Tesla owners are only beginning to receive FSD (Supervised) v14.3.9, but Tesla is already talking about its next major self-driving upgrade: FSD V15.
And this one could be much more than another incremental update.
Tesla VP of AI Ashok Elluswamy recently revealed that V15 is being developed around three particularly important improvements: earlier hazard prediction, faster reaction times, and significantly better safety and collision avoidance. (Not a Tesla App)
That matters because FSD 14.3.9 already introduced some of Tesla's most substantial active-safety improvements yet, including Automatic Collision Evasion and a rewritten AI compiler/runtime that Tesla says delivers 20% faster reaction time. (Not a Tesla App)
So if 14.3.9 is already reacting faster, what exactly is Tesla trying to accomplish with V15?
The answer may be less about adding another feature—and more about making the car understand dangerous situations earlier.
This weekend, Tesla's FSD saved me from an accident.
— Steve 🤙 (@SteveKBark) September 8, 2026
I was driving on a busy road with FSD activated, and a car swung out in front of me from a parking lot. I don't know if they didn't see me, but there was NOT room.
Before I could react, FSD swerved to safety. I didn't even…
Tesla Says FSD V15 Will Predict Hazards Earlier
The latest V15 comments came after a Tesla driver shared an FSD encounter involving another vehicle entering the Tesla's path.
Responding to the incident, Elluswamy said Tesla is working toward:
“Even earlier prediction of hazards, even faster reaction time and overall significantly better safety and collision avoidance…”
He identified those improvements as part of Tesla's next big upgrade, V15. (Not a Tesla App)
The distinction between detecting a hazard and predicting one is important.
A conventional collision-avoidance system may respond once another vehicle, pedestrian or obstacle has already entered a dangerous trajectory.
A more capable driving model needs to recognize the behavior leading up to that danger.
For example, it could potentially evaluate:
-
A vehicle creeping out of a blind driveway
-
A car approaching an intersection too quickly
-
A pedestrian moving toward the road
-
A vehicle beginning to drift out of its lane
-
Traffic ahead behaving unusually
The goal isn't simply to brake or steer faster.
It's to recognize that something is becoming dangerous before the situation fully develops.
FSD 14.3.9 Already Started Moving in This Direction
V15 isn't appearing out of nowhere.
Tesla's latest FSD 14.3.9 release already contains several changes focused on rare situations and proactive safety.
Tesla says the update improves responses to emergency vehicles, school buses, right-of-way violators, small animals, complex traffic lights and unusual objects extending into the vehicle's path.
The company also says it completely rewrote its AI compiler and runtime using MLIR, resulting in a 20% faster reaction time. (Not a Tesla App)
Another major addition is Automatic Collision Evasion.
Under specific conditions, the system can go beyond straight-line emergency braking and potentially steer around an imminent collision when Tesla determines braking alone may not be sufficient.
We've covered that feature separately in our Tesla Automatic Collision Evasion guide, including when it can activate, its limitations and how drivers can override it.
But V15 suggests Tesla wants to push the concept further.
Instead of asking:
“How quickly can the car react to danger?”
Tesla increasingly appears to be asking:
“How early can the car understand that danger is developing?”
V15 Is Already Being Tested in Tesla's Robotaxi Program
There is another reason V15 is particularly interesting.
It's not purely theoretical.
Tesla has previously confirmed that early versions of the V15 software are already operating within its Robotaxi fleet.
During Tesla's Q2 2026 earnings discussion, Elluswamy said V15 was being developed across roughly seven major improvement tracks, with early Robotaxi builds already incorporating around 40% of those planned improvements. (MarketBeat)
That makes Tesla's Robotaxi program something of a real-world development environment for the architecture that may eventually reach customer vehicles.
However, there's an important distinction.
The V15 software currently operating in Robotaxi vehicles should not be assumed to be identical to the eventual consumer release.
Tesla is still integrating and validating additional parts of the V15 stack.
Early Robotaxi V15 Rides Offer Some Clues
We're also starting to see anecdotal reports from people riding in Tesla's Robotaxi fleet.
One recent rider described an early V15 Robotaxi as appearing more confident and smoother than current consumer FSD v14.3.x in several situations, including following distance, lane positioning, lane changes and reacting to obstacles.
The rider also reported the vehicle avoiding a pothole and slowing early for ducks crossing the road. (Zamantika)
Those observations are interesting, but they need context.
A single Robotaxi ride isn't a controlled comparison between V14 and V15, and Tesla hasn't published enough standardized V15 performance data to conclude that these behaviors represent universal improvements.
For now, they should be viewed as early real-world observations—not proof of final V15 performance.
What Tesla Owners Actually Want V15 to Fix
Tesla owners aren't only asking for spectacular emergency maneuvers.
Many of the most common requests are much more ordinary.
In current community discussions about V15, owners are asking Tesla to improve behaviors such as unnecessary highway lane changes, remaining in the left lane too long, HOV-lane decisions, acceleration behavior and pothole avoidance. (Reddit)
That's significant.
The difference between a technically capable self-driving system and one people actually want to use every day often comes down to these smaller behaviors.
A system can successfully navigate a complicated intersection and still frustrate its driver five minutes later by repeatedly changing lanes for little benefit.
For V15, Tesla therefore has two different challenges:
Make FSD safer in rare, dangerous situations.
And:
Make it more predictable and natural during completely ordinary driving.
The second may ultimately matter just as much to owners.
Could V15 Finally Improve Pothole Avoidance?
This is one of the most interesting owner requests currently appearing around V15.
Potholes remain a particularly difficult problem for automated driving.
Avoiding one requires more than detecting an object.
The vehicle needs to determine:
Is it actually a pothole?
How deep or dangerous is it?
Can I safely move around it?
Is another vehicle beside me?
Would avoiding it create more risk than driving over it?
Reddit discussion following the V15 news quickly focused on pothole avoidance as one of the features owners want most. (Reddit)
Interestingly, an early Robotaxi V15 rider has already reported seeing a vehicle avoid a pothole that previous versions had driven over. (Zamantika)
That's encouraging, but it isn't enough evidence yet to say:
“FSD V15 adds pothole avoidance.”
Until Tesla formally documents the capability or it appears consistently across broader testing, it should remain an early observation.
What About Highway Lane Changes?
This may be an even bigger everyday issue.
Current owners continue to complain about FSD making unnecessary highway lane changes or staying in the wrong lane longer than a human driver would.
One current V15 discussion includes owners specifically asking Tesla to improve left/middle-lane camping, excessive lane changes and HOV-lane behavior. (Reddit)
This is exactly where a more capable prediction model could matter.
Good highway driving isn't simply about recognizing the vehicle immediately ahead.
It requires understanding traffic several vehicles into the future.
A human driver might see:
The left lane is moving faster now—but there's a truck 500 feet ahead, so changing lanes won't actually help.
An improved FSD model needs to make the same kind of longer-horizon judgment.
That's one area we'll be watching closely as more V15 testing becomes available.
FSD V15 vs. FSD 14.3.9
At this stage, the easiest way to understand the two versions is:
| FSD 14.3.9 | FSD V15 |
|---|---|
| Rolling out to customer vehicles | Still under development |
| Automatic Collision Evasion | Broader collision-avoidance improvements promised |
| Faster AI runtime | Even faster reactions targeted |
| Improved rare-event handling | Earlier hazard prediction targeted |
| Improved driver monitoring | Major architecture improvements still being integrated |
| Current consumer FSD generation | Next major FSD generation |
Tesla lists FSD 14.3.9 as part of software releases 2026.27.5/2026.27.6, with the broader 2026.27.6 rollout beginning September 8. (Not a Tesla App)
V15 does not yet have a confirmed general consumer release date.
That's an important distinction when interpreting headlines about the new software.
Will FSD V15 Come to HW4 Teslas?
Tesla hasn't published the complete consumer compatibility list for V15 yet.
However, the company's current FSD development makes HW4 vehicles the most important platform to watch.
FSD 14.3.9 is already associated with HW4 vehicles, while Tesla's older HW3 fleet is receiving the reduced FSD 14.2 Lite branch rather than the full 14.3.9 feature set. (Not a Tesla App)
That makes the hardware question increasingly important:
How much of Tesla's next-generation AI stack can HW3 realistically run?
Until Tesla officially announces V15 consumer compatibility, owners should be cautious about assuming every current Tesla will receive exactly the same version or feature set.
V15 Could Matter Beyond Consumer FSD
V15 is also closely tied to Tesla's Robotaxi ambitions.
Tesla has said the same V15 models being developed across its vehicle platforms are intended to work with Cybercab as well. (MarketBeat)
That creates an important development relationship:
Consumer FSD → Robotaxi → Cybercab
Instead of building three completely separate driving systems, Tesla is increasingly trying to unify the underlying AI driving model.
FSD 14.3.9 already reflects this direction: Tesla's release notes say the model has been unified across FSD, Robotaxi and Actually Smart Summon. (Not a Tesla App)
V15 appears set to push that convergence further.
Does V15 Mean Tesla Is Close to Fully Autonomous Driving?
Not necessarily.
This is where it's important to separate improved software from autonomy classification.
Tesla currently describes customer FSD as Full Self-Driving (Supervised).
Tesla's own guidance says the system requires active driver supervision and does not make the vehicle autonomous. (Tesla)
Earlier hazard prediction, better collision avoidance and faster reactions could make FSD substantially more capable.
But those improvements alone don't automatically turn a supervised consumer system into an unsupervised one.
Regulatory approval, validation, hardware capability and demonstrated safety performance all remain relevant.
For Tesla owners, the safest interpretation is simple:
V15 is the next major FSD architecture—not a confirmed switch that suddenly makes every Tesla autonomous.
What Should Tesla Owners Watch Next?
The biggest question isn't whether Tesla can produce an impressive V15 demonstration.
It's whether V15 consistently improves the situations owners encounter every day.
The things worth watching as more tests appear are:
-
Highway lane selection
-
Unnecessary lane changes
-
Pothole and road-debris avoidance
-
Complex intersections
-
Construction zones
-
Bad weather
-
Difficult lane markings
-
Pedestrian and cyclist prediction
-
Sudden cut-ins
-
Vehicles emerging from blind areas
-
Emergency collision avoidance
These mundane and unpredictable situations will tell us much more about V15 than a carefully selected demo route.
Yeslak's Take
FSD 14.3.9 feels like an important bridge.
Tesla has already started moving from reactive safety toward more proactive intervention with Automatic Collision Evasion.
V15 appears to take the next logical step:
Don't just react faster. Predict the danger earlier.
If Tesla can combine earlier hazard prediction with more natural everyday driving—fewer unnecessary lane changes, better road-surface awareness and more consistent decisions—V15 could be much more meaningful than a typical numbered software update.
But we're not calling it a breakthrough yet.
The most important evidence will come when the completed V15 stack reaches broader real-world testing and thousands of ordinary owners begin driving the same roads, intersections, construction zones and bad-weather conditions they encounter every day.
That's when we'll know whether V15 represents another impressive FSD improvement—or a genuine architectural leap.
FAQ
What is Tesla FSD V15?
FSD V15 is Tesla's next major self-driving software architecture. Tesla AI VP Ashok Elluswamy says it is being developed to provide earlier hazard prediction, faster reactions and significantly improved safety and collision avoidance. (Not a Tesla App)
Has Tesla released FSD V15?
Not as a general consumer release. Early V15 builds are already being used within Tesla's Robotaxi development, but Tesla is still integrating major parts of the architecture. (MarketBeat)
When will Tesla FSD V15 be released?
Tesla has not announced a firm general consumer release date for V15.
Is FSD V15 better than 14.3.9?
Tesla expects V15 to be a major improvement, but there isn't enough broad consumer testing yet to make a reliable real-world comparison. Early Robotaxi observations are promising but remain anecdotal.
Will HW3 get FSD V15?
Tesla has not confirmed full V15 compatibility for HW3. HW3 vehicles are already receiving FSD 14.2 Lite while HW4 vehicles receive the newer 14.3.9 branch, making hardware compatibility an important issue to watch. (Not a Tesla App)
Does FSD V15 make Tesla fully autonomous?
No such conclusion can currently be made. Tesla still describes consumer FSD as Supervised, requiring active driver attention. (Tesla)
