Energy density is the underrated property of fossil fuels. Wood contains about 16 MJ per kilogram; coal, 24-32; oil, 42-47 — and oil is liquid at room temperature, the easiest energy carrier ever discovered. Natural gas pushes higher still and burns cleaner. That density physically enabled the Industrial Revolution. The steam engine (Newcomen 1712, Watt 1769) needed coal to be commercially viable; charcoal was too expensive. Internal combustion (Otto 1876, Diesel 1893) needed liquid hydrocarbon fuel to be portable. Twentieth-century agriculture is built on natural gas (for Haber-Bosch ammonia) and oil. The story of the modern world is, in significant part, the story of finding, extracting, and burning increasingly large quantities of buried sunlight. About 80% of all primary energy consumed by humanity in 2024 still comes from fossil fuels.
Every previous energy transition added a source rather than retiring one. Britain in 1900 still consumed as much wood as in 1700; coal was added to wood, not substituted. Oil was added to coal; world coal use continued to rise as oil grew. Natural gas, accelerated by the shale revolution of the 2000s, has added to oil rather than retired it. Fossil fuels in aggregate now supply roughly four-fifths of all primary energy, and total consumption has risen more than fiftyfold since 1750. The pattern is structural: each new fuel class brought new applications, expanded the economy that demanded energy, and left the older fuel still needed for its existing uses. The climate target now requires the first energy transition in history that is genuinely a substitution — fossil-fuel use must fall to near zero, not just be diluted. The substitution faces a stack of obstacles. Energy density is the first: a liter of gasoline contains ~33 MJ; a liter of lithium-ion battery contains ~2.5 MJ at fourteen times the cost. For aviation, long-haul trucking, shipping, and high-temperature industrial heat, battery substitutes are difficult. Storage is the second: oil and gas store easily and indefinitely; renewable electricity requires active storage with cost and efficiency penalties. Sunk infrastructure — pipelines, refineries, distribution — represents trillions of dollars of capital that wants to be amortized. Political economy is the fourth: fossil-fuel industries employ ~30 million people worldwide and generate roughly $5 trillion per year. Vaclav Smil's pessimism (How the World Really Works, 2022) is grounded in the historical observation that energy substitutions of this scale take 50 to 80 years; the technologist's optimism is grounded in solar, wind, and battery cost curves that have exceeded every 2010-era projection. Both views are doing legitimate work, and the outcome is genuinely uncertain.
Coal use peaked globally around 2014 and has plateaued; China drove the post-2014 plateau, deploying renewables aggressively while still building coal plants. Oil consumption is near peak in OECD countries and rising in non-OECD; aggregate global demand may peak this decade depending on EV adoption. Russia's 2022 invasion of Ukraine and the weaponization of natural-gas exports to Europe accelerated European decarbonization in some respects and delayed it in others. The Inflation Reduction Act (US, 2022) is the largest single climate-policy commitment by a major economy — roughly $370 billion in clean-energy tax credits. China dominates manufacturing of solar PV (~80% of global capacity), wind turbines, batteries, and EVs. EU emissions-trading prices reached €60-90 per tonne in 2024, up from ~€5 in 2017. Stranded-asset risk is now priced into capital allocation. The transition is real, slow, and contested.