The EU Has a New Plan to Phase Out Animal Testing. What Does It Actually Change?

If you've followed cruelty-free beauty for any length of time, the latest announcement from the European Union can sound strangely familiar.

Animal testing on finished cosmetic products was banned in the EU years ago. Restrictions on testing cosmetic ingredients followed, and the full cosmetics marketing-ban milestone arrived in 2013. So when the European Commission adopted a new roadmap on June 1, 2026 to phase out animal testing for chemical safety assessments, the obvious question was: wasn't this already supposed to be over? ^{1 2}

The answer isn't that the earlier ban was fake, and it isn't that the new roadmap ends every remaining test. The two developments deal with different parts of a much larger regulatory system.

What the EU actually announced

The official package is called the Roadmap towards phasing out animal testing for chemical safety assessments. It includes Commission Communication C(2026) 3497 and an accompanying Staff Working Document, SWD(2026) 144. ¹

The roadmap spans 15 areas of EU legislation, organizes the work under three pillars, and sets out 22 actions. It also points to more than 30 endpoint-specific recommendations, a public progress dashboard expected by the end of 2026, and a high-level stocktake in 2029. ¹

That is a real shift. The EU now has named actions, responsible institutions, reporting mechanisms, and dates that can be checked. The European Chemicals Agency has already created a collaborative platform to help move non-animal approaches into regulatory use. ³

But the legal status matters. This is a Commission roadmap, not a new regulation that instantly amends REACH, the Cosmetics Regulation, or every testing requirement across the EU. It describes how change is meant to happen and where future legal proposals may be needed.

Why the 2013 cosmetics milestone didn't end every test

The confusion starts because the Cosmetics Regulation and the EU's wider chemicals law don't answer the same safety questions.

Article 18 of the Cosmetics Regulation restricts animal testing and the marketing of cosmetics tested on animals to meet the requirements of that regulation. It was a major change in how cosmetic safety could be demonstrated, and it remains one of the strongest legal protections of its kind. ⁴

A cosmetic ingredient can also be a chemical used in manufacturing or in products outside beauty. That can bring it under REACH, which looks beyond the person applying the finished cosmetic. REACH may also consider workers handling a substance during production and risks to the environment. Vertebrate testing under REACH is supposed to be used only as a last resort, but it hasn't been removed from the system entirely. ⁵

The Commission and ECHA have previously clarified that animal testing may still be permitted under REACH for worker-safety questions, environmental endpoints, or non-cosmetic uses when other evidence isn't enough. ⁶ The General Court's 2023 Symrise judgments confirmed that the cosmetics ban doesn't automatically cancel those separate REACH obligations, even when a substance is used in cosmetics. ⁷

What non-animal testing can already do

The phrase “alternative testing” can make it sound as though scientists are waiting for one perfect replacement. In practice, replacing an animal study often means combining several forms of evidence.

That can include reconstructed human skin, cell-based tests, chemical reactivity assays, computer modelling, read-across from similar substances, exposure data, and structured weight-of-evidence assessments. You may see these grouped under the term New Approach Methodologies, or NAMs, although definitions vary and not every use of the term means a method is fully animal-free.

Some areas are already well established. OECD test guidelines support non-animal methods for skin corrosion, skin irritation, phototoxicity, and important parts of skin sensitisation. ⁸ For skin sensitisation, defined approaches can combine in chemico, in vitro, and computer-based evidence through a fixed interpretation process. The OECD says these approaches can provide the same level of information as, or more information than, the mouse Local Lymph Node Assay for hazard identification. ⁹

This matters to beauty because irritation, corrosion, phototoxicity, and sensitisation are closely connected to the kinds of local effects that cosmetic safety assessors need to understand. In these areas, the conversation isn't only about promising future technology. Regulators already have validated non-animal tools they can use in defined contexts.

Why the harder safety questions take longer

Local skin effects are only part of chemical safety. Some questions involve metabolism, long-term exposure, reproduction, development, immune effects, several organs interacting at once, or what happens after repeated doses. Those are much harder to recreate in a single laboratory model.

The Commission openly identifies repeated-dose toxicity and reproductive toxicity as difficult areas. Developmental neurotoxicity and complex environmental effects also need stronger methods, better interpretation frameworks, and wider regulatory confidence before full replacement is realistic. ¹

Even a method that reduces animal use may not yet replace an older test in every legal setting. The zebrafish embryo acute toxicity test, for example, can reduce the use of juvenile and adult fish, but European scientific guidance still notes limits involving metabolism, chemical uptake, and regulatory acceptance. ¹⁰

None of this means animal models are a flawless default. Species differences, uncertain translation to humans, time, cost, and reproducibility can all limit their value. The responsible question isn't whether old methods deserve automatic trust or new methods deserve automatic praise. It is whether the evidence is reliable enough for the safety decision being made.

Consumer safety and animal welfare shouldn't be treated as competing goals. The strongest transition is one that improves the relevance of safety science while reducing and replacing animal use wherever the evidence can support it.

What changed in June 2026, at a glance

What changes now

The roadmap starts building the machinery needed for a real transition. That includes:

• A public dashboard and progress indicators expected by the end of 2026.
• New coordination across the Commission, ECHA, EFSA, EMA, Member States, researchers, and international partners.
• Work on validation, standards, data sharing, training, computational models, and endpoint-specific recommendations.
• Reports and, where appropriate, legislative proposals for shorter-term actions by the end of 2029.

The political origin matters too. The roadmap followed the European Citizens' Initiative Save Cruelty Free Cosmetics, Commit to a Europe Without Animal Testing, which was submitted in 2023 with 1,217,916 verified statements of support. ² The Commission's 2023 response committed to developing the roadmap that has now been published. ¹¹

The Commission also reports that more than 15 million animals were used for regulatory testing in the EU from 2015 to 2023, with almost 40 percent used for chemical safety assessments. That scale helps explain why the roadmap extends far beyond the beauty industry. ¹

What doesn't change yet

The roadmap doesn't immediately remove an animal-testing requirement from REACH. It doesn't prevent a regulator from requesting an animal study where current law still permits one as a last resort. It doesn't create a final binding phase-out date, and it doesn't automatically change what a beauty brand can claim.

It also doesn't prove that every promised action will happen on schedule. The public dashboard, agency cooperation, accepted methods, changes in guidance, and any future legislation will have to be watched separately. The plan is now easier to hold accountable because its milestones are visible. That is valuable, but visibility isn't the same as completion.

Why cruelty-free certification still matters

Legislation and certification do different jobs.

EU law tells companies and regulators what is prohibited, permitted, or required under specific legal systems. A certification program can add an ongoing standard for how a brand monitors its own products and suppliers.

Cruelty Free International's Leaping Bunny program requires approved brands to use a fixed cut-off date, monitor suppliers, examine the supply chain down to ingredient-manufacturer level, and maintain an independently audited supplier-monitoring system. ¹²

That doesn't mean certification can erase an ingredient's entire scientific history. It doesn't replace legal compliance, prove environmental sustainability, or guarantee that no animal-generated data has ever existed for a substance. What it can do is create clearer present-day rules, documentation, review, and accountability beyond a brand's own unsupported wording.

Why VHUES is paying attention

VHUES Beauty is an independent Canadian brand offering vegan and cruelty-free makeup and skincare. The collection is Leaping Bunny Certified and made in North America.

That doesn't make VHUES the centre of a European regulatory story. It does make the story relevant to how we think about claims. “Cruelty-free” should be specific enough to verify, narrow enough to defend, and open to scrutiny when laws, science, or supply-chain expectations change.

The useful takeaway isn't that one certification solves the entire system. It is that legal reform, better science, supplier accountability, and independent standards can reinforce one another when each is described honestly.

Final thought

The EU's new roadmap is worth taking seriously. It moves the conversation beyond a general promise and gives the transition a structure that can be measured.

It is also incomplete. The hardest legal and scientific questions haven't disappeared, and the roadmap shouldn't be presented as though every animal test connected to chemical safety has already ended.

A more honest reading can hold both truths at once: Europe has already made meaningful progress in cosmetics, and the wider system still has further to go.

That distinction may be less dramatic than a victory headline. It is also much more useful.