Can You Estimate a Battery's Carbon Footprint for the EU Battery Passport?
Short answer: Not in the declared row. The carbon footprint on an EU battery passport is a declared value under Article 7 of Regulation (EU) 2023/1542, calculated from the maker's own production data under a Commission method that is still a draft. No maker we have checked has published one. None is legally due yet for any category. What you can do is show an indicative estimate beside the blank, clearly labelled as yours, built from chemistry, capacity, factory country and warranty, with a range. This guide explains the law, the unit trap that catches almost everyone and the method we use.

- The declared carbon footprint is per kWh of total energy delivered over the service life, not per kWh of capacity.
- Industrial batteries: due the later of 18 February 2026 and 18 months after both Commission acts take effect. No method act is drafted.
- An indicative estimate is shown in its own block with a range; no performance class is derived from it.
Jump around this page
- What does the Regulation ask for?
- When is it due?
- The unit trap: per kWh of what?
- How we build an indicative estimate
- A worked example: 10 kWh LFP home battery made in China
- How the estimate is shown and how it is not
- What a maker has to give us to replace the estimate
- What would change this page
- Frequently asked questions
- Get a free passport check
- Keep exploring
- Primary sources
What does the Regulation ask for?
Article 7(1) requires "a carbon footprint declaration" for "each battery model per manufacturing plant" containing, among other things, "the carbon footprint of the battery, calculated as kg of carbon dioxide equivalent per one kWh of the total energy provided by the battery over its expected service life", the same figure "differentiated according to life cycle stage" and "a web link giving access to a public version of the study supporting the carbon footprint values" (Regulation (EU) 2023/1542, Article 7). It applies to electric vehicle batteries, rechargeable industrial batteries over 2 kWh and light means of transport batteries. Not to SLI or portable batteries.
Annex II sets the method. The functional unit is "one kWh (kilowatt-hour) of the total energy provided by the battery system over the battery's service life". The four stages are raw material acquisition and pre-processing, main product production, distribution, then end of life and recycling. The use phase is excluded. Point 5 is the one that decides whether an estimate can ever be a declaration: "all activity data related to the battery's anode, cathode, electrolyte, separator and cell-casing shall refer to a specific battery model produced in a specific production plant. Accordingly, no default activity data shall be used."
Annex XIII point 1(c) puts "the carbon footprint information referred to in Article 7(1) and (2)" in the public part of the passport. Article 7(2) adds a performance class from A to E on a label; Article 7(3) adds a maximum threshold later still.
When is it due?
Each Article 7 date is the later of a fixed date and a lag after both the delegated act (the method) and the implementing act (the format) enter into force.
| Category | Declaration (Art 7(1)) | Performance class (Art 7(2)) | Status of the acts on 6 November 2026 |
|---|---|---|---|
| Electric vehicle batteries | 18 Feb 2025 or 12 months after both acts | 18 Aug 2026 or 18 months after | Method: draft published 30 April 2024, feedback closed 28 May 2024, not adopted (Have your say 13877). Format: draft, not adopted. Classes: no draft |
| Industrial batteries over 2 kWh (home and commercial storage included) | 18 Feb 2026 or 18 months after both acts | 18 Aug 2027 or 18 months after | Method: no draft act published (Have your say 14313). JRC technical rules exist |
| Light means of transport batteries | 18 Aug 2028 or 18 months after both acts | 18 Feb 2030 or 18 months after | JRC first draft rules of 23 June 2025; no act |
| Industrial batteries with external storage | 18 Aug 2030 or 18 months after | 18 Feb 2032 or 18 months after | Nothing published |
The arithmetic is plain. For an industrial battery, the declaration cannot be due before 18 months after an act that has not been drafted. The Commission guidance of 19 October 2026 marks data points 17 and 18 "Not to be filled/displayed as of February 2027- format still to be specified in the upcoming implementing act" (Commission guidance, data points by category).
The unit trap: per kWh of what?
Every published battery life cycle study you will find reports kilograms of CO2 equivalent per kilowatt hour of capacity. Values for lithium iron phosphate cells sit roughly between 35 and 115 kg CO2e per kWh across the literature. The most cited recent European figures sit in the 50 to 80 band.
The passport does not use that unit. It divides the whole footprint by the energy the battery delivers over its service life. The JRC's draft rules for industrial batteries set that energy as capacity multiplied by 365 full cycles a year multiplied by the warranty in years, with a five year fallback where no warranty applies (JRC draft rules for industrial batteries, June 2024). For light means of transport batteries the draft uses 30 cycles a year (JRC draft rules for LMT batteries, June 2025). The draft EV act uses 60 cycles a year for cars and asks for the result "with a resolution of 0,001 kg CO2-equivalent/kWh" (draft delegated act, April 2024).
So a 10 kWh home battery with a 10 year warranty divides its footprint by 36,500 kWh. A footprint of 74 kg CO2e per kWh of capacity becomes about 0.02 kg CO2e per kWh delivered. Both numbers are correct. Only one of them is what the passport asks for. A passport that prints 74 under the Article 7 label is wrong by a factor of several thousand.
How we build an indicative estimate
We call the method CF-1. It runs from data a passport already holds and from public factors with licences that allow commercial use. It produces an ActivateDigital estimate with a range, never a declaration.
Inputs from the battery record: chemistry, capacity in kWh, mass, cell factory country, pack factory country, warranty years and declared cycles where the datasheet gives them, plus the battery category, which picks the lifetime rule.
Inputs held once, versioned:
- A chemistry benchmark per kWh of capacity with a low and high, from published studies. For lithium iron phosphate we use a central 65 kg CO2e per kWh with a 50 to 80 range, anchored on a 2025 primary data study that found 50.0 kg CO2e per kWh including recycling (Jasper et al. 2025, EES Sustainability, CC BY), a 2024 cell level comparison in the 59 to 70 band and a 2022 pack level figure of 77.9 on the EU grid (Minviro for Transport and Environment, 2022). High nickel chemistries use a central 80 with a 55 to 110 range, drawn from European gigafactory modelling that found "from 46.5 to 126.5 kg CO2 eq." depending on chemistry and grid (Kolahchian Tabrizi et al. 2024, Energy, CC BY).
- Cell manufacturing energy of 30 kWh of electricity per kWh of cell, range 20 to 65, from the Fraunhofer finding that "it takes 20 to 40 kilowatt-hours of energy to produce a battery cell with a storage capacity of one kilowatt-hour" (Fraunhofer FFB on Degen et al. 2023), plus 10 kWh of gas.
- Grid intensity by country from Ember's yearly electricity data, released under CC BY 4.0: in 2024, China 559, Germany 340, South Korea 410, Poland 608 and the EU 212 grams CO2e per kWh (Ember yearly electricity data).
- A natural gas factor of 0.183 and a container ship factor of 0.0161 kg CO2e per tonne kilometre from the UK Government's 2025 conversion factors, under the Open Government Licence (DESNZ conversion factors 2025).
The calculation:
- Grid adjustment = cell electricity × (factory country grid − the 0.25 kg CO2e per kWh reference grid the benchmarks imply).
- Per kWh of capacity = benchmark + grid adjustment, split into the four Annex II stages: main production is electricity and gas; distribution is pack mass × distance × the freight factor; end of life is the study's recycling adder; raw materials is the remainder.
- Absolute footprint = per kWh of capacity × capacity.
- Lifetime energy = capacity × cycles a year under the draft rule for the category × warranty years. Per lifetime kWh = absolute ÷ lifetime energy.
- Repeat at the low and high ends of the benchmark and the electricity figure to get the range.
A worked example: 10 kWh LFP home battery made in China
Inputs: LFP, 10 kWh, 91 kg, cells and pack made in China, 10 year warranty, 6,000 cycles on the datasheet, 20,000 km by sea.
| Step | Result |
|---|---|
| Grid adjustment | 30 × (0.559 − 0.25) = +9.3 kg CO2e per kWh (range +6 to +20) |
| Per kWh of capacity | 65 + 9.3 = 74 kg CO2e per kWh (range 56 to 100) |
| Stage split per kWh | Raw materials about 56, main production about 19, distribution about 3, end of life about 0.5 |
| Absolute | About 780 kg CO2e for the battery |
| Lifetime energy, draft industrial rule | 10 × 365 × 10 = 36,500 kWh |
| Per lifetime kWh | 0.021 kg CO2e per kWh (range 0.015 to 0.028) |
| Same, using the 6,000 declared cycles instead | 0.013 kg CO2e per kWh |
Two things to notice. The capacity based number carries about a 30 % uncertainty from the benchmark spread alone. The lifetime based number moves by 40 % depending on which lifetime rule applies, which is why the estimate block always names the rule it used. For a light means of transport battery the two draft routes differ by a factor of eight, so for those we show both or neither.
How the estimate is shown and how it is not
The declared row reads "Not yet required (Article 7). Not declared by manufacturer." A separate block, headed as an ActivateDigital estimate, carries the central value with its range in both units, the method version, the inputs used, the factor sources and the date. No performance class letter is ever derived from it, because the classes do not exist yet. When the maker declares, the estimate block is replaced, not merged.
This mirrors what the calculation rulebooks themselves require of secondary data. The draft delegated act asks for a data quality rating on the declared value and for every secondary dataset to be reported in the study. The Global Battery Alliance rulebook says a proxy should be documented and "ideally overestimates the GHG impact" (GBA Greenhouse Gas Rulebook v2.1). The Catena-X rulebook says "When using secondary data, a conservative estimate shall be applied" and that proxy data "shall be documented and made transparent" (Catena-X PCF Rulebook v4). None of the passport data models, including DIN DKE SPEC 99100, has a flag for "estimated", which is exactly why the estimate needs its own block rather than a footnote on the declared value.
What a maker has to give us to replace the estimate
The carbon footprint declaration for the model and plant with the seven items in Article 7(1), the public version of the study with its data quality rating, the EU declaration of conformity number, the split across the four Annex II stages, the lifetime energy calculation (capacity, cycles a year, warranty or declared cycles) and written confirmation under Article 77(4) that we may publish it. Until the Commission acts exist, a study under the JRC draft rules, the GBA rulebook or the batteries PEFCR can be shown as declared by the maker under that rulebook. It still goes in the declared row with the rulebook named, because it is the maker's number, not ours.
No code? Request access
What would change this page
- Adoption of the delegated act on the carbon footprint method for EV batteries or the first draft for industrial batteries.
- Adoption of the implementing act on the declaration format, which fixes the unit and resolution on the passport.
- Adoption of the delegated act on performance classes, which would let a letter be shown.
- Publication of the final JRC rules for industrial batteries in a form that changes the 365 cycle lifetime rule.
Frequently asked questions
What unit is the battery passport carbon footprint in?
Kilograms of CO2 equivalent per kilowatt hour of the total energy the battery delivers over its service life, under Annex II of the Regulation. Not per kilowatt hour of capacity, which is what almost every published study reports.
Is a 10 kWh home battery an industrial battery?
Yes. The Regulation defines a stationary battery energy storage system as an industrial battery with internal storage, so the industrial dates and the industrial lifetime rule apply.
Can we use a green electricity contract to lower the estimate?
No. The draft delegated act does not allow contractual instruments in the method, so the country grid mix applies regardless of certificates or power purchase agreements.
Can an estimate ever become the declared value?
No. A declared value needs the maker's own production data for the main production stage, under Annex II point 5. An estimate is replaced by a declaration; it is not upgraded into one.
Why does the estimate carry a range?
Because its inputs are published averages. Showing a single number would overstate what public data can tell you about one battery from one plant.
No code? Request access
Get a free passport check
Bring one battery model and the documents you already hold. We'll show what an indicative carbon footprint looks like for it, in both units and what the maker would need to send to declare one. Request a free passport check.
Keep exploring
The questions this page usually raises next.
- CompareNext questionWhat Goes in a Battery Passport When the Manufacturer Hasn't Published the Data?Carbon footprint, recycled content, renewable content and who declared what are blank on every real battery passport.
- CompareNext questionIs Recycled Content Required on a Battery Passport in 2027?Recycled cobalt, lithium, nickel and lead shares are not due before August 2028 and cannot be estimated.
- CompareNext questionWho Is Responsible for the Data in a Battery Passport and How Do You Show Where It Came From?Article 77(4) puts passport accuracy on the operator placing the battery on the market.
Primary sources
-
EU law
Does this reach your products?
ActivateDigital builds the battery record from the documents you already hold, and says which of the 32 fields in our battery model your evidence answers.