The EU Liability change the Jewelry Trade
Twenty-Five Years: The EU Liability Change the Jewelry Trade Hasn’t Priced In
On 9 December 2026, the European Union replaces a product liability regime that has stood since 1985. Most of the commentary written so far about Directive (EU) 2024/2853 has been aimed at the automotive and software industries, where the headline change — liability for AI and connected products — is the obvious story. The jewelry trade has largely looked away.
That is worth reconsidering. Three provisions of the new Directive interact in a way that lands on our products harder than on almost anything else sold in Europe — and the reason has nothing to do with technology. It has to do with skin, time, and what a production batch is.
What actually changes
The Directive keeps the principle that has governed European product liability for forty years: strict liability. A claimant does not have to prove fault or negligence — only that the product was defective, that they suffered damage, and that one caused the other.
Three things are new, and all three matter here:
- The chain of liable parties is longer. Where the manufacturer sits outside the EU, liability runs first to the importer, then to the manufacturer’s authorised representative, and only then to the fulfilment service provider. Under the old regime, an authorised representative was not in scope for damages claims at all. It is now a named party.
- Proving a defect becomes easier. A product may be presumed defective where the defendant cannot produce evidence a court has ordered disclosed, or where the claimant demonstrates that the product did not comply with EU product safety requirements. Causation is presumed where the damage is typically consistent with the defect alleged.
- The clock runs much longer. The long-stop extends from ten years to as much as twenty-five for injuries that appear late. The €500 minimum threshold for property damage is removed.
And one provision deserves its own line: a court may find a product defective without establishing its actual defectiveness, where it belongs to the same production series as a piece already shown to be defective.
Nickel sensitisation is the textbook latent injury. Under a ten-year long-stop, most of the exposure window fell outside the claim period. Under twenty-five, it does not.
Why the timing provision lands on jewelry
A consumer does not develop contact dermatitis on the day of purchase. Sensitisation builds through repeated exposure, and a reaction may appear months or years after a piece was bought. That is precisely the profile the extended long-stop was written for.
Now add the second provision. Demonstrating non-compliance with EU product safety rules triggers a presumption that the product was defective. In practice, a nickel release result above the permitted limit stops being a purely administrative matter and starts shifting the burden of proof in a civil claim.
And add the third: a finding against one piece can extend to the production series it came from. For a collection produced in volume, those provisions compound rather than simply add.
The part solid-gold brands assume doesn’t apply to them
There is a widespread assumption in the trade that this is a costume jewelry problem, or a vermeil problem. It is not, and the reason is worth stating plainly.
White gold is a plated surface
Almost all white gold sold in Europe is rhodium-plated. Rhodium wears. What sits beneath it is an alloy whose composition the brand may never have verified — and a worn surface is a different article from a new one. For coated articles, the relevant assessment is made after simulated wear and corrosion, because the question is not whether a piece is safe in the box, but whether it remains safe on the wrist.
Findings and solder are not always the same alloy
Clasps, posts, springs, jump rings and solder joints frequently come from a different supplier and a different alloy than the body of the piece. In a solid gold necklace, the component in longest continuous contact with skin is often the clasp — and the clasp is the part nobody tested.
A supplier certificate answers a different question
A certificate showing a compliant result on a new sample is not evidence that the article as sold complies. It is evidence about a different question. Many brands are holding certificates that will not do the work they believe they will do — and almost none of them know it.
Where the trade stands today
Here is the honest picture, and it is more encouraging than the above may suggest. The old Directive continues to apply to products placed on the EU market before 9 December 2026. The new rules apply from that date forward. Nothing is retroactive, and nothing has gone wrong yet.
What that means is that the sector has a defined window — measured in months, not years — to move from a compliance posture built around market access to one built around evidence. The work involved is smaller than it sounds. Three gaps are common, and all three are fixable:
- Certificates that test the wrong thing. If your documentation does not address the article as worn, it does not address the requirement.
- Batches that cannot be separated. If a finding can extend across a production series, the boundaries of your production series matter enormously. A brand that cannot show which pieces came from which alloy, which plating run and which subcontractor has no way to contain a claim. Batch segregation is not housekeeping — it is the difference between a contained problem and a general one.
- Files that exist but were never read. A folder of supplier PDFs is not a technical file. If disclosure is ordered and what emerges is incoherent, the presumption of defectiveness becomes available to the claimant.
Five things worth doing before December
None of this requires testing every piece — which is neither possible nor asked for. It requires a defensible position at the level of the batch.
- Sample to your own plan, not the supplier’s. Define product families by alloy, plating specification and subcontractor, and treat any change to those as a new family requiring fresh evidence.
- Test coated surfaces as worn. Rhodium on white gold, and any plated finish, through an accredited laboratory and a method that addresses wear — not a supplier’s internal certificate on a new sample.
- Keep retention samples. One physical piece from each batch, retained. Cosmetics and food have done this for decades; jewelry almost never does. In year eight of a latent claim, a retained article is worth more than any document.
- Record the placing-on-market date for every reference. From 9 December you are running two liability regimes at once, and only accurate dating tells you which applies to which piece.
- Ask your European representative what they have actually seen. A representative who stores documentation without having examined a product cannot help you produce evidence. From December, that distinction has consequences.
The gemologist at the centre
For most of the past decade, product compliance in jewelry has been treated as a market access question — paperwork to clear customs and satisfy a platform. From 9 December it becomes a liability question, which is a different discipline with different standards of proof.
The evidence that discipline requires is not documentary. It is material. Which alloy, which plating, which finding, which stone, which batch. These are gemological questions before they are legal ones — and they cannot be answered by a service that has never handled the pieces. This is exactly the intersection where Gems & Jewelry Tech Lab works: keeping gemology at the pace of regulation, without surrendering its rigor. The brands that come through this period well will not be the ones holding the most certificates. They will be the ones who can explain, in a form a court accepts, what is in their pieces and how they know.
If you are selling jewelry into Europe, the best time to build that file is before anyone asks to see it.
The European Jewelry Desk Get in touch