Tesla Promises Big Safety Gains With FSD v15 — Here's What the Data Actually Shows
Share
Tesla is lining up its next major Full Self-Driving release, and the headline feature isn't a new UI or a faster route — it's crash avoidance. Tesla AI lead Ashok Elluswamy said this week that FSD v15 will deliver "even earlier prediction of hazards, even faster reaction time and overall significantly better safety and collision avoidance." For owners who use FSD (Supervised) on a daily commute, that's the most consequential kind of over-the-air update there is: one that changes how your car behaves in the half-second before something goes wrong.
What happened
The comment came on September 11, 2026, in reply to a Tesla owner whose car had swerved away from another vehicle pulling out of a parking lot. Elluswamy said he was glad the owner was safe, then previewed v15 as the next big upgrade in prediction and collision avoidance. Internally, Tesla describes v15 as a larger architectural step rather than a routine point release, with a much bigger neural network and tighter coupling between the prediction and control halves of the driving stack.
Some of that work is already on the road. Tesla has begun running early v15 software in parts of its robotaxi operations, and it has been pushing safety features like Automatic Collision Evasion into current customer cars — a system that lets the driving stack intervene even when the driver is in manual control.
Why it matters
Tesla leans on published telemetry to make its safety case, and the numbers are worth understanding in full rather than in headline form. In recent North American Vehicle Safety Report data, vehicles with FSD (Supervised) engaged traveled roughly 5.1 million to 5.7 million miles between major collisions, defined as airbag-deployment events. Tesla's estimate of the U.S. average over the same period is about 699,000 miles per comparable crash — the comparison Tesla often frames as roughly seven times fewer major collisions.
The tighter comparison is the one that uses the same fleet. Teslas driven manually, with active safety features like automatic emergency braking still running, recorded a major collision about every 2.1 million miles. Measured against that baseline, FSD's advantage narrows to roughly 2.4 to 2.7 times fewer severe crashes — the figure independent researchers argue is the more apples-to-apples number. European data released in 2026 pointed the same direction: Tesla reported FSD as 3.5 times safer than manual driving in the Netherlands, and 4.1 times fewer collisions than manually driven Teslas with active safety across more than 100 million kilometers in five approved countries.
None of this settles the debate. NHTSA's Standing General Order still shows Tesla accounting for the large majority of U.S. Level 2 driver-assist crash reports, in part because the fleet simply logs far more assisted miles than its rivals. Critics also point out that Tesla's "U.S. average" mixes crash definitions and driving mix. The honest read: the same-car comparisons, plus lower rates of automatic emergency braking events and harsh maneuvers when FSD is engaged, are the evidence Elluswamy is pointing at — and the claim isn't that software eliminates risk, only that each major version widens the gap against an unaided human driver.
"Even earlier prediction of hazards, even faster reaction time and overall significantly better safety and collision avoidance coming as part of the next big upgrade (v15)." — Ashok Elluswamy, Tesla AI lead, September 11, 2026
What this means for EV owners
The practical takeaway is that your car's safety envelope now improves on a software cadence, not a model-year cadence — but only the parts Tesla controls. The hardware side is still yours to maintain, and it matters more than most owners realize. Every prediction and evasive maneuver the stack plans has to be executed through your tires and brakes, and an underinflated tire quietly lengthens stopping distance and reduces the grip available for exactly the kind of swerve that started this whole conversation. A set like the Bluetooth TPMS sensors for Model 3/Y/S/X gives you per-wheel pressure at a glance instead of waiting for a dash warning that only fires once you're already low.
It's also a good moment to think about your own record of what happened. Tesla's near-miss clips get shared because owners had footage — Sentry Mode and dashcam recording need reliable storage, and a dedicated USB port beats juggling a thumb drive in the console. Cybertruck owners can tidy that up with the 4-in-1 glove box USB hub with a dedicated dashcam port. And if you spend a lot of time watching the FSD visualization, a center console screen protector for Model 3 and Model Y keeps glare and fingerprints off the display you're actually using to supervise the drive.
Not sure what fits your car? Start at the shop-by-vehicle collection and filter to your exact trim — Highland, Juniper, Cybertruck, or classic Model 3 and Model Y all take different parts.
The bottom line
FSD v15 is shaping up to be a genuine architectural step rather than a tweak, with earlier hazard prediction and faster reaction time as the stated goals. Tesla's own data supports a real safety advantage — closer to 2.4x to 2.7x on the fairest same-fleet comparison than the 7x figure that circulates most widely — and that gap is what v15 is meant to widen. Keep supervising, keep your tires and cameras in good shape, and watch your Tesla app for the release notes. The best safety upgrade is still the one that combines good software with a well-maintained car.