6B Miles Driven (www.tesla.com)

Tesla published a detailed methodology for its Vehicle Safety Report, describing how it counts and compares collision rates for FSD (Supervised) versus manually driven cars. Collisions are tracked per 49 C.F.R. §563.5: any event that deploys non-reversible restraints or exceeds a Delta‑V of 8 km/h within 150 milliseconds. Tesla separates “deployment” (major) and “non‑deployment” (minor) collisions, and attributes a collision to FSD (Supervised) if the system was active at any point within five seconds before impact. Data come from automated telemetry (e.g., on shift‑to‑park transmissions and VIN-associated collision packets); Tesla reports receiving 2.5 billion telemetry packages in Q3 2025 (ex‑China). Metrics are updated quarterly on a rolling 12‑month basis. The report claims materially higher miles‑per‑collision for vehicles using FSD (Supervised) versus manually driven Teslas and versus an estimated U.S. average. To build that U.S. baseline Tesla combines FHWA vehicle‑miles traveled with NHTSA sampling systems (CISS/CRSS/FARS), using its large, geographically distributed fleet to apportion highway/non‑highway rates and minor/major ratios. Tesla does not assign fault and notes limitations—connectivity, unreported crashes, sample differences and necessary assumptions can bias estimates—but argues these do not undermine the clear real‑world safety advantage shown in direct comparisons within its telemetry pipeline.