Elon Musk provided a sneak peek into “Cortex”, Tesla’s new AI training center that has been under construction for quite a while at Giga Texas.
Previously, we saw some shots of the new wing for the massive water-cooled supercomputer cluster at Giga Texas, which was being built at the southern end of the Gigafactory.
AI4 Training
Cortex appears to be the name of the supercomputer cluster, and we can assume that it will be referred to that way from now on. Cortex will primarily focus on exploring and expanding FSD’s performance envelope to include AI4, which will begin to diverge from HW3 sometime soon. Tesla hasn’t yet unlocked the full potential of AI4, as it still runs HW3 in emulation mode, but with a few extra tricks that HW3 doesn’t have access to. These are mostly hardware compiler changes, but they make a significant difference in the speed of processing the AI model.
Tesla intends to continue supporting HW3, but it's already apparent that the FSD model for HW3 is smaller than the model for AI4. However, Ashok Elluswamy, Tesla’s Autopilot program director, expects performance to be similar for the most part.
One other thing that Cortex will likely be tackling is not just FSD in cars – but FSD in Optimus. Optimus uses the same principal technology behind FSD for movement, as well as understanding and learning tasks, so we can expect to see plenty of improvements as Tesla is now getting Cortex online.
Video of the inside of Cortex today, the giant new AI training supercluster being built at Tesla HQ in Austin to solve real-world AI pic.twitter.com/DwJVUWUrb5
What’s most interesting is that this new supercomputer cluster isn’t Tesla’s Dojo Supercomputers. Instead, this Cortex cluster is instead based on Nvidia’s H100 GPUs. Dojo was supposed to be Tesla’s in-house designed and specialized AI GPUs, meant to focus on optimizing FSD.
Dojo isn’t dead – but delayed from the sounds of it. We’re excited to see where Cortex takes Tesla in the future, and when Dojo shows up to join the party.
Tesla’s been on a roll with Supercharger improvements lately, from the 325kW charging update for the Cybertruck, to 500kW with V4 Superchargers coming next year. While those improvements have been limited to the Cybertruck, Tesla didn’t put all their focus on their new flagship vehicle, but looked at their more affordable vehicles as well.
LFP Battery Heating
Tesla’s Superchargers can now heat LFP Batteries - those that are in the Model 3 and Model Y Rear Wheel Drive variants. This applies to Long Range and Standard Range models, which saw a limited run. This is another update included as part of the 2024 Tesla Holiday Update - which really arrived with a lot of unannounced new features and capabilities.
The change is pretty interesting - Superchargers of the V3 and V4 variety can now pre-heat batteries for Model 3 and Model Y vehicles equipped with LFP battery packs. That means those vehicles are able to get back on the road faster when it's extremely cold. Of course, Tesla still advises you to precondition before you arrive, saving drivers time and money.
Max de Zegher, Tesla’s Director of Charging, also commented on the new feature. Essentially, Tesla is inducing an AC (alternating) ripple current through the battery to warm it up. Keep in mind that Superchargers are DC charging. That means it is possible to get a cold-soaked LFP vehicle on the road 4x faster than before, assuming that it didn’t precondition at all and that it is in the worst-case scenario (below 0ºF).
In essence, Tesla is using some engineering magic to turn the circuits inside the LFP battery into an electric heater - and powering that heater through the Supercharger. An AC ripple current is a small oscillation in the DC charging current that generates heat through electrical resistance, warming up the battery. Those ripples are a byproduct of converting AC to DC and back - so Tesla is using the onboard charger to induce those ripples to warm up the battery. Definitely an innovative technique that’s really only possible with the versatility of the NACS connector.
We’re hoping Tesla can implement this across their full lineup of vehicles, but we’ll have to wait and see how it is trialed across LFP vehicles first and if it is even possible on vehicles with 2170 or 4680 battery packs.
Tesla launched two FSD updates simultaneously on Saturday night, and what’s most interesting is that they arrived on the same software version. We’ll dig into that a little later, but for now, there’s good news for everyone. For Hardware 3 owners, FSD V12.6.1 is launching to all vehicles, including the Model 3 and Model Y. For AI4 owners, FSD V13.2.4 is launching, starting with the Cybertruck.
FSD V13.2.4
A new V13 build is now rolling out to the Cybertruck and is expected to arrive for the rest of the AI4 fleet soon. However, this build seems to be focused on bug fixes. There are no changes to the release notes for the Cybertruck with this release, and it’s unlikely to feature any changes when it arrives on other vehicles.
FSD V12.6.1 builds upon V12.6, which is the latest FSD version for HW3 vehicles. While FSD V12.6 was only released for the redesigned Model S and Model X with HW3, FSD V12.6.1 is adding support for the Model 3 and Model Y.
While this is only a bug-fix release for users coming from FSD V12.6, it includes massive improvements for anyone coming from an older FSD version. Two of the biggest changes are the new end-to-end highway stack that now utilizes FSD V12 for highway driving and a redesigned controller that allows FSD to drive “V13” smooth.
It also adds speed profiles, earlier lane changes, and more. You can read our in-depth look at all the changes in FSD V12.6.
Same Update, Multiple FSD Builds
What’s interesting about this software version is that it “includes" two FSD updates, V12.6.1 for HW3 and V13.2.4 for HW4 vehicles. While this is interesting, it’s less special when you understand what’s happening under the hood.
The vehicle’s firmware and Autopilot firmware are actually completely separate. While a vehicle downloading a firmware update may look like a singular process, it’s actually performing several functions during this period. First, it downloads the vehicle’s firmware. Upon unpacking the update, it’s instructed which Autopilot/FSD firmware should be downloaded.
While the FSD firmware is separate, the vehicle can’t download any FSD update. The FSD version is hard-coded in the vehicle’s firmware that was just downloaded. This helps Tesla keep the infotainment and Autopilot firmware tightly coupled, leading to fewer issues.
What we’re seeing here is that HW3 vehicles are being told to download one FSD version, while HW4 vehicles are being told to download a different version.
While this is the first time Tesla has had two FSD versions tied to the same vehicle software version, the process hasn’t actually changed, and what we’re seeing won’t lead to faster FSD updates or the ability to download FSD separately. What we’re seeing is the direct result of the divergence of HW3 and HW4.
While HW3/4 remained basically on the same FSD version until recently, it is now necessary to deploy different versions for the two platforms. We expect this to be the norm going forward, where HW3 will be on a much different version of FSD than HW4. While each update may not include two different FSD versions going forward, we may see it occasionally, depending on which features Autopilot is dependent on.
Thanks to Greentheonly for helping us understand what happened with this release and for the insight into Tesla’s processes.
![Andrej Karpathy considers returning to Tesla to work on Optimus [video]](https://www.notateslaapp.com/images/news/2022/karpathy-podcast_300w.jpg)
