
A Tesla sign is displayed at a dealership on March 20, 2025 in Austin, Texas. Brandon Bell/Getty Images
Slovenia has joined five other European Union member states in authorizing Tesla's Full Self-Driving (Supervised) driver-assistance system on its public roads — an announcement Tesla made September 7, 2026, on X, with CEO Elon Musk amplifying the post and Slovenian Energy and Infrastructure Minister Jernej Vrtovec reposting it alongside the words "Developing Slovenia." Reuters reported the approval on September 8. What the announcement does not say, and what matters more for the bloc-wide vote Tesla is pressing for on October 6, is that all six of these national clearances derive from a single Dutch safety assessment — one that the regulator sealed at Tesla's request as a commercial secret, meaning the record the EU is being asked to ratify across 27 member states and roughly 450 million people is a file that neither those member states nor the public can read.
The approval mechanism is not complicated, but it is routinely misread by the way these announcements are covered. The Dutch vehicle authority RDW granted provisional type approval in April 2026, under UN Regulation 171 — a global standard governing driver-control assistance systems that came into force in September 2024 and which FSD only meets in part, requiring separate Article 39 exemptions for behaviors existing rules do not permit, including hands-free lane changes and certain urban navigation modes. Under EU Regulation 2018/858, other member states can then accept that certification without repeating the technical review. Lithuania took that path in May, Estonia in late May, Denmark in June, Belgium on June 10. EVSHIFT confirmed Slovenia as number six.
Not one of the five countries that followed the Netherlands conducted an independent safety review of FSD. They adopted a certification. The Dutch certification — the actual safety reasoning, the test results, the data that produced the April 10 approval — was sealed by the RDW at Tesla's request as confidential business information. The RDW General Manager, Bernd van Nieuwenhoven, acknowledged the situation plainly: "Trust us on this, we tested it extensively." No research has been released. No data explaining the decision has been made public. The Next Web reported on both the sealed Dutch approval file and two fatal US crashes linked to stopped Tesla vehicles.
This means a European driver traveling through the Netherlands, Lithuania, Estonia, Denmark, Belgium, and now Slovenia can activate FSD Supervised the entire way. The system is pre-installed as software on compatible Tesla vehicles — Hardware 4 (HW4) models sold from 2023 onward — and approval is a remote software unlock rather than a hardware change, which is why the gap between regulatory clearance and on-road operation is typically weeks, not model years. The patchwork is real and functional. It also rests entirely on one sealed file.
Tesla has said publicly that a Technical Committee on Motor Vehicles (TCMV) vote — the EU-level process that would convert these six national patchwork approvals into a single authorization across all 27 member states — is expected on October 6, 2026. That vote requires a qualified majority: at least 55% of member states and at least 65% of the bloc's total population. Six countries with a combined population of roughly 41.9 million cannot by themselves reach the population threshold. The EU's total population is approximately 450 million; 65% of that is roughly 293 million. The Netherlands (~18 million), Lithuania (~2.8 million), Estonia (~1.4 million), Denmark (~6 million), Belgium (~11.6 million), and Slovenia (~2.1 million) represent about 9% of EU population collectively.
To reach the threshold, at least one of France (68 million), Germany (84 million), or Italy (60 million) must vote yes. None has approved FSD nationally. None appears close to doing so.
What Tesla is counting on is a different dynamic: that having six countries where FSD already operates legally makes it politically harder for undecided members to vote no. The argument is momentum, not arithmetic. A country that declines to block a system its neighbors are already running faces a different kind of question than a country that declines to approve something entirely untested in Europe.
The specific technical objection that most animated regulatory holdouts earlier in 2026 was not about the system's overall performance record. It was about a deliberate product design choice: FSD allows drivers to set a "Speed Offset" that lets the vehicle exceed posted speed limits by a configurable margin. On motorways in the EU, the system is capped at 120 km/h (75 mph) regardless of the offset — but on surface streets, the feature allowed a Tesla to reach 70 km/h (43 mph) in a 50 km/h (31 mph) zone, according to French regulatory documentation.
Sweden's Transport Administration sent a formal TCMV letter on April 30, stating the committee should vote against approving FSD unless Tesla removes the ability to exceed posted speed limits. The letter stated that "allowing automated systems to systematically exceed legal speed limits ... risks undermining both the legal framework and the expected safety benefits of vehicle automation." Sweden's position, as of that letter, was a stated condition for a yes vote, not merely a reservation.
However, by September 2, 2026, Sweden's government appeared to shift. Minister for Infrastructure and Housing Andreas Carlson stated publicly that the government's "starting point and ambition" is that EU-wide approval of advanced driver-assistance systems such as Tesla FSD Supervised "should be put in place" — a notable divergence from the April 30 Transport Administration letter. Whether the Speed Offset concern has been resolved, addressed through technical modifications, or simply deprioritized at the political level is not yet clear.
French Transport Minister Philippe Tabarot said on July 22, 2026, that safety trade-offs were insufficient to authorize FSD as it currently stands, naming the speed issue and adding concerns about driver attention monitoring in urban environments, particularly at roundabouts and lane changes. France has since conducted on-road tests with two FSD-equipped vehicles under what Tesla described as a "constructive exchange" with Musk. Whether testing has altered France's official position remains unclear as of publication.
One week before Slovenia's approval, Tesla published a safety dataset from 100 million-plus kilometers of FSD Supervised driving by the roughly 70,000 European customers who have activated the system since April. The headline claim: FSD Supervised is 4.1 times less likely to be involved in a crash than a Tesla driven manually. By road type, the company reported 45% fewer major crashes on highways, 69% fewer on arterial roads, 50% fewer in cities, and 40% fewer on local roads.
The limitation is structural: the data is Tesla's own. A Reuters investigation published June 15, 2026, documented that Tesla had previously provided European regulators — including Sweden's transport authority — with safety statistics that independent traffic-safety researchers described as deeply misleading. One slide Tesla shared with Swedish officials claimed FSD could have prevented 32,000 deaths and 1.9 million injuries — a figure, researchers told Reuters, that assumed every vehicle in the United States, including freight trucks and motorcycles, would be replaced by an FSD-enabled Tesla, and that each such Tesla would be at least seven times safer than the vehicle it replaced. Tesla also compared crash rates counting only airbag-deployment incidents in its own fleet against broad US crash statistics that include far less serious accidents — an apples-to-oranges comparison that, the researchers said, manufactured an impression of safety the underlying data did not support.
Dudley Curtis of the European Transport Safety Council told Reuters his organization was "certainly concerned" that Tesla had presented "unreliable safety data" to government officials whose job is to determine whether FSD is safe to widen on European roads.
The approval that produced these six clearances is provisional. It runs 36 months from April 2026 and can be withdrawn. It was granted under an exemption mechanism designed for technologies that do not yet fit existing regulations, not under the standard type-approval process. The driver behind the wheel in any of the six countries remains legally responsible for every maneuver the system makes — SAE standard J3016 Level 2 driver-assistance does not transfer liability from the human operator to the technology. A car with FSD Supervised active is not a self-driving car in any regulatory, legal, or practical sense.
Two features of the system's design are particularly relevant for drivers to understand. First, the speed-offset behavior is active by default and configurable; drivers who do not actively reduce the offset can find the car exceeding posted limits in urban areas. Second, when FSD detects that a driver has stopped paying attention — repeated ignored attention warnings — the system performs a "controlled stop": it parks the vehicle in its current position in the roadway. In a live freeway lane, a controlled stop produces a stationary car in moving traffic.
In September 2026, The Next Web and Electrek matched two fatal US freeway crashes to redacted Tesla incident reports: a 2020 Model 3 that stopped in the center lane of I-4 in Seminole County, Florida, on October 6, 2025, before being struck by a semi-trailer; and a 2020 Model 3 that stopped on the eastbound Loop 202 in Mesa, Arizona, on October 31, 2025, and was struck from behind by a Ford F-350. Both reports confirmed a driver assistance system was engaged and the car was stopped at 0 mph in a live lane at the time of impact. Both drivers died. In both cases, Tesla redacted the crash narrative, the software version, and the field recording whether the road was inside the system's approved operating area — all withheld as confidential business information, making it impossible to confirm from outside which system was active.
The Dutch regulator that approved FSD in April sealed its own safety reasoning under the same commercial-confidentiality arrangement Tesla requested. Regulators voting on October 6 will weigh a company-produced evidence dashboard against a public record that is thinner than anyone involved would prefer.
Finland's transport authority, Traficom, said in June it could approve ahead of the EU-wide vote if it received additional information from Tesla on specific assessment issues — including how quickly drivers can resume control and how the system behaves in the kind of low-visibility winter conditions common across Nordic markets. Greece's transport ministry said a pending bill would create a legal pathway for national authorization. Both countries moving before October 6 would add to Tesla's momentum argument without resolving the qualified-majority population arithmetic.
For Tesla owners in approved countries — or in countries that may approve before the end of September — the practical change is a toggle in the car's settings and a monthly subscription of €99 (approximately $114 USD at current exchange rates). Enhanced Autopilot holders qualify for a €49 (approximately $57 USD) discounted rate. (Exchange rate conversions are based on an ECB reference rate of approximately $1.155 per euro as of September 2026 and are approximate.)
For everyone else in Europe, the question the October 6 vote is actually asking is not whether FSD is safe — it is whether the sealed record of one Dutch agency testing a system over 18 months and 1.6 million kilometers (994,000 miles) of European roads is sufficient grounds for 27 countries to open their roads to it. Six of them have already answered yes, without reading the file.
It means six countries have allowed Tesla to unlock FSD Supervised software on compatible Hardware 4 Tesla vehicles via a remote over-the-air update. In five of those six cases, no independent safety review occurred: those countries accepted the type approval issued by the Dutch vehicle authority RDW in April 2026 through the EU's mutual recognition mechanism. The Dutch safety assessment — the actual test data and reasoning behind the April decision — was sealed as commercial confidential information at Tesla's request and is not publicly available.
The SAE defines six levels of driving automation. Level 2 means the car can simultaneously control both steering and speed, but the human driver must supervise at all times and is legally responsible for every action the vehicle takes. Level 3 would shift the monitoring obligation to the system itself. FSD Supervised is Level 2, regardless of how Tesla's product name describes it. If a driver allows FSD to make a lane change that results in a collision, the driver is at fault — not Tesla.
A positive vote at the Technical Committee on Motor Vehicles would convert the patchwork of six national approvals into a single EU-wide authorization, allowing Tesla to roll out FSD across all 27 member states at once. The threshold is a qualified majority: 55% of member states and 65% of the EU's total population. The six currently approved countries hold roughly 9% of EU population collectively — France, Germany, or Italy must vote yes for the threshold to be met. If the October 6 vote fails or is delayed, individual member states could still grant national approvals one by one, and Tesla has indicated it expects to continue that country-by-country strategy in parallel.
Sweden's transport authority formally recommended a no vote unless Tesla removes its Speed Offset feature, which allows the car to exceed posted speed limits — though Sweden's government subsequently signaled support for EU-wide approval in September 2026, introducing uncertainty about Sweden's final position. France has cited the speed concern alongside worries about driver attention monitoring in complex urban scenarios. More broadly, several EU regulators have expressed unease that they are being asked to rely on a safety assessment conducted by one national authority whose reasoning is sealed from view, supported by safety statistics that independent researchers have described as methodologically flawed. The RDW has not released the data underlying its April 2026 approval decision.
