Battery Passport: what the Commission's 71 data points mean for 2027

The battery passport is not a new duty. It was written into Regulation (EU) 2023/1542, the EU Batteries Regulation, and Article 77(1) has carried the same sentence ever since: from 18 February 2027 each LMT battery, each industrial battery with a capacity greater than 2 kWh and each electric vehicle battery placed on the market or put into service shall have an electronic record, the battery passport. What changed in August 2026 is not the duty. It is how concretely a company can now prepare for it.

On 21 August 2026 the European Commission published an updated version of its guidance document, "Digital Batteries Passport - data points by category". The Commission's own Digital Product Passport for Batteries page records that document as last updated on 15 August 2026, and asks readers to check whether a more recent version exists before relying on it. The document brings together 71 data points, names the legal source of each one, and shows how each applies to the battery categories the passport covers.

It is worth being precise about what that document is. It is preparatory guidance. It was not published in the Official Journal, it is neither a delegated nor an implementing act, and it creates no obligation of its own: everything it describes traces back to the Batteries Regulation and to Annex XIII. What it adds is structure. A company that wants to know which fields it will owe, and which of them have to be answered on the first day, now has a Commission-authored map to work from instead of reading Annex XIII cold.

The guidance covers three categories: electric vehicle batteries, light means of transport (LMT) batteries, and industrial batteries. Article 77(1) puts a capacity threshold on the third of those, bringing industrial batteries into scope where capacity is greater than 2 kWh. LMT batteries and electric vehicle batteries carry no such threshold.

For each category, the guidance marks each of the 71 data points in one of four ways: mandatory, optional, applicable only in specific circumstances, or not required to be completed or displayed as of February 2027. That fourth marking is the one most likely to be misread. A data point that does not have to be displayed at the first stage is still a data point in the model. It is simply not something the passport has to show on the day the obligation starts.

The practical consequence is that 71 is the size of the vocabulary, not the length of a checklist every battery must complete. Applicability varies by category, several points apply only in defined circumstances, and a manufacturer reading the list will find fields that do not concern its product at all. Planning on the basis that all 71 fields are mandatory for every battery means planning for work the regulation does not ask for, and it is a good way to still miss the fields that genuinely do apply.

Publishing guidance neither postpones nor accelerates a statutory date. The battery passport requirement still begins on 18 February 2027, exactly as Article 77(1) has said since the regulation entered into force. A guidance document has no power to move that date, and nothing in this one attempts to. Read the two together and the position is unchanged: the law sets the date, the guidance helps companies arrive at it ready.

The access structure does not change either. Article 77(2) already splits passport information into three tiers, each keyed to Annex XIII: information accessible to the general public; information accessible only to notified bodies, market surveillance authorities and the Commission; and information accessible to any natural or legal person with a legitimate interest, for the purposes the article defines, principally dismantling, repair, remanufacture, second life and recycling. Whatever a company builds has to answer the question of who may see which field, not only the question of what the field contains.

It is tempting to read a list of 71 data points as a database schema and to treat the work as finished once the columns exist. The columns are the easy part. The difficulty sits in everything that has to happen before a value is allowed into one of them.

For each field, someone has to decide whether it applies to this product at all, find the source that establishes its value, and be able to say where that source came from. Values do not originate inside the company that owes the passport. A carbon footprint figure rests on a calculation performed somewhere upstream. A recycled content percentage rests on measurements taken at a plant the passport's author has never visited. Due diligence information on raw material sourcing rests on reporting from further up the chain again. Each of those has passed through several hands, and each hand copied, summarised or re-keyed it.

Then the record has to stay true. A battery passport is not a snapshot taken at the factory gate: Annex XIII carries fields that change during the battery's life, including its state of health and a status that moves through original, repurposed, re-used, remanufactured and waste. Someone has to own each field, keep the identifiers stable so that the record can still be found years later, and record what changed and when. And because the access tiers are real, the system has to tell disclosed information apart from restricted information, field by field, for the whole of that life.

The failure mode worth designing against is a familiar one: the passport becomes the first place a company discovers that its evidence is incomplete. By then the date is fixed, the field is visible, and the missing document is somewhere in a supplier's archive in another country.

The alternative is to treat the passport as the last step rather than the first. The sequence that makes that possible runs roughly like this: a source document arrives; an attribute is extracted from it; the attribute stays linked to the document it came from; that link carries provenance, meaning who supplied it and when; applicability is decided against the product's actual category; a human reviews what needs reviewing; the result is a coverage status that says plainly which fields are evidenced and which are not; and only then is any of it rendered as a passport.

Read in that order, the passport stops being a publishing problem and becomes a reporting one. The interesting output is not the finished passport. It is the coverage view produced long before it: the list of fields that apply, with the ones that have nothing behind them still showing as gaps, while there is time to close them.

None of the following is a certification checklist, and working through it establishes nothing legally. It is a preparation sequence, and its useful property is that every step can be taken before February 2027.

Determine which battery category applies to each product, and whether the 2 kWh threshold brings an industrial battery into scope. Map the data points that follow from that category, separating the ones marked mandatory from those that apply only in specific circumstances and those not required at the first stage. Identify, for each attribute that applies, the document or system that is its source. Then separate the evidence that already exists from the evidence that does not, and treat that second list as the actual project.

After that, decide who owns each field for the life of the record, establish persistent identifiers for the product and for the economic operator, and test that a machine can read the output rather than only a person. Preserve provenance and a change history, so that a value can be traced back to where it came from. Prepare the access rules Article 77(2) requires, field by field. And run a coverage review well before February 2027, while a gap is still something that can be closed rather than something that has to be explained.

AnyDPP organises product attributes around the evidence underneath them, rather than treating a passport as an isolated web page behind a QR code. In practice that means mapping attributes to the fields a regulation asks for, keeping each attribute linked to the source document it came from, carrying provenance on that link, scoring evidence coverage so that gaps stay visible, routing to human review where that is appropriate, and producing a portable passport built on open standards and identifiers. It prepares and structures evidence. It does not replace the legally responsible economic operator: the manufacturer or importer placing the battery on the market remains responsible for what the passport asserts.

The 71-data-point guidance makes the battery passport more concrete, and in doing so it makes visible a distinction that was easy to overlook while the passport was still abstract. Having a field in a data model is not the same as having defensible evidence behind that field. The companies best prepared for February 2027 will be the ones that start from evidence, ownership and provenance, rather than waiting to generate the passport interface and finding the gaps at the end.