Right to Repair Is Law Now. That Was the Easy Part.

After a decade of campaigning, the right to repair has largely won the legislative argument. The EU has a directive. Several US states have statutes. Manufacturers who spent years insisting repair was a safety hazard now run their own parts programmes.
Which makes this a good moment to ask a less celebratory question: what has actually changed for someone trying to get a device fixed?
The answer is uneven, and the most useful evidence comes from the jurisdiction that has been at this longest.
What the EU directive requires
Directive (EU) 2024/1799 entered into force in July 2024. Its core obligations:
- Manufacturers must repair covered goods on request at a reasonable price and within a reasonable time, even after the legal guarantee has expired
- Spare parts and tools must be available at prices that do not deter repair, with indicative repair costs published publicly
- A standardised European repair information form must be provided on request, stating the defect, estimated cost and timeframe
- Choosing repair over replacement earns a 12-month warranty extension
- Contractual terms and technical measures — software locks, parts pairing, diagnostic restrictions — that impede independent or self repair are prohibited
The product scope is an exhaustive list in Annex II: smartphones, washing machines, dishwashers, refrigerators, vacuum cleaners, displays, tablets, servers. It grows as new ecodesign regulations are adopted, but it is a list, not a general principle.
That last bullet contains the most consequential provision and the most consequential caveat. Parts pairing is banned — unless the manufacturer can demonstrate objective justification relating to safety or intellectual property. How narrowly national regulators read that exception will determine most of the directive’s practical effect.
The deadline passed. Transposition didn’t.
Member states had until 31 July 2026 to write the directive into national law. That deadline is now ten weeks past, and according to Right to Repair Europe only a handful had formally notified the Commission that they had finished.
This matters more than a missed administrative date. A directive is not directly binding on manufacturers; it obliges member states to legislate. Until a country transposes, consumers there have no new enforceable right. Germany’s Reparaturgesetz cleared the Bundestag about five weeks before the deadline. Many others did not.
And because the directive leaves penalties to member states — requiring only that they be “effective, proportionate and dissuasive” — the same European right will carry very different consequences for ignoring it depending on which of 27 national laws applies.
What happened to New York
The United States has no federal statute. It has a growing patchwork of state laws, and the first one is the cautionary tale.
New York’s Digital Fair Repair Act passed the legislature in 2022 as a strong bill. It was signed as a considerably weaker one. Following lobbying by the industry trade group TechNet, amendments added before signature:
- Let manufacturers supply pre-assembled components rather than individual parts — at the manufacturer’s discretion
- Removed any obligation to provide passwords, security codes or materials needed to get past security features
- Excluded enterprise electronics of the kind used by schools, hospitals and data centres
- Applied only to newly manufactured devices, not to products consumers already owned
Repair advocate Louis Rossmann called the result functionally useless. That is a campaigner’s verdict, not a neutral one — but the carve-outs are real and they are in the statute.
The more interesting observation came from iFixit, which noted that several large manufacturers — Apple, Google, Valve, Samsung — launched parts programmes partly to comply. Even a weakened law changed behaviour, because running one parts programme is cheaper than running fifty variants. That dynamic is the strongest argument for state-level legislation in a federal system.
Parts pairing is the actual battleground
Spare-part availability is close to meaningless if the device rejects the part once fitted. Parts pairing — serialising components so software disables features, or displays warnings, when a non-authorised part is installed — converts a hardware repair into a permission problem.
Oregon was the first US state to ban it outright. Colorado followed, with its law prohibiting manufacturers from reducing functionality or showing misleading warnings about unidentified parts, and treating violations as a deceptive trade practice — which makes them privately actionable, with civil penalties reported up to $20,000 per violation.
Reported effective and enforcement dates vary between sources, so check the statute before relying on a particular one. The direction, though, is clear: the newer laws target software gates, while the earlier New York and California statutes are largely silent on them.
Apple lobbied against the bans in both states, then began opening up its parts pairing system. Both facts are worth holding together.
Enforcement has barely started
This is where the celebration gets ahead of the evidence.
Oregon’s penalty regime — reported at $1,000 per day per violation — is not scheduled to begin until mid-2027. As of recent reporting, the states with laws on the books had seen essentially no enforcement actions, despite no shortage of apparent non-compliance.
The first known suit under Colorado’s law was filed in September 2026 by a repair business, Acme Revival, alleging it could not complete repairs because a manufacturer refused to supply the necessary materials. Note the shape of that: a private business suing, not a state agency enforcing.
A law that is never enforced is a disclosure requirement with optional compliance.
The breakthrough came from antitrust, not repair law
The most substantial win for repairers in years did not arrive through a right to repair statute at all.
The FTC and five states sued Deere & Company, alleging it had unlawfully maintained monopoly power by restricting full repair software to its authorised dealers. The case settled in July 2026. Its terms give farmers and independent repair providers the same repair resources and software capabilities as authorised dealers for ten years, require future tools to be released to independents once deployed to more than half the dealer network, and bar retaliation against those who use them. Deere separately settled a farmers’ class action for $99 million in April 2026, with no finding of wrongdoing.
Competition law delivered, within one sector, something more concrete than most repair statutes have. That is a strategic observation worth noting: where repair restriction functions as a monopoly on servicing, antitrust may have sharper teeth than consumer-protection drafting.
The uncomfortable evidence from France
France has run a repairability index since January 2021 — a 0–10 score displayed at point of sale. It is the only such scheme with enough history to evaluate, which makes it the closest thing to a natural experiment.
The French government’s own impact evaluation found average scores rose between 2020 and 2022: roughly 3% for computers, 4% for front-loading washing machines, 9% for televisions and 12% for smartphones. Encouraging — except the evaluation explicitly describes this as a descriptive trend, not a causal effect. The rise could equally reflect newer models scoring higher, or low-scoring models leaving shelves faster. On the question that matters most — whether consumers actually bought more repairable products because of the index — the effect was positive but not statistically significant.
Then there is how the scores are produced. Manufacturers calculate their own index using a grid supplied by the ministry. Oversight of those calculations is minimal.
The advocacy group HOP tested six products against their declared scores. All but one scored lower than claimed, with three off by 1.3 to 1.5 points out of 10. HOP also reported that documentation and spare-part availability appeared overrated relative to what it could actually obtain. Six products is a small sample from a campaigning organisation, and should be read as such. But a wider HOP analysis of more than 700 devices found most scores clustered above 5 — the bottom half of the scale going essentially unused, which is what you would expect from self-assessment.
A structural criticism compounds it: the five scoring criteria carry equal weight, so a phone with parts glued or welded in place can still post a respectable overall grade by scoring well on documentation and parts pricing.
France is now replacing the repairability index with a broader durability index that adds reliability criteria. No published evaluation of the replacement exists yet.
Why this is the test case for everything else
The Digital Product Passport leaves one question conspicuously open: who verifies the data? France has been running that experiment for five years, and the result is not reassuring. Where a score is self-declared and unaudited, it drifts upward, the scale compresses toward the top, and independent testing finds the published figure flattering.
This is the same pattern as chemical recycling’s mass balance accounting: a number that is officially recorded, widely cited, and not independently checked.
The lesson is not that transparency rules fail. France’s scores are better than no scores, and manufacturers did launch parts programmes in response to even weak legislation. The lesson is that the verification provision is the law. Everything else is formatting.
What to watch
- Mid-2027 — Oregon’s penalties begin. The first real enforcement test of a parts-pairing ban.
- Colorado’s private suits. Deceptive-trade-practice framing may make litigants more effective than regulators.
- National penalty regimes as EU member states finish transposing. The directive’s practical force will be set by whichever country enforces it hardest.
- Whether any jurisdiction audits a declared repairability score. On current evidence, none systematically does.
Winning the legislative argument took ten years. Finding out whether it mattered will take longer, and requires somebody to check.
Sources: Council of the EU, Right to repair products; Right to Repair Europe; US Federal Trade Commission, Deere settlement; French government impact evaluation of the repairability index; HOP, Rapport indice de reparabilite.