Tesla has reached another significant milestone in its pursuit of autonomous driving, reporting more than 27 billion kilometres travelled with its Full Self-Driving (Supervised) technology engaged. The growing figure reflects widespread use of the system across highways, city streets and everyday commutes. However, the achievement comes as regulators and independent safety researchers continue questioning how reliably the technology performs in challenging driving conditions.
While Tesla points to its expanding driving data and improving collision statistics as evidence of progress, federal investigations remain focused on potential safety concerns. The milestone highlights an important distinction in the autonomous-driving debate: accumulating billions of kilometres of experience is significant, but demonstrating that the technology can consistently make safe decisions remains a separate challenge.
A Major Driving Milestone Reflects Tesla’s Growing Reach
Tesla’s online Full Self-Driving safety dashboard has surpassed approximately 17 billion miles of cumulative driving, equivalent to more than 27 billion kilometres. The figure represents distance travelled while the company’s supervised driver-assistance software was engaged, rather than the total distance covered by all Tesla vehicles. Tesla has tracked FSD usage since the system’s introduction in 2020, with accumulated mileage increasing substantially as its capabilities and availability have expanded. The company’s third-quarter 2026 update showed cumulative FSD distance approaching 16 billion miles by the end of September, demonstrating how quickly the total has continued growing.
The milestone also reflects increasing commercial adoption. Tesla reported approximately 1.74 million active FSD subscriptions and purchases at the end of September 2026, compared with 1.04 million a year earlier. More than 75% of new vehicle deliveries in North America during the third quarter included FSD subscriptions, according to the company. However, Tesla’s reported subscription metric also includes customers who previously purchased the software outright. The kilometre counter provides evidence of extensive real-world use, but it does not independently establish the safety of every driving situation encountered.
Full Self-Driving Still Requires an Attentive Human Driver
Despite its ambitious name, Tesla’s Full Self-Driving (Supervised) system is classified as Level 2 driver assistance, meaning the person behind the wheel remains responsible for operating the vehicle safely. The technology can handle steering, acceleration, braking, lane changes and navigation through many intersections. It can also negotiate roundabouts, merge onto highways and follow routes through residential neighbourhoods. These capabilities allow drivers to experience extended periods when the vehicle manages most routine driving movements. Nevertheless, Tesla explicitly states that the software does not make its consumer vehicles fully autonomous and requires constant driver supervision.
That distinction becomes especially important when the system encounters unexpected circumstances. A construction worker redirecting traffic, an unusual intersection or a vehicle making an unpredictable manoeuvre may require immediate human intervention. Tesla uses an interior camera to monitor driver attentiveness and displays escalating warnings when drivers repeatedly look away from the road. The system can also disable automated assistance when warnings are ignored. The National Highway Traffic Safety Administration emphasizes that Level 2 technology does not transfer driving responsibility from the human to the vehicle. Even when the software appears to be performing confidently, the driver must remain prepared to intervene.
City Driving Provides an Important Test of the Technology
A significant portion of Tesla’s accumulated FSD experience comes from urban environments rather than controlled-access highways. Its safety dashboard has reported approximately 6.75 billion city miles, equivalent to roughly 10.9 billion kilometres. This distinction matters because urban driving involves a much broader range of potentially unpredictable situations. Highway journeys often feature consistent lane markings, fewer intersections and traffic travelling in similar directions. City streets introduce pedestrians, cyclists, delivery vehicles, turning traffic, temporary obstructions and complicated signal arrangements. A system capable of maintaining speed and position on a highway must make considerably more varied decisions when navigating a busy downtown intersection.
Tesla has also published more detailed research separating its FSD mileage across different categories of roads. Its 2025 North American data included hundreds of millions of miles on urban collector streets, arterial roads and local access routes. These environments provide opportunities to observe how the software handles situations beyond straightforward highway cruising. However, exposure to complicated conditions is not the same as consistently successful performance. Researchers still need to understand how frequently drivers intervene, which circumstances cause difficulties and whether certain road designs produce unusually high numbers of safety-related incidents. Such information becomes increasingly important as Tesla seeks to demonstrate that its technology can operate reliably across different communities.
Tesla’s Safety Statistics Point to Fewer Reported Collisions
Tesla argues that Full Self-Driving can reduce collision risk compared with conventional driving. At the time of review, its online safety report advertised approximately nine times fewer major collisions, seven times fewer minor collisions and six times fewer off-highway collisions under its reported comparisons. The company uses vehicle telemetry to identify driving conditions, calculate distance travelled and record collision events. Tesla defines major collisions primarily through airbag or other non-reversible restraint deployments, while its minor-collision category includes certain impacts that meet a specified change-in-velocity threshold without triggering those restraints. These definitions are important because they determine which incidents appear in the calculations.
Tesla also counts a collision as involving FSD when the software was active at any point during the five seconds preceding the event. This approach is intended to capture cases in which a driver disengages the system shortly before an impact. Its safety report uses rolling 12-month collision statistics, while the headline mileage counter covers accumulated use since 2020. Tesla considers these results evidence that its technology provides meaningful safety advantages. However, the figures remain company-generated statistics, and the comparisons depend on the collision definitions, vehicle populations and driving environments selected. They should therefore be understood as Tesla’s reported findings rather than definitive independent proof that FSD is safer under every condition.
Independent Researchers Question How Safety Is Measured
The broader debate surrounding Tesla’s safety claims centres on whether different driving populations are being compared fairly. In June 2026, Reuters reported that independent researchers had challenged safety statistics Tesla presented to European regulators. Their concerns included comparisons involving differently defined collision categories and the use of older vehicles in the general driving population as a benchmark against newer Teslas. Researchers warned that these differences could exaggerate the apparent safety advantage. Tesla’s more recent reporting includes comparisons with manually driven Tesla vehicles equipped with active safety features, but questions about representative driving conditions, driver behaviour and independent verification remain important.
The Insurance Institute for Highway Safety has separately found limited evidence that partial driving automation generally reduces crashes beyond the benefits already provided by conventional collision-avoidance technologies. That research addresses partial automation broadly and does not establish that Tesla’s FSD has no safety benefits. Meanwhile, NHTSA cautions that federal crash-reporting data cannot reliably compare manufacturers without accounting for differences in reporting capabilities and driving exposure. A system recording more crashes may simply have more vehicles, more kilometres travelled or better telemetry. Establishing meaningful safety improvements requires consistent measurements, comparable operating conditions and enough independent information to distinguish technological progress from statistical differences.
U.S. Regulators Are Investigating Alleged Traffic Violations
One of the most significant investigations involving Tesla’s Full Self-Driving system began in October 2025, when NHTSA opened a preliminary evaluation covering approximately 2.88 million vehicles. The agency identified 58 reports involving possible traffic-safety violations, including 14 crashes and 23 reported injuries at the time the investigation opened. Allegations included vehicles proceeding through red traffic signals, entering opposing traffic lanes and making improper movements through intersections. The investigation focused on whether FSD could unexpectedly command manoeuvres that drivers might struggle to recognize or correct quickly enough. The allegations prompted questions about both the software’s decision-making and the warnings provided to drivers.
Among the incidents identified by investigators were multiple collisions associated with an intersection in Joppa, Maryland. NHTSA reported that several FSD-equipped vehicles entered the intersection against red signals, with injuries reported in some cases. The agency stated that Tesla had taken action to address the issue at that location. Investigators also examined complaints involving vehicles crossing double-yellow lines or entering lanes intended for traffic travelling in the opposite direction. These concerns highlight why even relatively uncommon system errors deserve careful examination. The investigation itself does not establish that every allegation represents a confirmed defect, but it demonstrates why regulators remain concerned about unexpected behaviour during ordinary driving situations.
Poor Visibility Creates Another Major Safety Challenge
Tesla’s camera-based approach to driver assistance has also attracted regulatory attention. In March 2026, NHTSA escalated an earlier investigation into an engineering analysis covering approximately 3.2 million Tesla vehicles equipped with FSD. The inquiry examines whether the system adequately detects reduced roadway visibility and provides drivers with enough warning to respond. Conditions under review include strong sunlight, glare, dust and other airborne obstructions. Investigators identified nine crashes in their initial engineering-analysis record, including one fatal incident. The agency expressed concern that the software had sometimes failed to recognize impaired camera visibility or provided warnings too close to the moment of a collision.
Tesla began transitioning toward its camera-only approach in 2021, relying on visual information and software to interpret surrounding traffic conditions. The company has subsequently developed updates intended to improve how the system detects degraded visibility. However, regulators have questioned whether those changes adequately address the circumstances associated with reported crashes. The challenge extends beyond Tesla because weather, visibility and rapidly changing road conditions affect the development of automated-driving technologies throughout the industry. A vehicle that performs reliably in clear daylight must still be assessed against difficult circumstances such as fog, intense glare or obscured road markings. These conditions can become especially demanding when the system has limited time to recognize a hazard and request human intervention.
Canadian Drivers Face Their Own Practical Considerations
Full Self-Driving (Supervised) is available to eligible Tesla owners in Canada, where the company currently advertises a monthly subscription priced at $99. Availability depends on factors including vehicle hardware, software configuration and model year. However, Canadian drivers are subject to the same fundamental distinction between assistance and autonomy. Transport Canada explains that Level 2 systems can support steering and speed control but do not replace an attentive driver. The federal regulator also cautions that driver-assistance technologies may perform poorly in rain, snow and certain road environments. Ontario’s guidance similarly states that drivers remain responsible for safely operating vehicles equipped with advanced assistance features.
Canada’s climate makes these limitations particularly relevant. Snow-covered lane markings, slush-obscured cameras, temporary winter road configurations and rapidly changing visibility can complicate the interpretation of driving conditions. A familiar suburban intersection may become more difficult when snowbanks restrict sightlines or road markings disappear beneath accumulated snow. These are situations in which drivers must evaluate whether automated assistance remains appropriate. Tesla’s growing global mileage provides useful information about widespread system use, but it does not establish equivalent performance across every Canadian season or road condition. For Canadian motorists, understanding the vehicle’s limitations and maintaining control remain more important than the size of the company’s cumulative distance figures.
Robotaxi Expansion Raises the Stakes for Tesla’s Next Steps
Tesla’s ambitions extend well beyond supervised driving in privately owned vehicles. During the third quarter of 2026, the company began deploying its purpose-built Cybercab in Austin, Texas, and reported nearly two million cumulative unsupervised kilometres? No—its quarterly disclosure measured nearly two million unsupervised miles across its Robotaxi service. That distinction matters because driverless Robotaxi operations represent a different operating model from consumer FSD, where a human must remain attentive and ready to intervene. The company views its supervised driving experience as part of its broader technological development, but success with a supervised system does not automatically demonstrate that completely driverless operation is safe.
Regulatory oversight is evolving alongside those ambitions. In September 2026, NHTSA opened an investigation into Tesla’s certification of the Cybercab’s compliance with federal vehicle safety standards, particularly because the vehicle lacks conventional driver controls. As Tesla expands its autonomous-driving efforts, the most meaningful evidence of progress will involve more than growing mileage totals. Clearer reporting on collisions, system limitations, driver interventions and differences between software versions could help regulators and researchers assess real-world performance. Tesla’s 27-billion-kilometre achievement demonstrates substantial experience with assisted driving. Whether that experience ultimately translates into consistently safer and more independent transportation remains one of the industry’s most consequential questions.
