Tesla Cybertruck FSD Delayed to September - FSD Analysis

On the morning of August 22, Tesla began rolling out the latest FSD v12.5 update to vehicles equipped with the older HW3.0 chip.

Tesla executive Ashok Elluswamy mentioned that the FSD driver assistance system is expected to be released for the Cybertruck in September, which is a delay from the timeline previously promised by Elon Musk, who had said the update would come in August.

What happened: The company will initially launch the parking assist feature on the Cybertruck, followed by the FSD rollout. Elluswamy noted on the social media platform X that the Cybertruck is "not the easiest vehicle to park."

He added that the parking assist feature is expected to be released for the first time this weekend, with the FSD release anticipated in September. Elluswamy is Tesla's Director of Autopilot Software.

Earlier this month, a Tesla Cybertruck was spotted in Florida undergoing FSD testing, indicating that the FSD feature is soon to be available on this unique electric pickup.

Due to the relatively low production volume of the Cybertruck, Elon Musk had previously mentioned that this model had a lower priority in FSD development. However, as production capacity at the Texas factory increases, deliveries of the Cybertruck are gradually ramping up, and Tesla began offering test drives in select areas last month.

Significance: The Cybertruck is Tesla's latest model. The company began deliveries in November, and since then, several customers have requested FSD technology for the vehicle. Elon Musk has previously provided multiple timelines for the rollout of FSD on the Cybertruck.

In May of this year, the CEO stated that FSD would be available on the Cybertruck by the end of June, coinciding with the release of FSD software version 12.5. While version 12.5 started rolling out to Tesla vehicles in July, the Cybertruck did not receive access as promised. At the end of July, Musk indicated that the stainless steel truck should gain FSD access in August, but it seems even that won’t be happening this month. Musk has previously mentioned that FSD will eventually enable full autonomous driving. However, for now, it still requires active driver supervision.

What exactly is the much-discussed FSD?

What is FSD?

Tesla refers to its driver assistance features as Autopilot or FSD (Full Self-Driving), but it clarifies that these features do not enable full autonomous driving and still require active driver supervision.

FSD is the most advanced version of Tesla's Autopilot software, launched in 2020. Its features include automatic parking, automatic lane changes, and traffic navigation.

What level is FSD at now?

Firstly, SAE International is an organization that sets standards for the automotive industry (similar to ISO), such as for power and torque measurement. SAE International has also established standards for various ADAS (Advanced Driver Assistance Systems), ranging from Level 0, the most basic, to Level 5, which represents full autonomy. Here’s a breakdown of what each of the six levels means:

Level 0: ADAS functions only provide warnings and brief assistance.
Level 1: ADAS can control either acceleration (adaptive cruise control) or steering (lane centering).
Level 2: ADAS can control both acceleration (adaptive cruise control) and steering (lane centering).
Level 3: Conditional hands-free driving is possible, but the driver must take over when prompted.
Level 4: Hands-free driving under certain conditions is possible without any driver input.
Level 5: The vehicle can drive itself under any condition without requiring any driver input.

So far, the highest level achieved globally is Level 4, which is essentially what driverless taxis operate at within pre-mapped areas. Think of Waymo or Cruise's autonomous taxis, and you'll get the idea. As for Level 3, Mercedes-Benz's Drive Pilot is the first such system in the U.S. Approved for use in California and Nevada, the system can be activated on certain highways in these states, illustrating what is meant by "conditional." Tesla’s system is classified as a Level 2 system because it constantly requires your attention, and in most cases, you need to keep your hands on the steering wheel.

How FSD Works

To understand the strengths, weaknesses, and limitations of Tesla's FSD, it's essential to grasp the key features of FSD and how it’s designed to operate. This will give us a clearer picture of its advantages and where it might fall short.

Unlike most ADAS systems, Tesla's FSD isn’t as complex as you might think. Since 2021, the system has relied solely on cameras, eliminating the more sophisticated radar and sensor-based systems found in previous versions of Autopilot. This change was primarily made to reduce costs and simplify the vehicles, which has been particularly useful in recent years as Tesla has been lowering the prices of its models.

FSD uses cameras mounted on the front, sides, and rear of the vehicle to create a visual map of the surroundings, helping it decide which lane to enter or whether to stop for a pedestrian. However, relying purely on cameras means the system is not as effective at detecting or mapping the 3D world compared to systems that include ultrasonic sensors and/or LiDAR.

Given that many Tesla vehicles on the road are already equipped with FSD, the company is collecting vast amounts of data, aggregating thousands of miles and hours of driving from users of the feature. Since all this data is gathered over the internet, FSD can also inform the servers if a road is suddenly closed due to construction or if a building has undergone renovations.

Everything the system learns is applied across the thousands of Tesla vehicles equipped with FSD. When the system makes a mistake, the AI's machine learning capabilities come into play, reducing or even eliminating the chances of similar errors occurring in the future. Initially, FSD was only available to a select group of drivers with excellent driving records as part of Tesla's data collection, but as of 2024, FSD is now open to everyone, allowing the neural servers to gather an even broader range of data.