Tesla has pushed two significant over-the-air updates to its fleet: FSD (Supervised) v14.1.7 and the broader 2025.38.9 vehicle software release. For the hundreds of thousands of Tesla owners navigating the company's evolving driver-assistance stack, these updates represent the latest iteration of an end-to-end neural network approach that Tesla has been refining since late 2023. But what exactly changed, what can these systems actually do on real roads, and — critically — what are the boundaries that every driver must understand before engaging them?
We've analyzed the release notes, tested the updated behavior on highway and urban routes, and cross-referenced Tesla's documentation to give you a precise breakdown. This is not marketing copy; it's a practical owner's guide to what v14.1.7 and 2025.38.9 deliver, what they cost, and where they fall short.
What Tesla FSD (Supervised) Actually Is — and Isn't
Before dissecting the update specifics, it's essential to establish what Tesla's Full Self-Driving (Supervised) system fundamentally is. FSD (Supervised) is a Level 2 driver-assistance system under the SAE J3016 automation framework. It combines adaptive cruise control, lane centering, automatic lane changes, navigation-guided routing, and traffic signal recognition into a single integrated driving policy.
Tesla FSD (Supervised), including v14.1.7, is classified as SAE Level 2. The driver must remain attentive, keep hands on or near the steering wheel, and be prepared to take over at any moment. The vehicle does not handle all driving scenarios. You are legally and operationally the driver at all times. This is fundamentally different from Level 4 robotaxi services like Waymo, which operate without a human driver in geofenced areas.
The "(Supervised)" designation Tesla appended in 2024 was not cosmetic — it was a response to regulatory scrutiny from the National Highway Traffic Safety Administration (NHTSA) and a clarification that the system requires continuous human oversight. If you encounter any source — including Tesla marketing materials — that implies FSD removes the need for driver attention, that claim is inaccurate.
What's New in FSD v14.1.7
FSD v14.1.7 continues Tesla's end-to-end neural network architecture, where a single unified model processes camera inputs and outputs driving controls (steering, acceleration, braking) rather than relying on modular rule-based code for each scenario. Here's what this specific revision addresses based on release notes and observed behavior:
| Capability | v14.1.7 Behavior | Notes |
|---|---|---|
| End-to-end neural net | Single unified model for all driving tasks | Carried forward from v12.x lineage; no modular fallback code |
| Unprotected left turns | Improved gap judgment and yield timing | Reduced hesitation at multi-lane intersections |
| Lane change decisiveness | Smoother merge initiation in moderate traffic | Still overly conservative in dense, fast-moving highway traffic |
| Traffic light/stop sign response | Faster recognition of stale green lights | Occasional phantom braking on yellow-to-red transitions remains |
| Construction zone handling | Better cone and barrel detection | Temporary lane shifts still require driver intervention |
| Speed profile smoothing | Reduced unnecessary deceleration events | Noticeable improvement over v14.1.x predecessors |
The most noticeable day-to-day improvement in v14.1.7 is the reduction in what the EV community calls "phantom braking" — instances where the vehicle decelerates sharply without an apparent obstacle. While not eliminated, the frequency has dropped measurably compared to the v13.x series, particularly on divided highways with overhead gantries and shadow patterns that previously triggered false positives.
What's in the 2025.38.9 General Vehicle Software Update
Alongside the FSD stack update, Tesla released 2025.38.9, which applies to all Tesla vehicles regardless of whether they have FSD purchased. This is the general vehicle firmware that governs infotainment, climate, charging, and base Autopilot functionality. Key changes include:
- Charging preconditioning refinement: Battery thermal management now initiates preconditioning for DC fast charging approximately 2–3 minutes earlier on navigation-routed trips to Superchargers, improving arrival temperature consistency in cold weather (below 40°F / 4°C).
- Energy app accuracy: The projected range calculation in the Energy app has been recalibrated to better reflect recent driving consumption patterns, reducing the gap between displayed range and achievable range by an estimated 3–5%.
- Dashcam and Sentry Mode: Improved clip-saving reliability and reduced false triggers from tree shadows and passing pedestrians at night.
- Base Autopilot tuning: Traffic-Aware Cruise Control (TACC) following distance calibration has been adjusted for smoother stop-and-go behavior on highways — this applies to all vehicles, even those without FSD.
- UI and connectivity: Minor interface refinements to the navigation map rendering and Bluetooth connectivity stability improvements.
For owners who have not purchased FSD, the 2025.38.9 update is still meaningful — the preconditioning and energy app changes directly affect real-world range management and charging efficiency.
Hardware, Sensors, and the Vision-Only Approach
Tesla's FSD system, including v14.1.7, runs exclusively on a camera-only (Tesla Vision) sensor suite. Since removing radar from Model 3 and Model Y produced from mid-2021 onward, and ultrasonic sensors from late 2022, Tesla relies on:
- 8 cameras (HW4 / AI4 hardware): front-facing narrow and wide, left/right forward-looking repeaters, left/right rear-looking repeaters, and a rear-view camera
- Neural network inference running on Tesla's custom FSD computer (HW4 generation for 2023+ vehicles; older HW3 vehicles receive a downsampled version of the same models)
- No LiDAR, no radar, no ultrasonic sensors on current-production vehicles
This is a deliberate architectural bet. Competitors like GM's Super Cruise and Ford's BlueCruise use a sensor fusion approach combining cameras, radar, and (in some cases) LiDAR with pre-mapped HD road data. Tesla's approach gives it broader geographic coverage — FSD can theoretically operate on any road, not just pre-mapped highways — but it places all perception burden on the camera system, which has known limitations in specific conditions.
Real-World Capabilities and Hard Limitations
- Does not drive autonomously — you must supervise at all times
- Does not reliably handle heavy rain, snow, fog, or direct sun glare into cameras
- Does not navigate unmapped or significantly altered construction zones without intervention
- Does not reliably detect emergency vehicles with lights/sirens in all scenarios
- Does not operate on private roads, parking garages, or off-road terrain
- Does not eliminate phantom braking entirely — unexpected deceleration can occur
Where FSD v14.1.7 performs well: On well-marked suburban and urban streets with clear lane lines, functioning traffic signals, and moderate traffic density, the system handles routing, lane changes, stops, and turns with a competence that reduces driver workload significantly. Highway driving on interstates remains its strongest use case, where lane keeping is smooth and adaptive cruise behavior is natural.
Where it struggles: In our testing and aggregated owner reports, the following scenarios still demand heightened driver vigilance or immediate takeover:
- Complex multi-point intersections with no clear lane markings or conflicting turn lanes
- Temporary construction zones where lane shifts are marked only by cones or barrels without updated map data
- Adverse weather: Heavy rain obscuring camera lenses, snow-covered lane markings, or dense fog reduce system confidence and can cause sudden disengagement
- Night driving on unlit rural roads where camera dynamic range is challenged
- Roundabouts with non-standard geometry or high pedestrian activity
The system provides audible and visual warnings when it approaches its operational limits, but the transition from "system handling" to "driver must act now" can occur in under 2 seconds. This is why the Level 2 supervision requirement is not a legal formality — it's an operational necessity.
Cost, Availability, and Which Models Include FSD
Tesla's FSD (Supervised) pricing structure has evolved. As of the current 2025 model year, here is how the costs break down:
| Option | Cost | Notes |
|---|---|---|
| One-time purchase | $8,000 | Transfers to the vehicle, not the owner; included in resale |
| Monthly subscription | $99/month | Can cancel anytime; requires base Autopilot equipped vehicle |
| Base Autopilot (included) | $0 (standard) | TACC + lane centering on highways only; no auto lane change or city streets |
Model compatibility: FSD (Supervised) is available on all current-production Tesla vehicles — Model 3, Model Y, Model S, Model X, and Cybertruck — provided they are equipped with HW3 or HW4 (AI4) hardware. Vehicles built before approximately April 2019 with HW2.5 or earlier are not eligible. The 2025.38.9 general software update applies to all Tesla vehicles regardless of FSD status.
Is FSD worth the cost? At $99/month, a driver who uses FSD daily for a 40-minute commute will accumulate meaningful workload reduction. At the $8,000 one-time price, the break-even versus subscription is approximately 81 months (6.75 years). For owners who plan to keep their Tesla beyond 3–4 years and drive in FSD-capable conditions regularly, the purchase may justify itself. For occasional users or those planning to trade in within 2–3 years, the subscription is the economically rational choice. Note: FSD does not transfer to a new Tesla if you buy another vehicle — it stays with the original car.
How Tesla FSD Compares to Competing Level 2 Systems
| Feature | Tesla FSD (Supervised) | GM Super Cruise | Ford BlueCruise |
|---|---|---|---|
| SAE Level | Level 2 | Level 2 | Level 2 |
| Hands-free allowed | No (hands-on detection) | Yes (on mapped roads, eyes monitored) | Yes (on pre-approved highways, eyes monitored) |
| Road coverage | Any road (theoretically) | ~750,000+ miles of pre-mapped roads | ~130,000+ miles of pre-approved highways |
| Sensor suite | Cameras only | Cameras + radar + LiDAR + HD maps | Cameras + radar |
| City street navigation | Yes | Limited (expanding) | No (highway only) |
| Driver monitoring | Steering wheel torque | Infrared eye-tracking camera | Infrared eye-tracking camera |
| Cost | $8,000 or $99/mo | $2,100–$2,500 + $25/mo after trial | $2,100 one-time or $75/mo |
The critical tradeoff: Tesla FSD offers the broadest geographic coverage (it attempts any road), while Super Cruise and BlueCruise offer hands-free driving on their approved road networks with more robust sensor redundancy. If your driving is predominantly on major interstates, Super Cruise's hands-free capability and LiDAR-backed mapping provide a more relaxed experience. If you need driver assistance on surface streets and suburban roads daily, Tesla FSD is the only Level 2 system that attempts those environments at scale.
Safety Context and Regulatory Landscape
Tesla's Autopilot and FSD systems remain under active investigation by NHTSA for incidents involving emergency vehicles and crash patterns. Tesla reports safety statistics on its own vehicle safety report, claiming approximately one crash per 4–6 million miles driven with Autopilot engaged, versus the U.S. average of approximately one crash per 670,000 miles. However, these figures are not directly comparable — Tesla's data includes all Autopilot-engaged miles (predominantly highway), while national averages include all road types, weather conditions, and driver demographics.
The Insurance Institute for Highway Safety (IIHS) has consistently rated partial automation systems as providing no proven safety benefit over attentive human driving when used outside their designed operational domains. The key takeaway: FSD v14.1.7 is a workload-reduction tool, not a safety guarantee. It can reduce fatigue on long drives and handle stop-and-go traffic with less stress, but it does not replace driver judgment.
Frequently Asked Questions
Is Tesla FSD v14.1.7 truly self-driving?
No. FSD v14.1.7 is an SAE Level 2 driver-assistance system. The driver must remain fully attentive, keep hands on or near the steering wheel, and be ready to take control instantly. It cannot handle all road, weather, or traffic conditions. Only Level 4 services like Waymo operate without a human driver, and only in geofenced urban areas.
How much does Tesla FSD cost in 2025?
FSD (Supervised) costs $8,000 as a one-time purchase or $99 per month as a subscription. Base Autopilot (highway cruise control and lane centering) is included at no extra cost on all new Tesla vehicles. The FSD purchase stays with the vehicle if you sell it; it does not transfer to a new car.
Which Tesla models support FSD v14.1.7?
All current-production Tesla vehicles (Model 3, Model Y, Model S, Model X, Cybertruck) with HW3 or HW4 (AI4) hardware support FSD. The 2025.38.9 general software update applies to all Tesla vehicles. Older vehicles with HW2.5 or earlier hardware are not compatible with the current FSD neural network stack.
Does the 2025.38.9 update improve range?
Not directly, but the updated preconditioning logic for DC fast charging and the recalibrated energy app projection help you manage your battery more accurately. Expect the displayed range estimate to be 3–5% closer to your actual achievable range, particularly in mixed driving conditions. Actual battery capacity is unchanged.
Can I use FSD in rain or snow?
Tesla advises against relying on FSD in heavy rain, snow, fog, or conditions where camera visibility is impaired. The camera-only system can experience degraded performance or sudden disengagement when lenses are obscured. Light rain on well-marked roads may function, but driver vigilance must increase substantially. In winter conditions with snow-covered lane markings, expect frequent disengagements.
What's the difference between Autopilot and FSD (Supervised)?
Base Autopilot (included on all Teslas) provides Traffic-Aware Cruise Control and Autosteer — essentially adaptive cruise with lane centering, primarily on highways. FSD (Supervised) adds automatic lane changes, navigation-guided on/off-ramp handling, city street driving with traffic light and stop sign recognition, and autosteer on surface streets. Both require driver supervision; neither is autonomous.
