Is Different This Time

Three things have no historical precedent, and each deserves stating without overstatement.

First, cost. Previous transitions ran on sources that were more expensive at first and won on capability. Solar and wind are now the cheapest source of new electricity in most of the world, which inverts the usual sequence - the new source arrives already cheaper rather than growing into it.

Second, deliberate policy. Earlier transitions were driven by private advantage, with governments following. This one has explicit targets, carbon prices, subsidies and phase-out dates. Whether policy is strong enough is arguable; that it is present at a scale without precedent is not.

Third, demand. Every historical transition happened while total energy use was climbing steeply, which meant new sources could grow without anything shrinking. In wealthy countries demand has been flat or falling for two decades, so new capacity now takes market share rather than absorbing growth. Globally, demand is still rising, which is why the addition pattern persists worldwide even where it has broken locally.

The reasonable conclusion sits between the two familiar positions. The historical record genuinely argues that transitions are slower and less complete than announcements suggest, and it would be careless to dismiss it. The three differences are real, and it would be equally careless to assume the past simply repeats. What the history establishes firmly is that the outcome depends on whether old assets are actively retired, because nothing in the record shows a new source pushing an old one out on its own.