Connect with us

News

Tesla Autopilot Director shares insights on FSD V12.5 Hardware 3 rollout

Credit: Arash Malek/X

Published

on

Tesla Director of Autopilot Software Ashok Elluswamy recently provided some insights behind the release of Full Self-Driving (FSD) V12.5 to Hardware 3 vehicles. As per the Tesla executive, the company was able to achieve similar performance as Hardware 4 vehicles running FSD V12.5 using a smaller model for HW3.

The release of FSD V12.5 to HW3 vehicles has long been anticipated by Tesla owners. Initially, Elon Musk noted in a post on X that HW3 vehicles should receive FSD V12.5 about ten days after the update was rolled out to HW4 vehicles. Musk’s estimate proved optimistic, resulting in quite a bit of frustration among owners of HW3 Teslas. 

As noted by Tesla software update tracker Not a Tesla App, speculations were high that software update 2024.26.15 would include FSD V12.5.1.4 support for Hardware 3 vehicles. Sure enough, the software tracker confirmed later that HW3 vehicles were indeed running FSD V12.5 after they received update 2024.26.15. 

The release of FSD V12.5 to HW3 vehicles was welcomed warmly by the Tesla community, with many thanking the EV maker for keeping its older fleet fully supported with the latest FSD systems. Amidst social media conversations about the update, the Autopilot Director noted that Tesla could probably run the same FSD V12.5 model on both HW3 and HW4 platforms in the future. 

“We were able to achieve similar performance as the AI4 12.5 release with a relatively smaller model for AI3. Deploying the larger model requires us to implement a few kernels in the compiler to emulate the same operations that are natively supported on the AI4 hardware. The team is actively working on implementing and verifying those operations. Combined with other model compression techniques, we think we can run the same 12.5 model on both the hardware platforms in the future.

“That said, model parameter count is just one of the many axes of performance improvement. Dataset distribution, dataset quality, model architecture, training compute, etc., also have tremendous impact on the final performance,” the Autopilot Director wrote in his post on X. 

Advertisement
-

FSD plays a huge role in Tesla’s future, as the technology is expected to be the backbone of the upcoming Robotaxi program. Tesla has also been pushing FSD hard to customers, as highlighted by the recent rollout of the Drive Anywhere Package for the Model S and Model X, which includes access to FSD (Supervised), Free Supercharging, and Premium Connectivity for three years.

Don’t hesitate to contact us with news tips. Just send a message to simon@teslarati.com to give us a heads up.

Simon is an experienced automotive reporter with a passion for electric cars and clean energy. Fascinated by the world envisioned by Elon Musk, he hopes to make it to Mars (at least as a tourist) someday. For stories or tips--or even to just say a simple hello--send a message to his email, simon@teslarati.com or his handle on X, @ResidentSponge.

Advertisement
Comments

Elon Musk

Elon Musk teases TSMC as potential Terafab partner

Published

on

SpaceX Terafab rendering
SpaceX Terafab rendering

Elon Musk has acknowledged that early discussions with Taiwan Semiconductor Manufacturing Company (TSMC) could bring the company into his ambitious Terafab semiconductor project, signaling a possible partnership with the world’s leading contract chipmaker.

Musk confirmed that early talks are underway, but as of right now, they are “just discussions.” There is no confirmation of a deal nor dismissal of the possibility of one, leaving open the prospect of one of the largest advanced-chip collaborations under discussion in the U.S.

The report that speculated on potential discussions between Terafab and TSMC comes from Tim Culpan, who outlined a few ways the collaboration could operate. One is TSMC using the project as an “anchor customer” for future facilities in Texas, potentially contributing process expertise, operational know-how, or capacity while Terafab provides capital, long-term purchase commitments, or both.

Tesla and SpaceX jointly developed the Terafab project, with Intel already participating on the tech side. Elon Musk announced the project in March, and it intends to produce more than one terawatt of AI compute capacity annually once fully built.

Elon Musk’s Terafab project locks up massive new partner

Company statements place the first phase at approximately $16.8 billion in cost, with later filings pointing to a total that could reach well into the tens of billions across multiple stages.

Intel joined the effort in April 2026 and is expected to supply its 14A manufacturing process for the full-scale plant.

Musk has said existing suppliers, including Samsung and TSMC, remain important for near-term needs; Tesla already has production arrangements with Samsung for AI5 and AI6 chips, but that future demand from Optimus robots, Cybercab vehicles, and planned space-based data centers will eventually exceed what the global industry can currently deliver.

Terafab is positioned as the long-term answer to that projected shortfall, and Tesla did something similar during COVID to avoid a chip shortage. This is just a much larger-scale solution.

If the partnership were to materialize, it would add TSMC’s industry-leading strategies to a project that already combines Tesla’s and SpaceX’s capital and offtake with Intel’s process technology. For now, the only public confirmation is Musk’s brief acknowledgement that conversations are occurring.

Continue Reading

News

Tesla reveals early Robotaxi charging strategy, showing scrappy DNA

Published

on

Credit: Tesla

Tesla’s early strategy for charging units operating within its Robotaxi fleet reveals that the company surely has not lost any of that scrappy DNA that took it from an unlikely success story to the most valuable carmaker in the world.

An observer at a Tesla Supercharger in Austin spotted ten total Robotaxi vehicles arrive: one Cybercab and nine Model Y units. A Tesla employee was waiting at the lot and allowed each unit to park itself; every car that arrived had nobody in it.

Tesla wins FCC approval for wireless Cybercab charging system

The Tesla employee would walk around and plug each car in, adjusting the parking if needed:

It’s a very interesting strategy, but extremely understandable at this early point in the Robotaxi program. It’s only been out for about 15 months, and Cybercab just entered the fleet in early September.

On top of that, Tesla is still working tirelessly on its wireless charging apparatus, and a new patent was just published regarding that product last week.

However, this is just another example of how Tesla still has plenty of that scrappy DNA leftover from the “production hell” days, when CEO Elon Musk slept on the floor of the factory, employees were working crazy hours, Tesla was building Sprung Structures to build cars in, and the company was tiptoeing on the brink of bankruptcy.

For now, Tesla is utilizing a simple system for recharging its ride-hailing vehicles, and that is a Tesla employee doing it manually until another solution presents itself. Sure, it’s not the most high-tech thing, and it certainly is not what people might have expected at this point in time, but it works, and it’s keeping the entire suite running.

Continue Reading

News

Tesla Robotaxi expands hours, Musk explains why it’s been a challenge

Published

on

Credit: Tesla

Tesla is expanding its Robotaxi service hours by pushing the time back by one hour, keeping the ride-hailing service operational until 11 p.m., one hour later than previously.

CEO Elon Musk confirmed the change and offered a specific reason the expansion has been gradual: the system still needs to reliably avoid small pets that are difficult to see after dark, as they commonly blend into the color of the road, especially when they’re grey.

The latest adjustment restores only a fraction of the operating window the service once held. When paid Robotaxi rides began in Austin on June 22, 2025, vehicles ran from 6 a.m. to midnight.

Tesla Robotaxi will be a 24/7 service: here’s when

In September 2025, Tesla lengthened the day to a 2 a.m. close, producing a 20-hour window that stayed in place for most of the following year. By early August of this year, the cutoff had already been pulled back; an August 26 update formalized hours of 6 a.m. to 10 p.m. across Austin and several other markets.

The October move to 11 p.m. therefore leaves the Austin day one hour shorter than the original launch schedule and three hours shorter than the 2025 peak.

Musk addressed the constraint directly after the announcement. “The main thing we’re trying to solve is making sure that we don’t run over pets when they’re hard to see at night,” he wrote. “Literally trying to avoid grey kittens on grey tarmac in the dark.”

The example points to a low-contrast perception problem in which a small animal can blend into the road surface under limited lighting.

Tesla’s vehicles rely on cameras and neural-network processing rather than lidar; Musk has previously argued that advanced vision software can extract useful information even in low light by analyzing photon counts, but the pet-detection case remains the stated limiter in later hours.

The modest schedule change arrives alongside faster growth in the purpose-built Cybercab fleet. Texas registration data tracked by observers showed the Austin Cybercab count rising sharply in recent weeks, reaching 169 vehicles after more than 100 were added in a short span.

Tesla has indicated that a broader shift toward 24-hour operation is tied to the upcoming FSD v15 software release expected this month on Robotaxi vehicles. Until that capability is validated for the edge cases Musk described, the company continues to add service time incrementally rather than jumping straight to overnight coverage.

The one-hour extension gives Austin riders a later option for evening trips while the underlying detection work continues.

Continue Reading