University of St Andrews finds metal carbon footprints may be 10x higher than thought

A new study from the University of St Andrews is throwing cold water on one of net-zero’s most important assumptions: the carbon footprint numbers used for metal production may be more than 10 times higher than previously estimated.

That is not a rounding error. In plain English, it means the emissions embedded in producing core industrial metals like steel and other metal products could be far larger than climate models, corporate reporting, and planning have been treating them as. If you have been building strategy around those figures, the strategy might be optimizing for a world that is smaller than reality.

Why does “10 times higher” matter so much? Because metal production is not a niche industry. Metals are the scaffolding of modern supply chains. They show up in construction, manufacturing, cars, power infrastructure, and a long list of the things people call “green” as long as nobody has to measure the emissions first. When a key input in those systems has a much bigger carbon footprint than expected, every downstream decision gets tugged: where you source from, what you buy, how you price risk, and how you justify decarbonization investments.

This study is also a reminder that carbon accounting is not just a spreadsheet exercise. Different estimation methods, boundaries, and datasets can move emissions results dramatically. In many cases, earlier estimates have been used in planning because they are available and usable at scale. But “usable” is not the same as “accurate,” especially when the gap is as large as an order of magnitude.

From a board and executive perspective, the practical question becomes: what happens to targets and timelines when the underlying footprint is wrong by a factor like this? Net-zero ambition is full of interconnected parts. Targets are set. Procurement commitments are made. Capital allocation plans are built. Supplier engagement programs get launched. If the footprint was underestimated, then the gap between current emissions and the path to net zero can be much wider than assumed. That affects the credibility of the entire narrative, not just one line item.

There is also a regulatory and reporting angle. Governments and regulators increasingly push companies toward more consistent, comparable, and verifiable emissions disclosures. Even when specific rules differ by jurisdiction, the direction is similar: less hand-waving, more traceable measurement. If the baseline carbon intensity for metal production is materially higher than previously estimated, disclosures based on those baselines may need to be revisited. That can trigger internal audits, revision of inventories, renegotiation of supplier data requirements, and renewed scrutiny from stakeholders.

Then comes the second-order effect that tends to sneak up on decision-makers: cost and risk modeling. Higher footprints often correlate with higher abatement needs and potentially higher compliance or transition costs, depending on the regulatory and market environment. Even if the industry ultimately finds ways to cut emissions, the investment timeline may be harder than expected. Assets built or contracts signed on the assumption of lower emissions intensity may face misalignment with later, corrected accounting.

For companies that buy metals, this research is not just “someone else’s problem.” Supply chain emissions are a central focus for many climate programs, and metal production sits near the core of industrial supply chains. When the footprint of a key upstream input jumps, procurement strategies and supplier engagement plans may have to be rebuilt. Executives may need to pressure suppliers for better data, validate methodology assumptions, and stress-test how supplier decarbonization claims translate into real emissions reductions.

For peers in the same universe, the message is simple: if your net-zero roadmap depends on metal emissions estimates, you should treat this as a signal to re-check the foundation. The study from the University of St Andrews does not just suggest incremental improvement to existing models. It suggests a potential order-of-magnitude underestimation, and that is the kind of discrepancy that can turn a “manageable transition” into a “major recalibration” fast.