Spanish 
English 
There is no single, universally accepted Tesla crash rate. The apparent answer varies depending on whether the measurement involves crashes per mile, insurance claims per vehicle-year, fatal crashes, federally reported driver-assistance incidents, or controlled crash-test performance.
Some evidence supports the crashworthiness of certain Tesla models and the potential safety benefits of driver-assistance technology. Other evidence identifies serious crashes, methodological limitations, open federal investigations, and unanswered questions. Understanding the difference requires looking beyond a headline number.
Direct answer: Tesla crash statistics do not establish that every Tesla is safer or more dangerous than every other vehicle. Tesla, federal regulators, insurers, and crash-testing organizations measure different outcomes using different definitions and driving populations. The result depends on the vehicle model, software, road type, driver behavior, crash severity, and comparison method.
What Do Tesla Crash Rates Actually Measure?
Different studies report different Tesla crash rates because they may not be measuring the same event or using the same denominator.
Crash Counts, Crash Rates, and Fatality Rates Are Different
A report may measure any of the following:
- Total reported crashes
- Crashes per million vehicle miles traveled
- Crashes per registered vehicle year
- Insurance claims per insured vehicle year
- Fatal crashes per mile
- Driver or occupant deaths
- Claims involving property damage
- Crashes involving airbag deployment
- Incidents reported while a driver-assistance system was recently active
These figures are not interchangeable. A fatal-crash rate cannot be directly compared with a rate that includes low-speed property-damage incidents. Similarly, insurance claim frequency measures a different outcome from crashes detected through vehicle telemetry.
NHTSA’s Fatality Analysis Reporting System, or FARS, is a census of qualifying fatal traffic crashes. Its Crash Report Sampling System, or CRSS, uses a nationally representative sample of police-reported crashes ranging from property damage to fatalities. Neither system is designed exclusively to measure Tesla or FSD performance. (NHTSA)
Every Reliable Rate Needs an Exposure Measure
A raw crash count is incomplete without knowing how much driving produced that count.
Researchers should ask:
- How many vehicles were operating?
- How many miles did they travel?
- Were those miles primarily urban, rural, or highway miles?
- What weather and traffic conditions were present?
- Which driver-assistance mode was active?
- What severity threshold triggered reporting?
- Were incidents self-reported, police-reported, insurer-recorded, or detected through telemetry?
One hundred crashes over one billion miles represents a very different rate from 100 crashes over ten million miles.
How to Read a Crash Rate
Crash rate = qualifying crashes ÷ a defined measure of driving exposure
Tesla Crash Data Sources Compared
| Fuente | Measured outcome | Typical denominator | Principal strength | Principal limitation |
| Tesla Vehicle Safety Report | Telemetry-detected collisions by driving mode and road classification | Miles driven | Large volume of connected-vehicle data | Company-controlled methodology and comparison assumptions |
| NHTSA Standing General Order | Certain reported crashes involving ADS or Level 2 assistance | Generally no comparable fleet-mile denominator | Identifies serious incidents and possible defect patterns | Not designed as a direct manufacturer ranking |
| FARS | Qualifying fatal traffic crashes | Vehicles, registrations, or miles added during analysis | Nationwide fatal-crash census | Covers only crashes involving a death within the defined period |
| CRSS | National estimate of police-reported crashes | Weighted national sample | Includes multiple severity levels | Excludes crashes not reported to police |
| IIHS and HLDI | Crash-test performance and insurance-loss experience | Test conditions or insured vehicle years | Independent testing and broad claims data | Does not isolate every effect of software or driver behavior |
What Tesla’s Own Safety Reports Show
Tesla publishes a Full Self-Driving (Supervised) Vehicle Safety Report using mileage and collision telemetry from its fleet.
Tesla Reports Crashes Using Miles Between Collisions
Tesla’s current methodology separates what it calls major and minor collisions. Major collisions generally involve deployment of an airbag or another non-reversible restraint. Minor collisions meet a specified change-in-velocity threshold without that deployment.
Tesla classifies a collision as involving FSD Supervised when the system was active at any point during the five seconds before the collision. Its report compares FSD Supervised driving with manually driven Teslas that have active safety features, older Teslas without those features, and an estimated national baseline. Tesla says the collision-rate information is updated quarterly using a rolling 12-month period. (Tesla)
Miles between collisions can be converted into crashes per million miles:
- One collision every five million miles equals 0.2 collisions per million miles.
- One collision every one million miles equals one collision per million miles.
A higher number of miles between collisions means a lower reported collision frequency, but only within the report’s definitions and population.
Important Limits of Manufacturer-Reported Telemetry
Tesla’s report provides valuable large-scale information, but it should not be treated as conclusive proof of comparative safety.
Questions to ask include:
- Does the collision threshold include minor police-reported crashes?
- Are injury outcomes included?
- Are highway and non-highway miles separated fairly?
- Is FSD driving compared with similarly situated manual driving?
- Could connectivity loss or crash damage interfere with transmission?
- Have definitions or methodology changed?
- Has an independent organization reproduced the result?
Tesla acknowledges that connectivity and damaged communications systems may affect event capture. It also acknowledges that comparisons with federal databases require assumptions because the datasets use different collection methods, crash thresholds, fleet distributions, and road classifications. (Tesla)
| What Tesla’s Report Can Show | What It Cannot Establish by Itself |
| Collisions detected under Tesla’s stated telemetry rules | That every qualifying crash was caused by FSD |
| Mileage attributed to different control modes | That Tesla drivers and national drivers encountered identical conditions |
| Differences within Tesla’s connected fleet | That every unreported or low-severity crash was captured |
| Trends under the published methodology | That every Tesla model is safer than every competing model |
| Whether FSD was recently active | Legal fault, negligence, or product liability |
What NHTSA Crash Data Shows About Tesla
The National Highway Traffic Safety Administration collects information about certain crashes involving automated driving systems and Level 2 advanced driver-assistance systems.
NHTSA Collects Reports Involving Driver-Assistance Systems
Under NHTSA’s Standing General Order, manufacturers and operators must report qualifying crashes when a Level 2 driver-assistance system was active within 30 seconds before the crash and the incident involved specified consequences, such as a fatality, an airbag deployment, hospital transport, or a vulnerable road user being struck.
These reports can help NHTSA identify:
- Serious incidents associated with driver-assistance use
- Fatalities and injuries
- Airbag deployments
- Crashes involving pedestrians, cyclists, and other vulnerable road users
- Recurring crash scenarios
- Potential defects or inadequate warnings
- Issues requiring additional investigation or enforcement
The presence of Autopilot or FSD shortly before a crash does not establish that the technology caused the crash. Causation requires a closer examination of driver conduct, vehicle data, warnings, road conditions, system behavior, and other evidence. (NHTSA)
Federal Report Counts Are Not a Direct Brand Ranking
NHTSA expressly warns that Standing General Order reports are not normalized by fleet size, miles traveled, or operating conditions. Manufacturers also have different telemetry capabilities and may learn about crashes through different channels.
Additional limitations include:
- A manufacturer with better telemetry may report more incidents.
- Initial reports may be incomplete or unverified.
- One crash may generate an initial report and later updates.
- More than one entity may report the same incident.
- Some crashes may never become known to the manufacturer.
- Reporting requirements do not cover every minor collision.
For these reasons, federal incident totals generally cannot establish which automaker has the highest fleet-wide crash rate.
Investigations and Recalls Provide Context, Not a Universal Rate
NHTSA can open a preliminary evaluation, advance a matter to an engineering analysis, oversee a safety recall, and review whether a recall remedy is effective. A federal investigation signals that the agency is evaluating a potential safety concern. It is not, by itself, a final defect determination.
As of June 2026, NHTSA had an open engineering analysis involving FSD performance under reduced roadway visibility conditions. The March 2026 opening document covered an estimated 3.2 million vehicles and focused on whether the system detected degraded visibility and warned drivers with enough time to respond.
A separate preliminary evaluation opened in October 2025 was examining reported traffic-law violations while FSD was engaged, including red-light and wrong-way scenarios. That investigation covered an estimated 2.88 million vehicles when opened. These proceedings provide important safety context, but neither opening document establishes a fleet-wide crash rate or final manufacturer liability.
Does NHTSA say Tesla has the highest crash rate?
No. Federal incident reports alone generally cannot establish which manufacturer has the highest crash rate because the data may lack comparable mileage, fleet size, reporting capability, road type, and operating-condition denominators.
Tesla Crash Rate vs. Other Cars: Why the Headlines Conflict
Two studies can use real data and still reach different conclusions because they measure different populations and outcomes.
Insurance Studies Measure a Different Population
Insurance data may involve:
- Claims per insured vehicle year
- Collision claim frequency
- Injury claim frequency
- Average payment per claim
- Overall insurance loss
- Drivers who requested quotes on a particular platform
- Incidents appearing on driver records
- A limited group of states or policyholders
A study of driver records does not establish a mechanical defect rate. Likewise, a high insurance loss may reflect expensive repairs rather than a high number of crashes.
The Highway Loss Data Institute explains that insurance losses are influenced by claim frequency, repair cost, vehicle size, driver demographics, vehicle use, and the type of driver attracted to a particular model. HLDI adjusts its vehicle comparisons for factors such as age, location density, gender, marital status, state, model year, and insurance risk classification, but no statistical adjustment eliminates every possible difference. (IIHS)
Driver and Vehicle Characteristics Can Change the Result
Potential confounding factors include:
- Driver age and experience
- Household income
- Annual mileage
- Urban versus highway use
- Vehicle acceleration and performance
- Model age and redesign history
- Regional road and weather conditions
- Repair and insurance-reporting behavior
- Driver-assistance adoption
- Seat-belt use
- Speed and distraction
These variables can influence both crash frequency and crash severity without proving that the vehicle itself is defective.
A Fair Comparison Must Match Like With Like
A stronger Tesla-versus-other-cars comparison should match:
- Similar road types
- Similar vehicle classes and weights
- Similar model years
- Similar crash thresholds
- Similar weather and traffic conditions
- Equivalent mileage exposure
- Comparable driver populations
- Comparable driver-assistance use
| Why Tesla Crash Studies Disagree | What the Dataset Primarily Measures |
| Manufacturer telemetry | System status, vehicle mileage, and telemetry-triggered collision events |
| Federal incident reports | Qualifying crashes reported under regulatory criteria |
| Insurance records | Claims, losses, drivers, and insured vehicle years |
| Fatality databases | Fatal crashes or deaths under a defined federal standard |
| Crash-test ratings | Occupant protection and crash-avoidance performance in controlled tests |
Crash Frequency Is Not the Same as Crashworthiness
A vehicle can perform well in a crash test while still appearing in real-world crash reports or federal investigations.
Crash Tests Evaluate Occupant Protection and Avoidance Features
IIHS defines crashworthiness as how well a vehicle protects its occupants in a crash. Its evaluations may examine:
- Passenger-compartment integrity
- Driver and passenger injury measures
- Restraint performance
- Frontal and side-impact protection
- Pedestrian crash prevention
- Headlights
- Child-restraint access
- Crash-avoidance technology
For example, IIHS currently gives the 2025 and 2026 Tesla Model 3 good ratings in several frontal and side evaluations, while its updated moderate-overlap test is rated acceptable. Ratings depend on the particular model year, build configuration, test, and redesign. IIHS also reports good pedestrian front-crash-prevention performance for the tested 2025 Model Y. (IIHS)
Real-World Crash Frequency Depends on More Than Vehicle Structure
Real-world outcomes can also be affected by:
- Driver behavior
- Vehicle speed
- Road design
- Software use
- Weather and visibility
- Conduct by another road user
- Seat-belt use
- Crash direction
- Impact energy
- Emergency response
Crash frequency: How often qualifying crashes occur.
Crashworthiness: How well a vehicle protects occupants once a crash occurs.
A strong crash-test score is meaningful, but it does not prove that a vehicle will rarely crash. Likewise, a reported collision does not show that the vehicle provided poor occupant protection. (IIHS)
Are Teslas Safer Than Other Cars? The Most Defensible Answer
Recent Tesla models can perform well in controlled crash testing, and driver-assistance features may help prevent certain crashes when used properly. No single public dataset proves that every Tesla is safer or more dangerous than every competing vehicle. Model year, software, road type, crash definition, driver behavior, and methodology materially affect the answer.
Autopilot and FSD Supervised remain driver-assistance systems. Tesla states that these features do not make the vehicle autonomous or replace the driver. NHTSA similarly classifies Level 2 systems as assistance that requires the driver to remain fully engaged and attentive. (NHTSA)
Tesla Crash Evidence Preservation Checklist
After a serious Tesla crash, some physical and digital evidence may be changed, overwritten, repaired, transferred, or lost. The steps below are general guidance and should be adapted to the circumstances.
- Photograph the vehicles, roadway, traffic controls, debris, skid marks, weather, and visible injuries.
- Obtain the police report number and witness contact information.
- Record whether Autopilot, FSD Supervised, Traffic-Aware Cruise Control, or another assistance feature was reportedly active.
- Preserve dashcam, Sentry Mode, mobile-app, account, and available trip information.
- Avoid authorizing destructive testing, salvage disposal, or major repairs before important evidence is evaluated.
- Keep medical, towing, repair, rental, and insurance records.
- Ask an attorney or qualified expert whether event data recorder information and other vehicle logs should be preserved.
NHTSA explains that an event data recorder may capture brief pre-crash and crash information such as vehicle dynamics, system status, driver inputs, restraints, and airbag deployment. It does not necessarily contain every camera recording, software log, or category of Tesla data. Consumers should not assume they can personally access every relevant data source. (NHTSA)
What Evidence Matters After a Tesla Crash in California?
A California crash involving a Tesla may require the same evidence used in other vehicle cases, along with additional electronic and technical information.
Preserve Vehicle and Digital Evidence Quickly
Evidence may include:
- Vehicle photographs and measurements
- Police and witness information
- Traffic-camera or business-surveillance footage
- Dashcam or Sentry Mode recordings
- Event data recorder information
- Vehicle inspection results
- Software and assistance-mode information
- Driver warnings or dashboard displays
- Repair, recall, and maintenance records
- Medical documentation
- Insurance communications
A preservation request does not guarantee that every category of information exists or is accessible. The available evidence depends on the model, software, crash severity, vehicle condition, account settings, and parties that possess the data.
Liability May Extend Beyond One Driver’s Conduct
Potential issues may include:
- Negligent driving by the Tesla operator
- Negligence by another motorist
- Pedestrian, cyclist, or commercial-driver conduct
- A defective component
- Inadequate warnings
- A driver-assistance response
- Roadway or traffic-control conditions
- Negligencia comparativa
California applies comparative-fault principles, so responsibility may potentially be allocated among more than one person or entity. A product-defect allegation also requires evidence connecting the alleged defect to the crash and resulting harm. The existence of FSD data does not automatically establish manufacturer liability.
When a city, county, state agency, or other public entity may be responsible for a roadway condition, separate government-claim procedures may apply. These matters should be reviewed promptly because public-entity claims follow procedures that differ from ordinary claims against private parties.
Readers can learn more about State Law Firm’s approach to Tesla accident cases in Los Angeles, general Sherman Oaks car accident claims, and cases involving an allegedly defective vehicle or consumer product.
Frequently Asked Questions About Tesla Crash Rates
Does Tesla have a higher accident rate than other cars?
No single public dataset conclusively shows that Tesla has a higher accident rate than all other cars. Different sources measure telemetry-detected collisions, insurance claims, fatal crashes, police reports, or qualifying driver-assistance incidents. A meaningful comparison needs equivalent mileage, road types, vehicle classes, driver populations, and crash-severity definitions.
Does NHTSA’s Tesla crash count prove that Autopilot is dangerous?
No. NHTSA reports can identify incidents and possible patterns requiring investigation, but raw report counts are not normalized for mileage, fleet size, telemetry capability, or operating conditions. An Autopilot or FSD report also does not independently establish that the system caused the crash.
Does Tesla include crashes after FSD is disengaged?
Under Tesla’s current public methodology, a collision is classified as involving FSD Supervised when the system was active at any point within the five seconds before the collision. Tesla states that this is intended to capture situations where the system or driver disengages shortly before impact. (Tesla)
Are Tesla safety ratings the same as Tesla crash rates?
No. Safety ratings generally evaluate occupant protection and crash-avoidance performance under controlled test conditions. Crash rates measure how often defined events occur within a population or amount of driving exposure. A vehicle can receive strong crash-test ratings while still being involved in real-world crashes.
Can a Tesla’s data show who caused a crash?
Vehicle data may help reconstruct speed, braking, steering, restraint deployment, system status, or driver-assistance use, but it rarely answers legal fault by itself. Investigators may also need photographs, witness statements, roadway evidence, medical records, software information, and applicable traffic laws.
Is Tesla FSD actually self-driving?
FSD Supervised is not a fully autonomous consumer driving system. Tesla states that the driver must remain attentive and ready to intervene. NHTSA treats these consumer features as Level 2 driver assistance, meaning the driver remains responsible for supervising the vehicle. (NHTSA)
Injured in a Tesla Crash? Have the Evidence Reviewed Early
A case-specific review may be appropriate when a Tesla crash involves injuries, disputed fault, inconsistent driver accounts, unavailable digital information, or questions about Autopilot or FSD Supervised.
Bufete de abogados estatal helps injured Californians investigate serious motor vehicle crashes, preserve available evidence, communicate with insurers, and evaluate the parties or entities that may be responsible. The firm’s main office is in Sherman Oaks, and it represents clients throughout California. Contact State Law Firm at (877) 659-9223 for a free consultation about the circumstances of a Tesla crash. (Bufete de abogados estatal)
Every crash requires an individual review. The presence of a Tesla, Autopilot, or FSD Supervised does not automatically establish negligence, a product defect, or a right to compensation.
