Tesla’s FSD V13 Pushes HW4 Hardware Capabilities; End of Line for HW3? (2025)

As Tesla continues to push the capabilities of its hardware as it prepares for autonomy with FSD Unsupervised, there has been a distinct divergence in features between HW4-equipped vehicles on FSD V13, and HW3-equipped vehicles on FSD V12.6.

There are definitely reasons for this, and Tesla hackerGreentheonly,over on X, went into great detail on some of the reasons why this divergence has continued and is likely to continue.

If you’re interested in some background, we recommend the following articles. Although they’re not necessary, you may enjoy reading them before diving right in if you enjoy this topic.

• How FSD Works - Part 1

• How FSD Works - Part 2

• How Tesla Optimizes FSD

• Tesla’s Universal AI Translator

Back in the summer of 2024, we reported that Tesla’s HW4 had yet to reach its full potential, and that HW3 still had some compute to work with. However, FSD has undergone a rapid evolution - there are now fundamental differences driven by the hardware capabilities and sheer scale of the latest builds of FSD.

FSD V13’s Gargantuan Neural Networks

FSD V13 is a substantial leap for Tesla. It features substantially larger neural networks - the core brain that processes all the environmental data and makes decisions. Those networks are far larger in V13 than in V12, which results in a much smoother and human-like expereience.

Tesla’s FSD computers contain two nodes. According to Green’s analysis of FSD V13, Node B, which handles the actual end-to-end driving logic, tripled in size from 2.3 GB in V12 to a staggering 7.5 GB in V13. That’s a substantial increase, and one of the main reasons why HW3 can’t handle FSD V13. While Interestingly, Node A on HW4 remains similar in size to FSD V12, using up about 2.3 GB.

However, this massive size is also pushing the limits of HW4. While this isn’t necessarily a big problem immediately - at the end of the day, there is a limited amount of compute and ram available and model sizes will likely grow in the future. In fact, Tesla is planning to scale model sizes in the next FSD version and has already said they’ll need to optimize them, as context size is limited by on-board memory.

HW3 Limitations

Since HW3 can’t run the latest FSD models, where does this leave millions of Tesla vehicles equipped with the now-legacy FSD hardware? Are they hitting a dead end? Not exactly - Tesla has promised to replace HW3 computers with new units during the Q4 2024 Earnings Call, but only for those who have purchased FSD and only if Tesla is no longer able to update HW3 vehicles.

There’s no doubt Tesla hopes to achieve autonomy and then work backwards to make it compatible with HW3, but that’s unlikely to happen with the limited ram available on HW3.

Tesla’s AI team has continued to say that they’re still working on updating HW3 vehicles - but that they will be taking a backburner to updates for HW4 until the team can optimize the model efficiently enough to run on the smaller footprint. That means V12.6 isn’t just V13 “lite” - it is a distinctly different software branch tailored specifically for HW3’s capabilities.

Green's analysis showed that HW3 under V12.6 runs NNs totaling 1.2 GB on Node A and 3.1 GB on Node B, which is significantly smaller than the footprint of V13. Back in November, we discussed how Tesla is running a smaller and optimized FSD model on HW3, with the specific goal of achieving similar performance to V13. FSD V12.6 appears to be the current implementation of that smaller model strategy.

Green also found that of the total neural nets used for FSD, 135 are shared between the current V12.6 and FSD V13.2 releases, which is the result of Tesla’s AI team working to optimize and refine the V13 model for HW3. So some neural nets are shared between the two sets of hardware, while others are specific to HW3 or HW4.

AI4 in the Future

The fact that V13 is already stretching HW4’s limits makes us wonder about Tesla’s imminent release of AI5. Last year, Tesla mentioned that AI5 would be 18 months away - which is now about 8 months away, and potentially even sooner to help solve the autonomy problem.

At the same time, Tesla is looking to launchUnsupervised FSD in Junein Austin, Texas, with presumably HW4, which would indicate that they’re confident in its capabilities.

Unless Tesla launches the Robotaxi network with AI5, we don’t believe AI5 will be a “must have” item. Undoubtedly, it will result in similar improvements to HW4 compared to HW3. However, the software needed to take advantage of the newer hardware could still be years away.

All eyes will be on Tesla’s Robotaxi network launch and the hardware required for it. For now, HW4 owners are in good hands with future updates are on the horizon, and we’re certainly excited for what’s next.

Following customer requests, Tesla has once again returned FSD Transfers for North America. This comes after some new details we received on FSD Supervised and Unsupervised at the recent Q1 2025 Earnings Call.

Vox Populi, Vox Dei … FSD Transfer is back

All countries (in NA), all S3XY + @cybertruck (excl Foundation Series & Launch Series)

— Tesla North America (@tesla_na) April 25, 2025

FSD Transfers

This latest round of FSD transfers doesn’t have a specific time limit - so if you’re in the market for a new vehicle and were worried about transferring FSD from a HW3 vehicle - now is the best chance.

This transfer opportunity is applicable for any new vehicle purchases, including the Model S, 3, X, Y, as well as the Cybertruck. The only vehicles it doesn’t apply to are the Foundation-Series Cybertruck or a Launch-Series Model Y, which already come with FSD included.

If you’re conducting an FSD transfer, you’ll need to reach out to your Tesla delivery coordinator through the Tesla app to confirm you’re eligible and to get the process started.

Note that once you take delivery, FSD access is removed from your older vehicle, regardless if you plan to trade in your vehicle or keep it.

Canada and Mexico

For those in Canada and Mexico, there’s some good news. This opportunity is available for customers in those countries as well, however, new vehicle orders are currently paused in Canada and Mexico for unknown reasons.

Running Offers

There are also several other offers running in North America, with the country flags indicating which countries their available in:

0% Financing on new Model 3 Orders (60mo, 0.99% for 72mo) 🇺🇸

Deep Blue Metallic and Pearl White are currently free on a new Model 3 Performance 🇺🇸

Original Model Y Owners are receiving $2,000 off the purchase of a Refreshed Model Y 🇺🇸 🇨🇦 🇲🇽

Free Supercharging on the Model S and Model X 🇺🇸 🇨🇦 🇲🇽

Free Supercharging on the Foundation-Series Cybertruck 🇺🇸 🇨🇦 🇲🇽

Inventory Vehicle Discounts 🇺🇸 🇨🇦 🇲🇽

Tesla’s long-awaited ambitions for a fully autonomous Robotaxi network powered by the futuristic, control-less Cybercab have finally started to take shape.

During their Q1 2025 earnings call, Tesla shared new details about its Robotaxi network, including that the network will start with new Model Ys. The Cybercab will join the network in 2026.

Robotaxi Plans

Tesla confirmed during the Earnings Call that it is on track to conduct the pilot launch of its Robotaxi network in Austin in June 2025. The first vehicles deployed for this service won’t be Cybercabs, as production for those is still in the prototyping stage.

However, Tesla did outline part of their plan for the Robotaxi network rollout. The initial aim is to begin driverless operations in Austin and then expand the service to other cities within the United States by the end of 2025. Interestingly, Tesla is also testing in California - albeit with safety drivers at this time.

Looking further ahead, there are ambitious plans to ramp up operations to have millions of vehicles operating autonomously within the network by the end of 2026. While substantial revenue generation won’t be immediate, Tesla anticipates that the Robotaxi network will start to “meaningfully move the financial needle” in the second half of 2026.

Tesla also highlighted the current real-world benefits that FSD is already offering today—and will be able to offer to many more people in the near future.

These include reducing effort and time spent on daily commuting or transit, as well as improving mobility for customers with disabilities. Tesla plans to emphasize some of these community stories and wants to highlight the advantages and capabilities of FSD to improve people’s lives.

Robotaxi App and Robotaxi Vehicle UI

After the call, Tesla AI also shared a video showcasing the Robotaxi interface in the Tesla app, as well as some footage that included the in-vehicle UI, at least as it currently stands.

The video shows Tesla employees using and testing Robotaxi features, although there are safety drivers in place. It looks like there is already a set of well-integrated UI mechanisms for both the front and rear screen when in Robotaxi mode.

Tesla is testing this today, and this network has already completed 1,500 trips, totaling approximately 15,000 miles. The system is being tested for vehicle allocation, mission control, and remote assistance operations.

Remote Operators

Inevitably, edge cases or unique situations will come up, and they will be difficult to handle. For these cases, where a Robotaxi vehicle gets stuck or requires assistance, Tesla will likely implement remote support.

While they didn’t completely confirm remote support operators on the Earnings Call, Tesla has previously posted positions for Teleoperator Engineers for Robotaxi, and an AI T=team member posted on LinkedIn about testing for remote assistance operations.

Scaling Up

Tesla initially intends to deploy a relatively small fleet in Austin, with its pilot beginning with about 10 to 20 vehicles on day 1. They will scale up gradually as the systems are validated and operations are refined, and the target is for anyone visiting Austin to be able to hail a robotaxi by the end of June or early July 2025.

Scaling Out

Tesla’s confidence in scaling out stems from its generalized approach with FSD. The team believes that once FSD is thoroughly verified and capable in a few diverse North American cities, deploying it to any other North American city should be technically feasible. After all, driving regulations don’t vary much from one major city to another, except for things like no right turns on red lights.

The primary limitation with scaling out geographically is expected to be securing regulatory approvals in different regions and jurisdictions. While the principles of Tesla’s FSD are expected to apply globally, as with theChina rollout, regulation will continue to be a considerable hurdle.

Cybercab

FSD Supervised ride-hailing service is live for an early set of employees in Austin & San Francisco Bay Area.

We've completed over 1.5k trips & 15k miles of driving.

This service helps us develop & validate FSD networks, the mobile app, vehicle allocation, mission control &… pic.twitter.com/pYVfhi935W

— Tesla AI (@Tesla_AI) April 23, 2025

Tesla’s fully autonomous, two-seater vehicle, theCybercab, is also being developed. While volume production for the Cybercab won’t begin until 2026, Tesla is already undergoing sample production validation. The first actual Cybercab builds are expected to be completed near the end of Q2 2025, and we’re excited to see if anything changes from the prototypes shown at We, Robot.

Tesla intends to build Cybercab at Giga Texas, and the production line is on schedule, according to Tesla. Cybercab production will not require a new building and will be built inside the existing Giga Texas factory.

Unboxed Method

Tesla's Shareholder Deck for Q1 2025 is now out, and we're 40 minutes away from the Earnings Call.

It seems Tesla won't be using the unboxed process for its more affordable model - which is a new vehicle.

Instead, they plan to use it for Cybercab soon! pic.twitter.com/0yddfWieJQ

— Not a Tesla App (@NotATeslaApp) April 22, 2025

The Cybercab’s production and low-cost methodology heavily rely upon Tesla’s innovative new unboxed manufacturing method. Tesla’s executive team mentioned that progress and implementation with this new manufacturing solution was going well, and it will be key to lowering the cost of production while considerably raising the level of automation.

Tesla has already achieved key milestones with this method, including successfully integrating large sub-assemblies and resolving challenges related to connecting portions of the vehicle's ceiling during the process. They have also successfully completed corrosion testing related to the new assembly method, as well as begun initial crash testing for the Cybercab.

All of that and more will eventually be incorporated into other vehicle production lines, and the Cybertruck is already benefiting from some aspects of these developments. With a final ambition of achieving a 5-second cycle time per Cybercab, Tesla has come a long way—but it still has quite a bit to go. They’ve managed to meet a 33-second cycle time for the Model Y at Giga Shanghai between two production lines, but 5 seconds per vehicle is quite ambitious.

With all that said, between Robotaxi and Cybercab, there is a lot of interesting news coming from Tesla. We expect this to continue throughout the year as Tesla launches its network and then aims to release FSD Unsupervised to customers, with a potential price increase.