How many died while TCMV deliberated?
NL approved Tesla FSD on 10 April 2026. Croatia became the eighth country to approve it on 28 September 2026, but TCMV still hasn't harmonised it EU-wide. The counter shows the net toll — after crediting measured use in the countries where it is live.
Since NL approval · – days of TCMV delay · Blocked EU-wide (net of patchwork)
The 8-country patchwork is real — people are being saved there (green). Blocked EU-wide shows what's still on the table: the net toll after crediting those savings. Even that number is only a few dozen per year — Teslas are ~0.5% of the fleet. The scary figures only appear in the whole-fleet fantasy.
Safety evidence in Tesla's thread
Supporting tests and telemetry cited alongside the new EU customer-fleet result.
Scope matters: the dashboard was assembled before the April 2026 customer rollout and mixes North American fleet telemetry with European engineering and route tests. These figures are Tesla-reported, use different denominators and baselines, and should not be combined into one rate.
Assumptions
Change any input — the link updates so you can share your settings.
The math
Blocked EU-wide
Before you @ TCMV
This is counterfactual, not a body counthonesty first
Same numbers as "lives saved by FSD" — flipped to "deaths from regulatory delay." Nobody is claiming TCMV literally killed these specific people. It's a sensitivity model: if FSD works as claimed, how many deaths happened while approval stayed fragmented?
Patchwork saves lives — blocked EU-wide is net of thataccounting
The patchwork rollout (NL, LT, EE, DK, BE, SI, CZ and, since 28 September 2026, HR) already saves lives — that's credited, not counted as delay. "Blocked EU-wide" subtracts those savings from the gross EU-Tesla figure. Same underlying math, two framings.
TCMV isn't sitting on thousands of bodiesfleet size
The entire EU Tesla fleet is ≈ 0.5% of cars on the road. Even instant EU-wide Tesla approval moves the needle only slightly — dozens per year, not thousands. The gap between patchwork and EU-wide is real but modest.
"Whole fleet" is the 32,000-lives constructionupper bound
Tesla's slide to Swedish regulators claimed 32,000 lives by assuming every vehicle becomes a safe FSD car. Researchers called it misleading. The whole-fleet scenario is the same shape — useful as an upper bound, useless as an indictment of TCMV.
4.1× is about collisions, not deathsseverity
Most collisions kill no one. Assuming deaths fall as much as collisions is unproven; fatal crashes turn on speed and road type. The factor applies most cleanly to minor collisions, less to serious injuries, least to deaths.
Highways vs where deaths actually happenroad type
Motorways account for only about 8% of EU road deaths; rural roads account for 53%. Tesla's road-class evidence reports 40–69% fewer major crashes, so using the all-collision 4.1× factor for fatalities may overstate the toll.
The safety data is Tesla's ownsource
Reuters flagged Tesla's earlier figures as potentially inflated; US senators asked NHTSA to review the method. RDW did its own testing, but the 4.1× is Tesla's telemetry, not an independent verdict on lives saved or lost.
The headline multiple keeps movingdrift
Tesla reported 3.5× for the Netherlands in June, 5.2× across 65m km in July and 4.1× across 100m+ km on 1 September. An analysis by Automotive World also found every dated measurement window in the published dashboard closes in 2025, months before the 2026 period the 4.1× covers.
US crash filings are at a record, and FSD can't be separated outus data
Tesla filed 236 driver-assist crash reports to NHTSA for July 2026, its highest month, four of them involving a death. The software field is redacted, so Autopilot and FSD cannot be told apart, and a system engaged within 30 seconds of a crash is not proof it caused the crash. These are US roads, not EU ones.
NHTSA is still investigating FSD in poor visibilityopen probe
Engineering analysis EA26002, opened in March 2026, covers about 3.2 million vehicles and names FSD Supervised explicitly. It lists nine crashes and one death and asks whether the system fails to detect and flag its own degraded state in reduced visibility. An engineering analysis is the last step before NHTSA closes a case or pushes a recall.
No EU collision has been reported publiclythe other way
Searches in English, Dutch, Danish, Slovenian and Czech turn up no publicly reported collision involving FSD Supervised in the countries where it is live. Absence of reports is not evidence of absence, and Tesla is not obliged to publish EU incidents. It still belongs in the ledger.
The per-km denominator is derived, not publishedmethod
Deaths per kilometre here use 2,500 bn EU car-kilometres a year. Official EU statistics publish passenger-kilometres, not vehicle-kilometres, so that denominator is a derivation rather than a figure you can look up. It scales the patchwork and EU-wide numbers directly.