Embodied Energy: What Was Spent Before You Bought It
Operational energy is metered and billed, which makes it visible. Embodied energy is spent by someone else, somewhere else, before the purchase, and appears in no bill the user ever sees. As operational efficiency improves, the invisible half grows relative to the visible one.
Where the Energy Sits
Three materials account for most of the embodied energy in the built environment. Cement is heated to around 1,450 degrees to produce clinker, which requires high-temperature process heat that is hard to electrify. Steel is reduced from iron ore, traditionally using coke as both fuel and chemical reducing agent. Aluminium is separated from its oxide by electrolysis, which takes a very large amount of electricity per tonne.
Cement has a feature the others lack. Roughly half its carbon dioxide comes not from the fuel but from the chemistry: heating limestone drives off carbon dioxide to leave calcium oxide, and that release happens regardless of how the kiln is powered. A cement plant running entirely on clean energy would still emit the process fraction, which is why cement is one of the strongest cases for carbon capture.
Aluminium is best understood as stored electricity. Producing a tonne takes on the order of 14,000 kilowatt-hours, so a smelter is effectively an electricity consumer that happens to output metal. This explains why smelters cluster around cheap hydropower and geothermal, why they are among the first industries to relocate when power prices rise, and why they are valuable participants in demand response.
Recycled material changes the arithmetic sharply for some of these. Remelting aluminium takes around five percent of the energy of primary production, because the electrolysis has already been done and cannot be undone. Steel from scrap in an electric arc furnace uses a fraction of the blast-furnace route. Cement has no equivalent shortcut, since the chemical change cannot be reversed.
Why the Ratio Is Shifting
In a poorly insulated building from the 1960s, operating energy over fifty years dwarfs the energy that went into construction, and embodied energy is a rounding error in the lifetime total.
Build the same building to a modern low-energy standard and operating energy falls by most of that, while embodied energy rises - more insulation, better glazing, mechanical ventilation and often more concrete. In a well-built passive-standard building, embodied energy can approach or exceed everything the building will consume in operation.
This does not argue against efficient buildings, and it is sometimes misused that way. It argues that once operation is efficient, the remaining question is what the structure is made of, and that a whole-life accounting is the only way to see the total. Regulations that set only an operational standard optimise half the problem and are increasingly being extended.
The same shift is happening in vehicles. A combustion car's lifetime energy is dominated by fuel. An electric car's is not: manufacturing, particularly the battery, is a much larger share, which is why lifetime comparisons depend heavily on how many kilometres the vehicle actually travels and on the grid it charges from.
The Problem With Measuring It
Embodied energy figures vary widely between sources, for the same reason EROI figures do: the system boundary. Counting the fuel at the factory gives one number, adding the energy in the raw materials gives another, adding the factory that made the machines gives a third.
A second problem is allocation. A refinery produces many products from one barrel, and a smelter produces metal and waste heat. Deciding how much of the input energy belongs to each output is a methodological choice, made differently by different practitioners, and it moves results substantially.
A third is data age. Manufacturing energy per unit has fallen for most products, and a coefficient from a decade-old database understates today's performance. Solar modules are the most affected: figures still in circulation reflect manufacturing that no longer exists.
The practical consequence is the same as elsewhere in this library. Numbers from a single consistently applied study can be compared with each other; numbers from different studies usually cannot. Anyone quoting an embodied energy figure should be able to say which boundary was used, and if they cannot, the number carries less information than it appears to.
What Reduces It
Using less material is the first and largest lever. Structural optimisation routinely finds that buildings contain more concrete and steel than they need, because over-specification is cheaper for the designer than analysis and carries no penalty. Design effort substitutes directly for material, and it is the cheapest reduction available.
Reuse beats recycling. A steel beam removed from one building and installed in another retains the energy already invested in it; the same beam melted and recast loses most of that and spends the remelting energy. Reuse is constrained by certification, documentation and the fact that buildings are rarely designed to be taken apart - which is a design choice rather than a physical limit.
Longevity is the same argument across time. An object that lasts twice as long halves its embodied energy per year of service, and this is where repairability, spare parts and modular design have an energy justification separate from waste.
Substitution comes last and is the most discussed. Timber replacing concrete in mid-rise construction, clinker substitutes in cement, hydrogen-based steel reduction - these are real and mostly early. They matter, and they matter less than not using the material in the first place, which is the consistent finding of whole-life studies and the one least represented in policy.
Frequently asked questions
What is embodied energy?
Everything consumed in extracting raw materials, processing them, manufacturing, transporting and installing a product, before it is used at all. It is invisible to the buyer because it was spent by someone else before the purchase, and appears in no bill the user ever sees.
Why is cement so hard to decarbonise?
Because roughly half its carbon dioxide comes from chemistry rather than fuel: heating limestone drives off carbon dioxide to leave calcium oxide, and that release happens regardless of how the kiln is powered. A cement plant running entirely on clean energy would still emit the process fraction.
Why is aluminium described as stored electricity?
Because separating it from its oxide by electrolysis takes on the order of 14,000 kilowatt-hours per tonne, making a smelter effectively an electricity consumer that outputs metal. This is why smelters cluster around cheap hydropower, relocate when power prices rise, and are valuable in demand response.
Can embodied energy exceed operating energy?
Yes, in efficient modern buildings. A 1960s building's operating energy dwarfs its construction energy, but a passive-standard building cuts operation sharply while embodied energy rises, and it can approach or exceed everything the building consumes in use. The same shift is happening in electric vehicles.
What reduces embodied energy most?
Using less material, since structural over-specification is routine and design effort substitutes directly for concrete and steel. Then reuse, which retains the energy already invested rather than spending remelting energy. Then longevity. Material substitution is the most discussed and matters less than not using the material at all.
