What Most Corporate Carbon Reports Get Wrong and How Better Data Can Fix Supply Chain Emissions

Corporate carbon reporting has become a cornerstone of modern sustainability strategies. Companies now routinely publish detailed emissions inventories, set net-zero targets, and promise cleaner supply chains. But a new Stanford-led study shows that many of these reports share a serious and largely invisible flaw: they significantly undercount emissions from global supply chains, sometimes by billions of tons.

The research, published in Nature Communications, takes a close look at how companies estimate their so-called Scope 3 emissions—the indirect emissions generated by suppliers, manufacturing partners, and upstream activities. These emissions often make up the largest share of a company’s climate footprint, yet they are also the hardest to measure accurately.

According to the study, a widely used accounting shortcut is at the heart of the problem.

The hidden assumption distorting corporate carbon numbers

Many companies rely on a popular statistical approach to estimate supply chain emissions. This method uses spending data—how much a company spends on steel, electronics, machinery, or services—and then applies average emissions factors for each sector. The idea is simple and practical, especially when supplier-specific data is unavailable.

The issue arises from where those averages come from.

For years, one of the most commonly used tools was a single-region U.S. economic model, maintained until recently by the U.S. Environmental Protection Agency. This model implicitly assumes that all suppliers operate under U.S. production conditions, including the U.S. electricity grid, industrial efficiency, and fuel mix.

That assumption does not reflect reality.

Modern supply chains are deeply global, spanning countries with vastly different energy systems, manufacturing practices, and emissions intensities. Treating foreign production as if it occurred in the United States systematically pushes reported emissions downward, sometimes by a large margin.

How big is the undercount?

To quantify the gap, the researchers compared results from the single-region U.S. model with those from a multi-regional model that accounts for where suppliers actually operate.

The difference was striking.

For more than 400 companies reporting emissions data in 2023, the U.S.-only approach missed roughly 2 billion tons of carbon dioxide emissions from supply chains. That shortfall represents about 10% of total upstream emissions for those companies.

To put that number in context, 2 billion tons of CO₂ is roughly equivalent to the annual emissions of entire countries like Russia or India.

This isn’t a rounding error. It is a structural blind spot baked into many corporate climate disclosures.

Why China and heavy industry matter so much

A large share of the missing emissions comes from China, which alone accounts for approximately 973 million tons of the undercounted emissions identified in the study.

The reason is straightforward. China remains heavily reliant on coal-fired power, especially in energy-intensive industries. When companies import manufactured goods—such as steel, cement, machinery, or electronic components—from China, those products typically carry a much higher carbon footprint than equivalent goods produced in countries with cleaner electricity grids.

The biggest gaps showed up in sectors including:

• Steel and concrete
• Construction machinery
• Fabricated metal products used in cars and infrastructure
• Electronic components

When companies rely on U.S.-based averages, these emissions are significantly underestimated.

Missed risks and missed opportunities

Undercounting supply chain emissions doesn’t just affect reporting accuracy. It can distort business decisions.

If companies believe their upstream emissions are lower than they actually are, they may overlook opportunities to cut both emissions and costs by changing sourcing strategies. For example, sourcing energy-intensive products from countries with cleaner grids, such as the United States, France, or Brazil, could reduce overall emissions without sacrificing performance.

There are also growing financial implications. Starting in early 2026, Europe’s expanded carbon border tariff comes into force. This policy increases the cost of importing carbon-intensive goods like steel, aluminum, and cement into the EU. Companies that underestimate their upstream emissions may also underestimate their exposure to these new costs.

Knowing where emissions occur—not just how big the total number is—has become increasingly important.

Why companies use flawed models in the first place

The researchers acknowledge that companies are not ignoring global emissions out of negligence. The main barrier has been accessibility.

Multi-regional models are far more complex than single-region ones. They require detailed trade data, regional emissions factors, and extensive maintenance. Until now, there has been no widely available, easy-to-use open-source global model that companies could readily adopt.

As a result, many organizations defaulted to the simpler U.S.-based tools, even though they were never designed for globally distributed supply chains.

Building better tools with Cornerstone

To address this gap, the research team is now working on an initiative called Cornerstone. The goal is to make high-quality global emissions data freely available and easier to use.

The project is integrating the former EPA database with a multi-regional model analyzed in the study. That model was developed by Watershed, a private climate analytics company. The lead author of the study chairs Watershed’s science advisory board and previously served as its head of climate science.

The merged model is expected to be released in late 2026. Future versions aim to expand beyond industrial emissions to include land-use change and deforestation, which can dramatically alter the carbon footprint of commodities like soybeans or palm oil.

The researchers also collaborated with experts from World Wildlife Fund and CDP, organizations focused on reducing greenwashing and improving the real-world impact of corporate climate action.

Are spend-based models fundamentally flawed?

Not everyone agrees that sector-average, spending-based models are the best way to estimate supply chain emissions, even when they are global. Critics argue that supplier-specific data would always be more accurate.

The study’s authors largely agree—but point out that perfect data remains out of reach for most companies. Gathering detailed emissions information from every supplier would require much stronger regulations and far greater transparency than currently exists.

Until that happens, improving modeling approaches is still valuable. Better models can help companies prioritize hotspots, decide where deeper data collection makes sense, and avoid making decisions based on misleading numbers.

Why corporate supply chains matter for the climate

Corporate supply chains represent one of the largest untapped opportunities for global emissions reductions. Many companies are investing heavily in sustainability, supplier engagement, and cleaner production methods.

If those investments are guided by more accurate data—especially data that reflects global realities—they have the potential to drive meaningful reductions in carbon pollution worldwide.

Getting the numbers right is not just about better reporting. It’s about directing effort and capital to the places where they can do the most good.

Research reference

The full research paper can be found here:
https://www.nature.com/articles/s41467-025-67759-5