A battery's carbon footprint is not one company's number. It is the sum of many companies' numbers, each calculated separately, then handed across a boundary and added together. Get that assembly right, and you have a figure that survives verification. Get it wrong, and you have a number no regulator or buyer will trust. The Global Battery Alliance pilots were built to test exactly this, and Minviro spent them as the LCA partner inside one of the consortia. What follows is what the work actually taught us.
This article is for the people who will have to produce and stand behind a battery's carbon footprint: cell manufacturers, cathode and anode material producers, and the upstream suppliers now being asked for primary data they have never reported. It covers what the GBA pilots were, why supplier data, not modelling, is the real bottleneck, and what the August 2026 Carbon Performance Class deadline demands.
The deadline that turns this from voluntary to mandatory
For most of the last decade, calculating a battery's carbon footprint was something a manufacturer did to stay ahead of the curve. That period is ending on a fixed schedule written into EU law.
The EU Battery Regulation phases carbon requirements in by battery type. EV batteries came first, with a mandatory carbon footprint declaration from February 2025. The next milestone is the one that changes commercial behaviour: from August 2026, EV batteries must declare a Carbon Performance Class, a rating of how carbon-intensive the battery is to produce. Carbon footprint thresholds, below which a battery cannot be sold at all, arrive after that.
A performance class is different from a declaration. A declaration is a number on a document. A class is a ranking against everyone else, visible to the buyer at the point of procurement. A battery that lands a worse class than a competitor's, on the same chemistry, loses on a criterion that has nothing to do with price or performance. That is the shift the industry is now pricing in.
The Battery Passport is the mechanism that carries this information: a digital record, mandatory for EV and larger industrial batteries from February 2027, that holds the carbon footprint, due diligence data, and recycled content, and passes them down the chain from producer to buyer to recycler. The carbon number is one field in that passport. Producing it credibly is the problem the pilots were built to test.
What the GBA pilots actually were
The Global Battery Alliance is a public-private collaboration founded at the World Economic Forum in 2017, working toward a sustainable and circular battery value chain by 2030. Its Battery Passport programme has run in waves: a proof-of-concept in 2023, a larger round in 2024, and Operational Trials running into 2026.
The 2024 round was the greatest pre-competitive effort of its kind. Ten consortia, led by cell manufacturers representing more than 80% of the global EV battery market share, mobilised their own supply chains, from mining companies through refiners to recyclers, to report against seven rulebooks covering greenhouse gas emissions, due diligence, forced and child labour, biodiversity, indigenous peoples' rights, and circular design.
Minviro joined as the LCA partner in a consortium bringing together Sunwoda Electronic, one of the world's largest battery producers, traceability provider Minespider, and EV manufacturer Li Auto. Our role was the Greenhouse Gas Rulebook: taking emissions data from each company in that chain and assembling it into an end-to-end product carbon footprint.
That context matters for what follows. We were not handed a clean dataset and asked to run a model. We were working with companies, many of which were calculating product-level emissions for the first time, across different countries, processes, and levels of data maturity. That is the realistic version of what a battery passport requires, and it is where the lessons sit. Minviro is also named in the GBA's 2026 Operational Trials, the third and largest wave, in which seventeen consortia are now working with suppliers to map chains and collect data. Results are expected in June 2026.
The footprint has to be assembled, not just calculated
Here is what the pilot made obvious. The GBA structure had each company report only its own in-house emissions, per unit of useful output. A cobalt mine calculates its emissions per tonne of cobalt. A precursor producer calculates its own and adds the inherited footprint of the cobalt it bought. The cell manufacturer does the same, one tier up. The full battery footprint is built by stacking these contributions in sequence, with the traceability provider tracking how much of each input ends up in the final product.
That sounds tidy on paper. In practice, the difficulty was rarely the arithmetic. It was getting each company to produce a defensible in-house number in the first place, in a consistent enough form that the next company up the chain could use it. Standardised templates helped companies running a greenhouse gas calculation for the first time follow the same structure, and allocate emissions correctly when one process produced several outputs. Without that consistency, the numbers do not add up, literally.
The practical implication for any battery maker: the bottleneck in your carbon footprint is almost never your own factory. It is the supplier two or three tiers upstream who has never measured their emissions, has no system to do it, and whose number, when it finally arrives, may not be in a form you can use. You cannot model your way out of that. You have to collect your way out of it.
Why supplier data is the real constraint
This pattern keeps appearing across the battery supply chain work, not just in the GBA pilots. The carbon impact of a battery is concentrated upstream, in mining and refining and the electricity used to process materials. Those are the stages furthest from the battery maker's direct control and hardest to get primary data from.
A cell manufacturer can measure its own assembly emissions accurately. But assembly is a small share of the total. The cathode active material, the refined lithium and nickel, the processing energy, these dominate the footprint, and they belong to other companies. If those companies report a generic industry-average figure instead of their actual production data, the final battery footprint inherits that imprecision, and a generic number is exactly what a Carbon Performance Class declaration cannot rest on.
This is the gap between a footprint that exists and a footprint that holds up. A spend-based estimate or an industry average will produce a number. It will not survive the verification the EU Battery Regulation requires, and it will not give a manufacturer a defensible position when a buyer compares their class against a competitor's. The difference is built on primary data, collected from real suppliers, in a consistent structure, with the methodology declared openly.
What does this mean for the manufacturer facing 2026
The pilots were in a controlled environment: companies volunteered, worked pre-competitively, and had the GBA's structure to lean on. The August 2026 Carbon Performance Class requirement will not be controlled or voluntary. It is the same data challenge, without the safety net, on a deadline. Three things carry over directly.
Start with the suppliers, not the model. The work that takes time is supplier engagement, getting upstream companies to measure and report their in-house emissions in a usable form. A manufacturer that waits until it has a tool before starting that conversation has the sequence backwards. The data collection is the long pole.
Build the chain once, then keep it. The value of the pilot structure was a repeatable way to assemble the footprint, not a one-off calculation. A passport is not a report you file once; the carbon number has to be maintained as suppliers change, grids decarbonise, and designs evolve. That is a living model, not a static document.
Declare the methodology, not just the result. The EU Battery Regulation uses LCA as its method, aligned to the Product Environmental Footprint approach and ISO 14040, 14044, and 14067. A number without a stated method, boundary, and data source is not verifiable, and verification is mandatory. The companies that did best in the pilots were the ones whose numbers a third party could follow back to their assumptions.
The battery passport is often framed as a compliance burden, and the data work behind it is real. But the same data that satisfies the regulation tells a manufacturer where its carbon actually sits, and therefore where to act. A Carbon Performance Class is a ranking, and a ranking can be improved. The manufacturers who treat the passport as a map of their supply chain, rather than a form to complete, are the ones who will move up the classes as the thresholds tighten. That was the real takeaway from the pilots. The hard part was never the life cycle assessment. It was building a supply chain that could feed one.
