The global energy balance today

Framing

Every year, a small number of institutions publish a single accounting document that tries to capture everything humanity did with energy over the preceding twelve months. The most widely cited is the Energy Institute's Statistical Review of World Energy, the annual successor to the review that BP published for seventy years. It is a spreadsheet of a few hundred rows. It records how much coal, oil, gas, nuclear, hydro, wind, solar and the rest the world consumed, broken down by country and by fuel, in a common unit. When an analyst, a minister or a chief executive wants to know what the energy system actually looks like, this is the document they open first.

In its 2025 edition, covering the year 2024, the headline number was 592 exajoules (Energy Institute 2025). That is the total primary energy the world supplied to itself in a single year. Lesson 1.2 gave the tools to feel the size of that figure: 592 exajoules is the energy that would be released by burning roughly fourteen billion tonnes of oil, or by running every power station on Earth for a year, which is, of course, exactly what it represents. It is the largest quantity of energy human beings have ever marshalled in one year, and it grew by about two per cent over the year before (Energy Institute 2025).

A balance sheet has two sides, and they must agree. In a company's accounts, assets equal liabilities plus equity. In an energy balance, supply equals demand: every joule extracted from the ground or harvested from the wind ends up consumed somewhere, or stored, or lost in transformation. The skill this lesson teaches is reading both sides of that identity and understanding why the two sides describe the same joules in completely different language. By the end you will be able to look at the world's energy in a single page, from the wellhead and from the wall socket, and know what you are looking at.

Key terms

Four terms are needed before the balance can be read. The first three were introduced in Lesson 1.3 and are recalled here only in their role within the balance sheet. The fourth is new.

1. Primary energy. Energy in the form in which it is first captured from nature, before any transformation. Crude oil at the wellhead, coal at the mine mouth, the kinetic energy of wind at the turbine, the heat of a geothermal reservoir. The supply side of the balance is counted in primary energy. Introduced in Lesson 1.3.
2. Total final consumption (TFC). The energy delivered to the final consumer, after the losses of conversion and transport have been taken out: the petrol in the tank, the electricity at the meter, the gas at the burner. The demand side of the balance is counted in final energy. The gap between primary energy and final consumption is the transformation system, principally electricity generation, and it is large. Global final consumption in 2024 was a little over 450 exajoules, against the 592 exajoules of primary supply (IEA 2024). The difference, roughly 140 exajoules, is mostly the heat rejected by power stations, exactly the second-law tax quantified in Lesson 1.1.
3. The substitution method. An accounting convention for counting the primary energy of electricity that was never burned into being: hydro, wind, solar and, under some conventions, nuclear. Because these sources produce electricity directly, with no combustion, they have no obvious primary-energy figure. Two conventions exist. The older input-equivalent or substitution method inflates their output to the amount of fossil fuel a thermal power station would have burned to generate the same electricity. The newer physical-energy-content method counts only the electricity actually produced. The choice changes the apparent renewable share by a factor of two to three, a point developed in Lesson 1.3 and revisited below.
4. Energy intensity. The quantity of primary energy consumed per unit of economic output, usually expressed as energy per unit of gross domestic product. It is the single most useful ratio for comparing the energy efficiency of whole economies and for tracking whether growth is becoming less energy-hungry over time. Example: global energy intensity has fallen by roughly one to two per cent per year for decades, meaning the world economy produces more output per joule each year, even as total joules rise. Non-example: energy intensity is not the same as carbon intensity, which measures emissions per unit of energy, and the two can move in opposite directions.

The supply side: what the world runs on

Begin with the side of the balance the industry is built around: where the 592 exajoules came from. The structure has been remarkably stable for a generation, and it is worth fixing the numbers in memory, because almost every argument about energy and climate is in some sense an argument about how fast this particular set of proportions can be made to change.

In 2024, oil supplied 199 exajoules, about 34 per cent of the total, the single largest source, as it has been since it overtook coal in the 1960s (Energy Institute 2025). Coal supplied 165 exajoules, about 28 per cent, and reached a record absolute level despite decades of predictions of its decline (Energy Institute 2025). Natural gas supplied 149 exajoules, about 25 per cent (Energy Institute 2025). Those three fossil fuels together accounted for close to 87 per cent of all primary energy. That figure is the most important single number in the energy transition debate. After two decades of rapid growth in wind and solar, after trillions of dollars of investment, the fossil share of global primary energy has fallen only a few percentage points from its level at the turn of the century.

The remaining 13 per cent divides among the non-combustion sources. Nuclear supplied about 31 exajoules, a little over 5 per cent (Energy Institute 2025). Hydroelectricity supplied about 16 exajoules, under 3 per cent (Energy Institute 2025). Everything else, the category the review labels renewables and which is dominated by wind and solar but also includes modern bioenergy and geothermal, supplied about 33 exajoules, under 6 per cent (Energy Institute 2025). Wind on its own provided about 9 exajoules and solar about 8 (Energy Institute 2025). These are the fastest-growing numbers in the whole table, expanding at around sixteen per cent in 2024, roughly nine times the growth rate of total demand (Energy Institute 2025). They are also, still, the smallest.

Two cautions attach to these renewable figures, and both were introduced in Lesson 1.3. The first is the accounting convention. The figures above use the physical-energy-content method, which counts the electricity wind and solar actually produce. Under the older substitution method, which inflates that electricity to the fossil fuel it displaces, the same wind and solar output would appear two to three times larger, and the apparent fossil share would fall correspondingly. The 2025 review changed to the physical-content method, which is why a reader comparing it against older editions will find the renewable share has apparently shrunk (Energy Institute 2025). Nothing physical changed. The accounting changed. The second caution is that a primary-energy comparison flatters fossil fuels in a deeper way, because two-thirds of the primary energy in coal and gas is lost as heat in the power station, while the electricity from wind and solar suffers no such loss. A megawatt-hour of solar electricity does the work of roughly three megawatt-hours of coal once the second-law tax is paid. The primary-energy table does not show this; it is the single most common way in which energy statistics mislead.

The demand side: where the energy goes

Turn the balance sheet over. The same 592 exajoules of primary energy, after the transformation system has taken its cut, arrives at the world's consumers as a little over 450 exajoules of final energy (IEA 2024). The demand side is organised not by fuel but by sector, and the conventional division is into four: industry, transport, buildings and a residual of non-energy and agricultural uses. The proportions are less widely memorised than the fuel shares, but they matter just as much, because the transition has to happen in these sectors, one combustion process at a time.

Industry is the largest consumer, taking a little under a third of final energy. This is the energy of making things: the heat to smelt iron, fuse cement, crack hydrocarbons into plastics, and drive the motors of manufacturing. Much of it is high-temperature heat, which is difficult to supply with anything other than combustion, and this is why industry is among the hardest sectors to decarbonise, a theme Module 25 onward develops in detail. Transport takes roughly another quarter to a third, and is overwhelmingly oil: petrol and diesel for road vehicles, kerosene for aircraft, fuel oil for ships. Transport is the reason oil sits at the top of the supply table. Buildings, the third sector, take around 30 per cent of final energy across the economy, used for heating, cooling, lighting, cooking and the growing electrical load of appliances and data (IEA 2025a). The residential share of that is about 70 per cent, with commercial and public buildings making up the rest (IEA 2025a).

Cutting across all three sectors is electricity, and electricity is where the structure of demand is visibly changing. Electricity is not a sector; it is a carrier, a way of moving energy from the power station to the point of use, and it shows up inside industry, transport and buildings alike. Its share of final consumption reached about 20 per cent in 2024, up from 18 per cent a decade earlier, and it is rising about twice as fast as total energy demand (IEA 2024; IEA 2025b). In 2024, global electricity demand grew 4.3 per cent, against roughly 2 per cent for energy as a whole, driven by air conditioning, electric vehicles, the electrification of industrial processes and the rapid growth of data centres (IEA 2025b). The slow climb in the electricity share is the statistical signature of the energy transition. Every heat pump that replaces a gas boiler, every electric car that replaces a petrol one, moves a unit of demand from the fuel columns into the electricity row. The transition, read through the balance sheet, is the story of that row growing.

Why the two sides speak different languages

A reader who lays the supply table beside the demand table for the first time is usually confused, because the two sides do not share a vocabulary. The supply side is labelled oil, coal, gas, nuclear, hydro, renewables. The demand side is labelled industry, transport, buildings. There is no row called oil on the demand side and no row called transport on the supply side. The two tables describe the same joules, but they classify them on different principles: supply by where the energy came from, demand by what it was used for. The transformation system sits between them, and it is in that middle layer that the relabelling happens.

The clearest case is electricity. On the supply side, the coal, gas, nuclear, hydro and wind that feed power stations are counted as primary energy under their own fuel names. Inside the power station, they are converted into a single, undifferentiated product: electricity. On the demand side, that electricity arrives at a factory or a home with no memory of the fuel that made it. A unit of electricity from a wind farm and a unit from a coal plant are physically identical at the socket. So the fuel identity that organises the supply side is dissolved in the transformation layer and reconstituted, on the demand side, as the sectoral identity of who consumed it. This is not a flaw in the accounting. It is the whole point of an energy system: to take the messy variety of primary sources and deliver clean, fungible energy services at the point of use.

The same dissolution happens, less visibly, in a refinery. A barrel of crude oil enters as a single primary commodity and leaves as petrol, diesel, jet fuel, naphtha and a dozen other products, each destined for a different demand sector. The primary-energy table counts the crude once, on the supply side. The final-consumption table counts the products separately, distributed across transport, industry and buildings. Following a single joule from the supply table to the demand table therefore means following it through the transformation system, losing some of it to the second law on the way, and watching its fuel name be replaced by a use name. Mastering the balance sheet is, in the end, mastering that middle passage, which is why Lesson 1.3 on primary, final and useful energy had to come first.

Figure 1. The 2024 global energy balance, read from both ends. The supply side is counted in primary energy by fuel; the demand side is counted in final energy by sector, and therefore sums to the smaller final-consumption total once transformation losses are removed. Reading across the two sides is the core skill of this lesson. Sources: Energy Institute 2025 (supply, primary energy, physical-content method); IEA 2024 and IEA 2025a (demand, final consumption by sector). Sector shares are approximate and the demand column sums to the final-consumption total, roughly 140 EJ below primary supply because of transformation losses. Electricity is a carrier shown within the sectors, not a separate sector.

A worked example: tracing one country through the balance

Abstract shares become concrete when run through a single country, so take Japan, an economy whose energy balance illustrates almost every feature of the global one in miniature. Japan is a large industrial economy with almost no domestic fossil fuel, which means nearly everything it burns is imported, and its balance sheet is therefore unusually legible: the supply side is essentially a record of what arrived at its ports.

On the supply side, Japan consumed roughly 17 exajoules of primary energy in 2024, a little under three per cent of the world total, for a population of around 124 million (Energy Institute 2025). Fossil fuels dominate even more than in the world as a whole. Oil, almost entirely imported from the Middle East, supplies around 36 per cent. Coal, imported from Australia and Indonesia, supplies around 27 per cent. Natural gas, imported as the liquefied cargoes that Lesson 1.2 followed across the Pacific, supplies around 21 per cent. Together the three hydrocarbons meet over four-fifths of Japanese primary energy. Nuclear, which before the Fukushima accident of 2011 supplied around a quarter of Japanese electricity, has recovered only partially and contributes a single-digit share. Hydro and the new renewables, led by a rapid build-out of solar, make up most of the rest.

Now follow those imports into the transformation system. A large fraction of the coal and gas, and almost none of the oil, flows into power stations. Japan generates roughly a thousand terawatt-hours of electricity a year. Recall the combined-cycle gas turbine from Lesson 1.1, running at about 60 per cent efficiency, and the older thermal plants at nearer 40 per cent: the gap between the primary energy that enters Japanese power stations and the electricity that leaves them is the national share of the second-law tax, tens of exajoules of rejected heat venting from cooling systems around the coast. The single LNG cargo costed in Lesson 1.2, 3.8 petajoules of methane yielding about 630 gigawatt-hours of electricity in a modern plant, is one teaspoon of this flow. Japan imports the equivalent of dozens of such cargoes every week.

On the demand side, the relabelling is complete. The oil that arrived as crude is now petrol and diesel in the transport column and naphtha feedstock in the industrial column. The gas and coal that entered the power stations are now electricity, distributed across the air conditioners and trains and semiconductor fabs of the buildings and industry columns, their fuel identities erased. An analyst reading only the Japanese demand table would see industry, transport and buildings in roughly the global proportions, with an electricity share somewhat above the world average, reflecting an advanced, electrified, service-heavy economy. An analyst reading only the supply table would see a near-total dependence on imported hydrocarbons. Both are true. They are the same 17 exajoules, described from the two ends. Japan's entire post-war energy strategy, from its LNG contracts to its nuclear restarts to its hydrogen ambitions, is an attempt to manage the distance between those two descriptions: to keep the import-dependent supply side secure while the demand side electrifies. That is the balance sheet not as a static record but as the field on which national strategy is played, which is where Module 2 begins.

Retrieval and generative practice

Answer the first three from memory, without returning to the text. The fourth asks you to apply the lesson to a new case.

1. Without looking back, name the three fossil fuels in order of their share of global primary energy in 2024, and give the approximate combined percentage they represent.
2. Global primary energy supply in 2024 was about 592 exajoules, but global final consumption was only about 450 exajoules. Explain in one or two sentences where the missing 140 exajoules went, using a term from Lesson 1.1.
3. Why does electricity appear on the demand side of the balance distributed across industry, transport and buildings, rather than as a sector of its own? What does its rising share signify?
4. Generative practice. Pick a country you know something about that, unlike Japan, is a large fossil-fuel exporter, for instance Norway, Saudi Arabia or Australia. Sketch in a paragraph how its supply side and demand side would differ from Japan's, and explain why an exporter's balance sheet tells a different national story from an importer's. You are not expected to know the exact figures; reason from the structure.

Further reading

• Energy Institute. 2025. Statistical Review of World Energy 2025. The annual balance sheet of the world's energy. Start with the summary tables for total supply by fuel and by region, then read the methodology note on the change to the physical-energy-content method, which explains the apparent shrinkage of the renewable share.
• International Energy Agency. 2024. World Energy Outlook 2024. The companion to the supply-side review, strongest on the demand side and on the transformation system. The Sankey-style energy-flow diagrams are the best single visualisation of how primary energy becomes final consumption.
• Smil, Vaclav. 2017. Energy and Civilization: A History. For the long view of how the fuel shares in the supply table came to be what they are, and how slowly such shares have ever changed. The corrective to any expectation of rapid transformation.
• International Energy Agency. World Energy Balances (database, updated annually). The underlying data behind the demand-side figures, free to browse online. Worth opening once to see a full national balance, supply and demand reconciled in a single matrix, for any country of interest.

References

• Energy Institute. 2025. Statistical Review of World Energy 2025. 74th edition. London: Energy Institute.
• IEA. 2024. World Energy Outlook 2024. Paris: International Energy Agency.
• IEA. 2025a. Energy Efficiency 2025. Paris: International Energy Agency.
• IEA. 2025b. Global Energy Review 2025. Paris: International Energy Agency.
• Smil, Vaclav. 2017. Energy and Civilization: A History. Cambridge, MA: MIT Press.

Answers to the retrieval questions

1. Oil (about 34 per cent), coal (about 28 per cent) and natural gas (about 25 per cent), in that order. Combined, they supplied close to 87 per cent of global primary energy in 2024.
2. The missing energy was lost in the transformation system, principally as waste heat rejected by power stations when they convert fuel into electricity. This is the second-law tax from Lesson 1.1: no thermal conversion of heat into work is fully efficient, and at the scale of the whole system the loss amounts to roughly 140 exajoules a year.
3. Electricity is a carrier, not a primary source: it is generated from many fuels and then used inside industry, transport and buildings, where it is counted. Its rising share of final consumption, around 20 per cent and climbing about twice as fast as total demand, is the statistical signature of the energy transition, as each electrified end use moves demand out of the fuel columns and into the electricity row.

Key insight. The world's 592-exajoule energy balance is the same set of joules described in two different vocabularies — fuels on the supply side, sectors on the demand side. The two never carry the same labels because the transformation system in between dissolves fuel identity into the fungible carrier of electricity. Reading the balance fluently means reading both columns at once, and watching every primary joule turn into either useful service or rejected heat along the way.