Tesla’s safety data splits European regulators

When one EU country approves a car technology, others can adopt that decision without repeating the full review. That makes the single market efficient. It also means that if the original evidence is weak, the weakness travels with it.

The Dutch vehicle authority RDW granted Tesla a provisional European approval in April for its "Full Self-Driving (Supervised)" software, the first such authorization on the continent (New Mobility). Within weeks, Lithuania, Estonia, Denmark and Belgium followed. Then on Sunday, a Reuters investigation reported that Tesla had presented safety statistics to Dutch and Swedish regulators that ten of eleven independent researchers called misleading.

What FSD does (and doesn't)

Despite the name, Full Self-Driving (Supervised) is not self-driving. Under the international regulation that governs it, it is classified as Level 2 driver assistance. Think of cruise control that grew far more ambitious: the car can steer, brake and handle many maneuvers, but the human behind the wheel must watch the road and remains legally responsible (Not a FSD Tracker).

That gap between the product name and its legal classification poses a practical safety risk. The European Transport Safety Council warns that drivers may overtrust a system branded as "Full Self-Driving," creating exactly the inattention the rules are designed to prevent (ETSC).

The data problem

RDW says it spent roughly 18 months testing the system, covering 1.6 million kilometres of road data and over 4,500 closed-track scenarios, without relying on Tesla's marketing claims (Reuters via Yahoo Finance, Electrek). The problem is what happened alongside the formal review.

Tesla separately told regulators its FSD-equipped cars could travel more than seven times farther between crashes than an average US driver, and that broad deployment could have saved 32,000 lives (Autoweek, Just Auto). Researchers told Reuters these comparisons were structurally unfair: Tesla counted only airbag-deployment crashes in its fleet while comparing against broader US statistics that include less severe collisions. It also compared relatively new Teslas against an older average vehicle fleet that benefits less from modern safety features regardless of software (Electrek). After the Dutch approval, Tesla emailed the same claims to Swedish regulators (Boursorama/Reuters).

RDW has not disclosed what specific metrics it validated, nor whether it assessed the US statistics Tesla circulated (Reuters via Yahoo Finance). That opacity is where the regulatory risk sits.

One approval, many countries

Belgium shows how the single market turns a Dutch decision into a continental one. Rather than running a full review, Flanders used an EU provisional-exemption rule to recognize the Dutch approval. Three Teslas drove in Flanders for about two weeks; officials rode along, consulted with RDW, and the Flemish mobility minister signed off the same evening the administrative advice arrived (EVwire Belgium, Tweakers). Under Belgium's federal structure, that regional decision applies nationwide, effectively importing the Dutch regulatory baseline without a separate review.

A split is forming. Lithuania, Estonia, Denmark and Belgium moved quickly. Germany, France, Czechia, Italy and Sweden have not. Germany's KBA shows only testing permits (KBA). Czechia said it would not recognize the Dutch approval because it wants a coordinated EU position first (EVwire Czechia). Sweden's transport agency said it looks beyond headline figures (Carup).

EU-wide approval would require support from member states representing 55% of countries and 65% of the EU population in the Technical Committee on Motor Vehicles, which discussed the Dutch file in May and meets again on June 30 (Tesla Mag, Reuters).

What to watch

Automated assistance could genuinely reduce collisions. Tesla's own early Dutch data claims fewer crashes over millions of kilometres of FSD use (Not a Tesla App). Those figures are company-generated, not independently verified.

The single market's exemption mechanism exists so regulation does not freeze innovation off European roads. That is a legitimate goal. But the system runs on mutual trust between regulators, and trust requires transparent evidence. When safety data is partly confidential, a manufacturer's public statistics are contested by researchers, and a product's name may mislead consumers about what it actually does, that trust is under strain.

The June 30 and October committee meetings, the pace of national recognitions, and whether any regulator demands clearer answers from RDW and Tesla will determine whether Europe builds a credible framework for advanced driver assistance or discovers its approval pipeline moved faster than its evidence base.