What a textile Digital Product Passport holds

What the regulation actually says

The Digital Product Passport is created by the Ecodesign for Sustainable Products Regulation, Regulation (EU) 2024/1781, in force since 18 July 2024. ESPR is a framework: it establishes that products placed on the EU market will carry a passport, and delegates to later acts what each product group must put in one. Textiles are among the first priority groups.

That structure matters more than it sounds. ESPR itself does not tell a brand which fields to fill in. The delegated act for textiles does, and it is expected in 2027. What ESPR does fix is the shape of the obligation: the passport is data attached to a product, reachable through a carrier on the product, addressable by a unique identifier, and registered centrally so that authorities can find it.

So the correct reading is not that nothing is required until 2027. The infrastructure is being built now, in public, and the parts that are already running are the parts that constrain what the delegated act can ask for.

The commitment is also no longer implicit. The Commission's ecodesign work programme, adopted in April 2025, names textiles and apparel among the first product groups in line for requirements, which turned the sector's inclusion from an expectation into a published plan. The delegated act's status, and every development around it, is tracked in a dedicated file in this cluster, dated and change logged, so this pillar can stay about the substance.

The dates that have already passed

Two of them, on the same day. On 19 July 2026 the central EU registry for Digital Product Passports went live, and the ban on destroying unsold apparel and footwear took effect. The pairing is not a coincidence. A registry that can be queried per item and a prohibition on quietly disposing of stock are the same policy seen from two ends: the EU wants the fate of a garment to be a matter of record.

• 18 July 2024 · ESPR, Regulation (EU) 2024/1781, enters into force and creates the passport obligation as a framework.
• 19 July 2026 · the central EU DPP registry goes live, alongside the ban on destroying unsold apparel and footwear.
• 2027 · the delegated act for textiles is expected, fixing the data requirements for the product group.

What a passport carries

The concrete data model is being worked out in the open. The CIRPASS-2 project has published a minimum viable Textile DPP: the smallest set of attributes that makes a passport useful rather than decorative. In parallel, CEN-CENELEC JTC 24 is producing the European standards the delegated act will lean on for identifiers, data carriers and access rights.

Reading both is the fastest way to see where the requirements are heading.

Look down the right hand column. Every row that a sorting or recycling operator actually needs is either printed on the garment or held by a producer who has no relationship with the person now holding the item.

The standards work underneath

The delegated act will not invent its machinery from nothing. Two public workstreams are producing the parts it is expected to assemble, and both can be read today. CIRPASS defined the passport concept across sectors; CIRPASS-2 is deploying it in pilots, textiles among them, and its minimum viable Textile DPP is the closest thing to a draft field list in existence. In parallel, CEN and CENELEC set up JTC 24 to write the European standards for the system itself: unique identifiers, data carriers, access rights management and interoperability between passport systems.

The division of labour matters when reading either. The pilot work answers what a textile passport should say; the standards work answers how any passport, whatever it says, must behave. A delegated act can adjust a field list late in the process with modest cost. It cannot easily depart from the identifier, carrier and access machinery being standardised underneath it, because the registry and every early implementation will already have been built against them.

That asymmetry is practical guidance. An implementation that binds itself to the standards layer, resolvable identifiers, standard carriers, event vocabularies, shape validated records, is robust to whatever the final field list says, because fields are content and the machinery is structure. An implementation that hardcodes today's best guess at the field list has bet on the one part still guaranteed to move.

Identifiers, carriers and events

A passport is only as good as the thing that resolves it. GS1 Digital Link turns the identifier printed on an item into a web address, so one QR code can serve a consumer a care page and serve an authority a compliance record from the same starting point. It is the mechanism behind the access tiers, and it is why the carrier and the data model have to be designed together.

The carrier itself is less settled than the identifier behind it. A QR code is the default expectation for garments because anyone can read one, but the standards work keeps data matrix codes and NFC in scope, and a woven or laundered carrier has to survive what a garment survives. What the regulation fixes is the function, a machine readable carrier on the item resolving to the passport, not the printing technology; the delegated act is expected to lean on the JTC 24 carrier standards rather than legislate a symbology.

Static attributes are only half of it. GS1 EPCIS 2.0 models what happened to an item: when, where and why. A passport that records only what a garment was when it left the factory says nothing about the eight years since. Event data is what turns a passport from a specification sheet into a history.

EPCIS matters for circularity because its vocabulary already covers the awkward parts of a second life. An object event records that this item was observed at this place for this reason, which is what a sorting line produces every time a garment crosses it. An aggregation event records items becoming a bale and a bale becoming a shipment, which is how custody actually moves in the used textile trade. A transformation event records inputs becoming outputs, which is what recycling is. A passport regime that adopts EPCIS inherits, for free, the grammar of everything that happens to a garment after its first owner is done with it.

Parsing is not conforming

A passport that loads is not a passport that complies. Shape validation, using SHACL, is how a receiving system checks that a record has the fields it must have, that values come from the permitted vocabulary, and that relationships between them hold. Without it, conformance becomes a manual reading exercise performed too late, on too few items, by people who did not write the record.

Who sees what

ESPR splits passport data into tiers. A public tier is readable by anyone who scans the item: composition, care, origin, recycled content, substances of concern. A restricted tier is available only to credentialed parties, typically authorities and recyclers, and holds the supplier chain, test evidence and treatment history that a brand will not publish.

The split is what makes the passport politically possible, and it is also where most implementation effort goes.

What to build before the act is final

The waiting position is a trap: the act's adoption starts a transition clock, and the organisations that spend it building will be the ones for whom application day is uneventful. The useful work is exactly the work that does not depend on the final field list.

• Inventory the data groups in the table above against your own systems: for each one, name the source system, the owner, and whether the value is asserted or evidenced. The gaps found this way survive any revision of the field list.
• Link claims to evidence now. Recycled content, origin and substance declarations will need substantiation behind them; a claim recorded without its source document is rework waiting to happen.
• Adopt the identifier and event machinery early: resolvable identifiers, standard carriers, EPCIS shaped events. This is the layer the standards work has already stabilised.
• Validate shape from the first record. A conformance check bolted on at the end finds every mistake at the moment it is most expensive to fix.
• Decide the strategy for garments that come back. Take back, resale and recycled input programmes run on stock the passport regime will never cover retroactively, and that data has to be originated, not fetched.

The blind spot

Every part of the design above assumes a producer who assigned an identifier, populated the fields and published a record. That assumption holds for a garment made after the delegated act applies. It does not hold for anything already in circulation.

German collectors handle roughly 1.3 million tonnes of textiles a year and sort them into as many as 350 fractions, and less than 1% of textiles are recycled fiber to fiber. Almost none of that material carries a digital identifier, and almost none of it ever will, because the producers who would have issued one are not in the room. A garment made in 2019 will still be moving through collection and sorting infrastructure in 2032. The passport regime says nothing about it.

This is the structural gap: the rules require item level data to exist, and they do nothing to create it for the stock that dominates real waste streams. A passport obligation is not a data source. Someone has to originate the record.

There is one source that is already there. Nearly every garment carries a composition and care label, and those labels are regulated documents: fibre names are fixed by Regulation (EU) 1007/2011, care symbols by ISO 3758. Read at the point where the material physically is, on the sorting line, with a calibrated confidence on every field and a link from each value back to the document it was read from, that label becomes a passport ready record for an item nobody ever issued one for.

Common questions