Tesla recently launched FSD in China, which has led many people to wonder exactly how they did it so quickly. Tesla isn't allowed to send training data out of China, meaning that it can’t leverage the capacity of the new Cortex Supercomputer Cluster at Giga Texas.
Instead, Tesla is using their generalized model, in combination with Synthetic Training Data, to train FSD for China. Of course, Tesla also uses this same synthetic data to supplement training for North America and for training for Europe. With European FSD on the horizon, we’ll likely see more and more use of synthetic training data for a sure-fire means to handle edge cases.
Simulated Content
Tesla officially refers to the synthetic training data as “Simulated Content” throughout their patent, which is titled “Vision-Based System Training with Synthetic Content.” Let’s break it down into easier-to-understand chunks.
Vision-Only Training
As you may well know, Tesla’s approach to autonomy focuses on using Tesla Vision. That means cameras providing visual data are the primary - and really only - means of acquiring data from outside of the vehicle. They no longer use radar and only use LiDAR to ensure vision sensor accuracy during training.
Capturing all the information from around the car builds a 3D environment that the vehicle uses to plan its path and conduct its decision-making. All that data is processed to build a fairly comprehensive view of what is around the vehicle and what is predicted to be around the vehicle in the future. All of that is also tagged and characterized to help the system prioritize various decisions.
Supervised Learning Model
Tesla’s FSD training is done through a supervised learning model. That means that the training model is fed data that is already labeled, either by humans or by Tesla’s unique AI model. The objects in the images that are being fed are identified and also tagged with position, velocity, and acceleration. This information acts as a ground truth for the AI model to learn from, allowing it to recognize and interpret similar objects and situations when encountered in real-world driving.
Ground Truth Label Data
The ground truth label data is a critical portion of this supervised learning process. The labeled data provides the model with accurate information about objects and their characteristics in the images. This enables Tesla to develop FSD’s robust understanding of the environment around it while it's driving. This data is typically collected from real-world driving scenarios and is either manually or automatically annotated with data.
Generating Simulated Content
Supplementing the real-world ground truth label data, Tesla employs a simulated content system to generate synthetic training data - which is really the key portion of this patent. This system generates synthetic training data that closely resembles the labeled ground truth data from above.
Content Model Attributes and Contextual Labeling
The generation of that simulated content is guided by what Tesla calls “content model attributes,” which are essentially the key characteristics or features that are extracted from the ground truth label data. These could include things like road edges, lane lines, stationary objects, or even dynamic objects like vehicles or pedestrians.
By varying these attributes, the system can create a wide array of simulated scenarios - which means that FSD’s training program is exposed to as many unique and normal situations as possible.
In addition to the attributes, the system also incorporates contextual labeling - which involves adding labels to the simulated content to help refine it with even more detail. These labels can include things like weather conditions, time of day, or even the type of road or environment the vehicle is driving in. All this information is useful context to help develop FSD’s understanding of driving environments.
Training Data Generation
Tesla’s simulated content system generates vast amounts of training data by creating variations of the content models. These variations generally involve making tweaks to the attributes of the objects in the scene - thereby changing environmental conditions, or introducing new types of driving scenarios, like heavy traffic or construction.
Training FSD
Wrapping it all up - the combined dataset of both real-world data and simulated data is then used to train FSD. By continuously providing new sets of both types of input, Tesla can continue to refine and improve FSD further.
Why Use Simulated Content?
It might seem counterintuitive that Tesla utilizes simulated content for training their autonomous driving system when their vehicles already collect vast amounts of real-world driving data. Their vehicles drive hundreds of millions of miles a month, all across the globe - providing them access to an unfathomable amount of unique data. Well, there are a few reasons to do so.
Not a Tesla App
Cost Reduction
One of the primary advantages of using simulated content is cost reduction. By not having to collect, transmit, sort, label, and process the incoming data from the real world, Tesla can instead just create data locally.
That cuts costs for data transmission, data storage, and all the processing and labeling - whether by human or machine. That can be a fairly significant amount when you think about just how much data goes through Tesla’s servers every single day from vehicles all around the world.
Simulating Challenging Conditions
Simulated content allows Tesla to train FSD in a wide range of environmental conditions that might be rare, difficult, or even dangerous to encounter consistently in real-world driving. This can include challenging conditions like heavy rain, fog, or snow - or even nighttime driving in those conditions.
By training the system on this type of content without trying to pull it from real vehicles, Tesla can ensure that FSD remains operable and fairly robust even in more difficult scenarios in the real world.
Edge Cases & Safety
Another crucial benefit of simulated content is the ability to train FSD on edge cases. While we sometimes jokingly refer to edge cases as things like stopping for a school bus, there are real edge cases that may not be frequently encountered in real-world driving scenarios but can pose real safety risks for drivers, occupants, pedestrians, or other road users. Think of things that you could see happening but have never actually seen, like a car falling off a transport trailer or a highway sign falling down.
As such, Tesla simulates many unique edge cases, including sudden pedestrian crossing, unexpected obstacles in the road, or even erratic behavior from other drivers. All these unique simulations are fairly hard to capture regularly in the real world, which means simulating and training on them is essential to ensure safety.
Efficient and Continuous Optimization
Finally, the vast amount of diverse training data that can be generated by Tesla on demand means that they can quickly and efficiently iterate on FSD without needing to wait for real-world data. This means they can keep a continuous learning process going, ensuring that FSD is always improving bit by bit.
If you’re interested in reading more about the guts that make FSD tick, check out our entire series on FSD-related patents from Tesla here.
We’d also recommend our deep dive into Nvidia’s Cosmos - which is a training system for autonomous vehicles that primarily uses synthetic data to train machine models. It's a different take on Tesla’s FSD training cycle that primarily relies on real data, but it does have some similarities to this particular means of using simulated content.
Subscribe
Subscribe to our newsletter to stay up to date on the latest Tesla news, upcoming features and software updates.
Tesla’s Dan W Priestley attended the Advanced Clean Transportation (ACT) Expo in Anaheim, California, and provided an update on Tesla’s Semi truck program. The presentation covered several key developments on the status of Tesla’s Nevada Semi Factory, refinements to the Semi, and Tesla’s plans for charging and ramping production through 2026.
Let’s dig in and take a look at everything that was captured by the Out of Spec team at ACT Expo. The original video is embedded below if you’d like to watch it.
Semi Factory & Production Ramp
Priestley reaffirmed the timelines mentioned during Tesla’s Q4 2024 Earnings Call that Tesla will scale Semi production in 2026. To achieve this, Tesla has been actively building and expanding the Gigafactory Nevada site, specifically to support the production of the Tesla Semi. The dedicated Semi facility will have a targeted annual capacity of 50,000 Semi trucks.
Following the beginning of production, Tesla will utilize the initial trucks to integrate into its own logistics operations. This will serve as both a final real-world testing ground as well as an opportunity for Tesla to gather data internally. Tesla plans to begin subsequent customer deliveries throughout 2026 as the ramp-up continues.
Reuters also reported that Tesla is hiring over 1,000 new employees at the Semi Factory to begin the rapid ramping of the program.
Semi has already amassed 7.9 million miles with Tesla’s current testing and operational fleets, providing some real-world data and testing. Feedback for the truck has been exceptionally successful, with many drivers praising the Semi’s performance and comfort.
New Tesla Semi Features
Of course, it wouldn’t be a Tesla keynote without showing off some new things. The Semi will be available in 500-mile and 300-mile range configurations, now featuring updated mirror designs and a drop-down glass section to improve visibility and allow easier interaction with external elements—such as control panels at ports, for example.
New Electric Power Take-Off (e-PTO)
The Tesla Semi will also feature a new capability called Electric Power Take-Off, or e-PTO system. Similar to the PTO systems found on other vehicles, this will allow the Semi’s high-voltage battery to power auxiliary equipment at variable voltages. That includes being able to power things like climate-controlled reefer trailers, potentially replacing the noisy and polluting diesel generators traditionally used for this purpose.
Charging and Batteries
Out of Spec BITS/YouTube
Tesla is also working on an updated battery pack design for the final production design of the Semi. This new pack is designed to be more cost-effective to manufacture. The battery pack itself is slightly smaller than before, but the truck maintains the same level of range through efficiencies. Dan also confirmed during his keynote that the battery cells for the Semi will be sourced domestically inside the United States, helping to alleviate potential burdens due to tariffs.
On the charging front, Tesla is using MCS - the Megawatt Charging System - capable of 1.2MW - and designed specifically for Semi. The system uses the same V4 charging hardware found at Supercharger sites but focuses on that larger power output. Alongside a smaller physical footprint, Tesla will be able to configure these V4 cabinets for either dedicated Semi charging or for shared power scenarios with regular Superchargers. Tesla is also working on an integrated overnight charging product, but Tesla isn’t ready to talk about it yet.
46 Semi Charger Sites Coming
The 46 new MCS sites coming soon.
Out of Spec BITS/YouTube
Finally, Tesla has made substantial investments in a public charging network for the Semi. There are currently 46 sites in progress throughout the United States, and plans for significant expansion throughout 2026 and 2027. These sites are strategically located alongside major truck routes and within industrial areas to support long-haul and regional operations. Tesla is aiming to offer the lowest possible energy costs to operators to help incentivize adoption.
This was one of the best updates to the Tesla Semi we’ve received since its initial unveiling. It seems that the Semi will receive a big portion of Tesla’s attention in 2026, while Robotaxi and FSD Unsupervised take the stage this year.
The Tesla Semi has the potential to transform transportation even more dramatically than EVs already have, serving as a testament to Tesla’s mission to electrify the world.
Sentry Mode is an invaluable tool for owners - capable of keeping the vehicle safe and secure even when you’re not around. This is especially true in recent times, with the misguided and unfortunate incidents surrounding Tesla ownership, including damage to Tesla vehicles, showrooms, and Superchargers.
B-pillar Camera Recording and Dashcam Viewer
With the 2025 Spring Update on 2025.14, Tesla is expanding Sentry Mode’s functionality for certain vehicles with some much-needed changes. Sentry Mode and Dashcam can now record footage from the vehicle’s B-pillar cameras. These cameras are located on the side pillars of the vehicle, between the front and rear doors.
This adds two crucially needed viewpoints, making Tesla’s Sentry Mode a truly 360-degree security system. These cameras also provide the best angles for capturing license plates when parked, so they will be greatly appreciated by owners in the event of an incident.
These vehicles are also receiving an improved Dashcam Viewer, which now displays the six camera feeds along the bottom and a new grid view. It also allows users to jump back or forward in the video in 15-second increments.
However, to the disappointment of many owners, not all vehicles are receiving these updates due to the additional processing power needed.
Limited to Hardware 4 Vehicles, Ryzen Isn’t Enough
We have confirmed that Tesla is only adding the additional camera recording and improved Dashcam Viewer on hardware 4 (HW4 / AI4) vehicles. The newer hardware presumably has the additional processing power and bandwidth needed to handle recording and saving the two additional video streams during Sentry Mode and Dashcam.
For the time being, owners of HW3 vehicles are not receiving this feature. This includes all vehicles with HW3, even those with AMD Ryzen infotainment systems. If you’re not sure whether your vehicle has HW3 or HW4, you can refer to our FSD hardware guide.
While there’s no doubt that recording two additional camera streams would be more computationally intensive, we hope that Tesla adds the improved Dashcam Viewer to HW3 vehicles in a future update.
Cybertruck Also Missing Improved Sentry Mode
Surprisingly, and most confusing for many - is the fact that the Cybertruck is also not receiving the improved Dashcam Viewer and B-pillar camera recording with this update. This struck us as odd, especially since the Cybertruck is currently the only vehicle with the improved, more efficient version of Sentry Mode.
Every Cybertruck is equipped with HW4 and AMD Ryzen infotainment units, so this clearly isn’t a hardware restriction. It’s possible the more efficient Sentry Mode is playing a role here due to the infrastructure changes. However, we expect Tesla to address this in a future update and eventually release these features for the Cybertruck as well.
Given the Cybertruck’s high visibility and its status as a frequent target for both positive and negative attention, many owners hoped that the Cybertruck would be one of the vehicles to receive this feature.
Adaptive Headlights
Tesla finally started rolling out its adaptive headlights in North America. While the new Model Y already came with the feature when it was released last month, other vehicles with matrix headlights are now receiving the feature in the Spring Update.
All vehicles with matrix headlights are receiving this feature, which includes the new and old Model 3, first-gen Model Y, and the new Model S and Model X.
If you’re not sure if your vehicle includes matrix headlights, check out our guide. What’s interesting here is that older vehicles that were retrofitted with matrix headlights due to an accident or user replacement are also receiving the adaptive headlights feature.
Legacy Model S & Model X
As with most updates, the older legacy Model S and Model X are not receiving all the features included in this update. Unfortunately, some of the features, which include the Blind Spot Camera on the instrument cluster, Save Trunk Height Based on Location and Keep Accessory Power On are limited to the new Model S and X.
Legacy S and X models will receive the Alternative Trip Plans feature, Avoid Highways (Requires Intel MCU) and the Keyboard Languages feature.
These vehicles are also receiving all the features in the Minor Updates section except for the visualization showing how far the door is opened, which is exclusive to the Cybertruck. These additions include improved music search results, contact photos in the phone app, automatic connecting to hotspots, the ability to show third-party chargers, view Supercharger amenities, and various improvements to music services.
While many users will be disappointed not to receive the B-pillar camera recording and Dashcam Viewer improvements, it’s important to remember that Tesla typically does a great job at bringing features to older vehicles, at least with the Model 3 and Model Y. If a feature isn’t added, it’s usually due to a hardware limitation.