Carbon loop

The barrel that is not in the tank

In 2018 the International Energy Agency named petrochemicals one of the key blind spots in the global energy debate. Fatih Birol said economies depend on them heavily, and the sector still gets less attention than it deserves. Oil is discussed as fuel. A rising share never reaches a tank. It becomes plastic, fertiliser, fibre, and parts of wind turbines, solar panels, batteries and cars.

The blind spot is the barrel that is not burned.

Oil is not simply disappearing. Part of it is changing jobs. Once that is admitted, a larger question follows. What happens to the materials needed to make the cleaner outlet?

An expanding material economy

The UN Environment Programme's Global Resources Outlook 2024 is blunt on scale. Resource extraction has tripled in fifty years. On the present course it could rise by about 60 percent from 2020 levels by 2060. This is not a transition inside a shrinking material world. It is a transition inside a growing one.

Efficiency does not automatically reverse that. The OECD finds that each unit of GDP can take less material, while total use still climbs. Under current policies, global materials use goes from about 96 gigatonnes in 2020 to about 145 in 2050. Relative decoupling is not an absolute fall. Throughput can rise while the charts look greener.

Fuel intensity, material intensity

Net zero does not cancel the pressure. It moves it. Oxford's EARTH programme was set up around that point. Jon Blundy, its research director, called the squeeze on critical raw materials an often overlooked part of the energy transition. Grids, storage, generation and vehicles all want more metal. In 2026 an Oxford Review of Economic Policy special issue put it more sharply. The world, the authors wrote, is in a new scramble for natural resources. This time not for fossil fuels. For critical minerals.

The Grantham Research Institute at LSE states the practical consequence. Low-carbon technologies are typically more materials-intensive than their fossil-fuel counterparts. An electric car, on the IEA comparison they use, takes about six times the materials of a conventional one. Fuel intensity falls. Material intensity rises. Dependence does not vanish. The object of dependence changes.

Cambridge adds a timing problem. Lukas Gast and Julian Allwood note that output of steel, aluminium and paper tripled in three decades. On a net-zero path, zero-emission electricity, biomass and carbon storage may not keep up. In their model, 2050 supply of those materials can sit up to 40 percent below demand. Clean machines can be designed faster than clean making of steel can be built.

Copper sits inside that lag. Electrification is copper-heavy. In a 2025 Oxford piece, Blundy and colleagues put it simply: in the coming decades the world will need more copper than has ever been mined. A new mine still takes about a decade. The question is not whether copper "runs out." It is how much of the transition plan assumes the material will arrive when the model needs it.

Oil that changes jobs

The IEA's recent oil outlooks are the same job-swap in barrels. In 2024, chemical feedstocks took about 70 percent of oil-demand growth by volume. China is the worked case. Over the five years to 2024, feedstock use there rose by more than the net increase in world oil demand. Gasoline can fall in one country while the factory still takes more oil.

Oil 2025 extends the split. From 2026, petrochemicals become the main source of oil-demand growth. By 2030, making polymers and synthetic fibres alone will take 18.4 million barrels a day. That is more than one barrel in six. Oil burned as fuel may peak around 2027 in that outlook. Oil used as material keeps growing through it. The same report has Chinese feedstock demand rising by about 1.1 million barrels a day over the forecast, offsetting a similar drop in gasoline. The barrel leaves the tank faster than it leaves the plant.

Plastic volumes follow. The OECD's baseline has production and use rising from 435 million tonnes in 2020 to 736 million in 2040. Recycled material stays about 6 percent of the total. Transport oil can stall. Virgin carbon chemistry can still grow.

The bill of materials

Carbon is not doing one job. In petrochemicals it becomes the product. In steel it is still part of how most iron is reduced. In cement, much of the CO₂ comes from heating limestone, not from the fuel under the kiln. The link is not a single molecule. It is carbon-intensive input under a low-carbon output.

The clean kilowatt still has a shopping list. The U.S. National Renewable Energy Laboratory puts steel in an offshore wind plant at 130 to 419 tonnes per megawatt. A one-gigawatt farm is then on the order of 130,000 to 419,000 tonnes of steel before it generates anything. Blades still need polymer and fibre. None of that is scrap left over from a coal station. It is the bill for the substitution.

Most of that steel still has to come from today's industry. About 70 percent of global production still uses the blast furnace–basic oxygen furnace route. Steel and cement together are 14 percent of direct energy and process emissions. Cement's direct intensity has not moved. Total cement CO₂ is higher than in 2015. The turbine can run without fuel. The kiln and the furnace that made it have not fully changed.

A cleaner use can also push carbon out of the picture that gets counted. Mining, refining, metals, polymers, factories. The outlet looks better. The chain outside the boundary may not.

The physical build-out of substitution therefore still needs large volumes of materials whose production has not yet transitioned.

Why progress does not yet break the loop

Recycling, efficiency and carbon capture exist. They are not matching the volume this decade. Recycled plastic stays a thin share while use is modelled to jump. Feedstock oil keeps growing after fuel oil, in the IEA outlook, has a chance of peaking. Clean steel and cement remain a small slice of what is poured and rolled. A local gain can be real. An electric car can displace petrol. The input to the whole machine can still be larger than last year.

That is a mismatch of rates, not a verdict that the transition cannot happen.

What current policy still produces

The World Energy Outlook 2025 Current Policies Scenario uses law that is already on the books. Energy-related CO₂ edges up, approaches about 40 gigatonnes a year in the early 2030s, and stays near that through 2050. In that scenario, total greenhouse gases point to about 2°C in 2050 and about 2.9°C by 2100. Under current policy, the physical system does not yet bend down.

More clean power and more electric miles can be added while total energy CO₂ stays close to record levels for decades. The outlet can change faster than the system underneath it.

The transition is not complete when the output becomes clean. It is complete when the system required to produce that output changes too.

illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.

The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.