Connect with us

News

Tesla begins retrofits of Hardware 3 Full Self-Driving computer on selected cars

Published

on

It appears that Elon Musk’s goal of achieving Full Self-Driving is getting closer to reality, with reports from members of the Tesla community indicating that the electric car maker is starting to retrofit selected vehicles with its custom-designed Hardware 3 computer. Tesla considers HW3 as a key to achieving autonomous driving, as indicated by the company in its Autonomy Day event this past April. 

Tesla owner-enthusiast Sofiaan Fraval mentioned on Twitter that his vehicle was given the HW3 retrofit after he took in his 2018 Model S P100D for a voluntary HEPA filter replacement. Upon receiving a report for his vehicle’s service, the Tesla owner was pleasantly surprised to see that service center had installed its HW3 Full Self-Driving computer on his Model S. “Performed the retrofit for the Full-Self Driving computer. Test drove and verified operation,” Tesla wrote. 

Prior to its HW3 retrofit, Fraval’s Model S P100D was equipped with Hardware 2.5. The vehicle had a build date of 04/2018 and was equipped with Tesla’s faster, upgraded MCU. It should be noted that Hardware 3 retrofits are reportedly exclusive for Model S and Model X vehicles for now, and are performed only for selected vehicles. Thus, at this stage, it is pertinent to note that Hardware 3 retrofits are not available on request yet. 

Elon Musk has been pretty open about the idea of retrofitting older vehicles with the company’s custom-designed Full Self-Driving computer. Prior to the company’s Autonomy Day event, Musk had emphasized that HW3 retrofits will be simple, considering that the new computer has the same footprint as its NVIDIA-powered predecessor. The Tesla CEO initially remarked that retrofits will be performed at a service center, though later tweets have indicated that the upgrade could be accomplished by the company’s mobile service team

Advertisement

Tesla’s Hardware 3 computer is specifically designed to achieve Full Self-Driving. Discussing the custom hardware during an appearance at ARK Invest’s For Your Innovation podcast, Musk estimated that Hardware 3 could provide the company’s electric cars with a 1,000% improvement in processing capability compared to the NVIDIA components equipped in Tesla’s HW2 and HW2.5 cars. 

“We can process on the order of 100 frames a second and we really need to do a lot of work in terms of cropping the frames, and sort of bending the pixels, and not going to full resolution on all cameras, that kind of thing with the current hardware. We’re at full frames, full resolution with the Tesla hardware. All cameras, at full resolution, full frames, and it still hasn’t tapped out,” Musk said.

H/T Vincent Yu.

Advertisement

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

Tesla owners surpass 8 billion miles driven on FSD Supervised

Tesla shared the milestone as adoption of the system accelerates across several markets.

Published

on

Credit: Tesla

Tesla owners have now driven more than 8 billion miles using Full Self-Driving Supervised, as per a new update from the electric vehicle maker’s official X account. 

Tesla shared the milestone as adoption of the system accelerates across several markets.

“Tesla owners have now driven >8 billion miles on FSD Supervised,” the company wrote in its post on X. Tesla also included a graphic showing FSD Supervised’s miles driven before a collision, which far exceeds that of the United States average. 

The growth curve of FSD Supervised’s cumulative miles over the past five years has been notable. As noted in data shared by Tesla watcher Sawyer Merritt, annual FSD (Supervised) miles have increased from roughly 6 million in 2021 to 80 million in 2022, 670 million in 2023, 2.25 billion in 2024, and 4.25 billion in 2025. In just the first 50 days of 2026, Tesla owners logged another 1 billion miles.

Advertisement

At the current pace, the fleet is trending towards hitting about 10 billion FSD Supervised miles this year. The increase has been driven by Tesla’s growing vehicle fleet, periodic free trials, and expanding Robotaxi operations, among others.

Tesla also recently updated the safety data for FSD Supervised on its website, covering North America across all road types over the latest 12-month period.

As per Tesla’s figures, vehicles operating with FSD Supervised engaged recorded one major collision every 5,300,676 miles. In comparison, Teslas driven manually with Active Safety systems recorded one major collision every 2,175,763 miles, while Teslas driven manually without Active Safety recorded one major collision every 855,132 miles. The U.S. average during the same period was one major collision every 660,164 miles.

During the measured period, Tesla reported 830 total major collisions with FSD (Supervised) engaged, compared to 16,131 collisions for Teslas driven manually with Active Safety and 250 collisions for Teslas driven manually without Active Safety. Total miles logged exceeded 4.39 billion miles for FSD (Supervised) during the same timeframe.

Advertisement
Continue Reading

Elon Musk

The Boring Company’s Music City Loop gains unanimous approval

After eight months of negotiations, MNAA board members voted unanimously on Feb. 18 to move forward with the project.

Published

on

The-boring-company-vegas-loop-chinatown
(Credit: The Boring Company)

The Metro Nashville Airport Authority (MNAA) has approved a 40-year agreement with Elon Musk’s The Boring Company to build the Music City Loop, a tunnel system linking Nashville International Airport to downtown. 

After eight months of negotiations, MNAA board members voted unanimously on Feb. 18 to move forward with the project. Under the terms, The Boring Company will pay the airport authority an annual $300,000 licensing fee for the use of roughly 933,000 square feet of airport property, with a 3% annual increase.

Over 40 years, that totals to approximately $34 million, with two optional five-year extensions that could extend the term to 50 years, as per a report from The Tennesean.

The Boring Company celebrated the Music City Loop’s approval in a post on its official X account. “The Metropolitan Nashville Airport Authority has unanimously (7-0) approved a Music City Loop connection/station. Thanks so much to @Fly_Nashville for the great partnership,” the tunneling startup wrote in its post. 

Advertisement

Once operational, the Music City Loop is expected to generate a $5 fee per airport pickup and drop-off, similar to rideshare charges. Airport officials estimate more than $300 million in operational revenue over the agreement’s duration, though this projection is deemed conservative.

“This is a significant benefit to the airport authority because we’re receiving a new way for our passengers to arrive downtown at zero capital investment from us. We don’t have to fund the operations and maintenance of that. TBC, The Boring Co., will do that for us,” MNAA President and CEO Doug Kreulen said. 

The project has drawn both backing and criticism. Business leaders cited economic benefits and improved mobility between downtown and the airport. “Hospitality isn’t just an amenity. It’s an economic engine,” Strategic Hospitality’s Max Goldberg said.

Opponents, including state lawmakers, raised questions about environmental impacts, worker safety, and long-term risks. Sen. Heidi Campbell said, “Safety depends on rules applied evenly without exception… You’re not just evaluating a tunnel. You’re evaluating a risk, structural risk, legal risk, reputational risk and financial risk.”

Advertisement
Continue Reading

Elon Musk

Tesla announces crazy new Full Self-Driving milestone

The number of miles traveled has contextual significance for two reasons: one being the milestone itself, and another being Tesla’s continuing progress toward 10 billion miles of training data to achieve what CEO Elon Musk says will be the threshold needed to achieve unsupervised self-driving.

Published

on

Credit: Tesla

Tesla has announced a crazy new Full Self-Driving milestone, as it has officially confirmed drivers have surpassed over 8 billion miles traveled using the Full Self-Driving (Supervised) suite for semi-autonomous travel.

The FSD (Supervised) suite is one of the most robust on the market, and is among the safest from a data perspective available to the public.

On Wednesday, Tesla confirmed in a post on X that it has officially surpassed the 8 billion-mile mark, just a few months after reaching 7 billion cumulative miles, which was announced on December 27, 2025.

The number of miles traveled has contextual significance for two reasons: one being the milestone itself, and another being Tesla’s continuing progress toward 10 billion miles of training data to achieve what CEO Elon Musk says will be the threshold needed to achieve unsupervised self-driving.

The milestone itself is significant, especially considering Tesla has continued to gain valuable data from every mile traveled. However, the pace at which it is gathering these miles is getting faster.

Secondly, in January, Musk said the company would need “roughly 10 billion miles of training data” to achieve safe and unsupervised self-driving. “Reality has a super long tail of complexity,” Musk said.

Training data primarily means the fleet’s accumulated real-world miles that Tesla uses to train and improve its end-to-end AI models. This data captures the “long tail” — extremely rare, complex, or unpredictable situations that simulations alone cannot fully replicate at scale.

This is not the same as the total miles driven on Full Self-Driving, which is the 8 billion miles milestone that is being celebrated here.

The FSD-supervised miles contribute heavily to the training data, but the 10 billion figure is an estimate of the cumulative real-world exposure needed overall to push the system to human-level reliability.

Continue Reading