In the race to deploy autonomous vehicles, there have been two schools of thought. One is led by sensor fusion, which means the more sensors and the more types, the better. The other is Tesla’s school of thought — vision.
So far, even Google’s CEO, Sundar Pichai, has described Tesla as the leader in the autonomy sector.
Google CEO on who is the leader in self-driving space: "I think obviously @Tesla is a leader in the space. It looks to me like Tesla and Waymo are the top two." pic.twitter.com/T0hlSICm8V
A new analysis from Bloomberg (paywall) offers a similar perspective, focusing on the numbers and real-world safety metrics. Tesla’s strategy isn’t just viable - it is far outpacing its competitors.
A Tale of Crash Rates
The most striking numbers from Bloomberg’s analysis are safety-related. According to their comparison, FSD reports approximately 0.15 crashes per million miles driven. In contrast, Waymo reported approximately 1.16 crashes per million miles.
That means that a Tesla using FSD is seven times less likely to be involved in a crash than a Waymo vehicle, even with its bevy of sensors. This is in line with Tesla’s latest vehicle safety report, which notes that a Tesla using FSD is 10 times less likely to be involved in an accident than a driver in any other vehicle.
Crash rates compared
Bloomberg
When it comes down to it, sensor fusion, while it can be fantastic, it simply provides too much data to process and analyze. While LiDAR, radar and cameras all have their unique advantages, cameras end up being the most versatile. Our roads and world were created around vision and audio, so a LiDAR-only vehicle can’t navigate our roadways since it would be unable to see signs or any other object that lacks depth. For LiDAR to be useful, it needs to be coupled with vision.
Vision works well because it applies to all situations, and it’s a system that continues to improve thanks to advancements in image processing and AI. While measurements with vision still lag behind LiDAR, they’ve reached a point where they’d “good enough,” and the millimeter-level accuracy of LiDAR isn’t needed.
When radar and vision disagree, which one do you believe? Vision has much more precision, so better to double down on vision than do sensor fusion.
Besides the difficulty of using sensor fusion, Bloomberg also points out that Tesla’s advantage is in the fundamental cost of the hardware. The Model Y costs just 1/7th of the total cost of a Waymo vehicle.
This enormous cost difference is a direct result of how Tesla and Waymo are approaching autonomy. Waymo’s vehicles are high-end, third-party electric cars, like the now-discontinued Jaguar I-Pace, which are then retrofitted with an expensive, custom-built suite of sensors. This sensor suite includes multiple LIDAR units, radars, and cameras.
Tesla’s ability to scale autonomous driving faster than its rivals gives it an edge in the self-driving race, says Bloomberg Intelligence's Steve Man https://t.co/B1x5Jhx6Lfpic.twitter.com/XYPCblWmXn
Tesla, meanwhile, includes all the hardware for autonomy as standard equipment on each of their vehicles, with a relatively inexpensive suite of cameras and its own in-house designed FSD computer. Using affordable hardware means it’s easy to produce and field more vehicles, resulting in more data.
On top of that, building more vehicles at a lower price creates a larger and larger economic difference as time goes on, as Tesla’s Robotaxis become profitable far quicker than Waymo’s.
3 Billion Miles… and Counting
The biggest advantage that Tesla has over any other entrant into the autonomy ring is simply just data. Tesla’s fleet has gathered over 3 billion miles of driving data globally, whereas Waymo’s fleet is just a minuscule 22 million miles.
Putting that into perspective, for every mile driven by a Waymo vehicle, a Tesla has driven over 135. Tesla’s advantage is also the fact that its data is global. It includes vehicles operating in a range of environments, from deserts to the Arctic, from cities to extremely rural areas, and is capable of achieving generalized autonomy.
Waymo’s data is extremely focused on urban and suburban areas and is effectively unusable for generalized vehicle autonomy. A larger, more capable fleet is the key to providing an effective robotaxi service, after all.
Scaling Manufacturing
Finally, Waymo doesn't produce vehicles. Tesla produces Robotaxis from scratch - every vehicle off the line has the ability to run Unsupervised FSD, and eventually join the Robotaxi fleet. Waymo needs to partner with other companies that have a good platform, and they must adapt their technology to that platform.
Waymo’s fleet is expected to be 2,500 vehicles by the end of 2025, while Bloomberg expects Tesla’s functional fleet to hit 35,000 by the same time. That’s not even counting the millions of AI4-powered vehicles that could also join the fleet by late 2026.
Overall, Tesla is a clear winner in the Robotaxi race - and it isn’t just because of one element. They’re winning through data, cost, and scalability, and the gap will only continue to grow.
Subscribe
Subscribe to our newsletter to stay up to date on the latest Tesla news, upcoming features and software updates.
For several months, we’ve been hearing rumors that Tesla is developing a new variant of the Model Y, one equipped with six seats rather than the five or seven we’re accustomed to seeing. Many people have been holding out for exactly such a variant, one that’s big enough to hold larger families without being as tight on legroom as the seven-seat variant.
Now, thanks to confirmation from Tesla hacker @greentheonly, we now have the strongest evidence yet that Tesla is making a six-seat variant, likely in a unique configuration, as he found in the latest batch of vehicle firmware.
Green found references to the six-seater Model Y in the 2025.20 update, validating the rumors we’ve heard. Given that the references are showing up in production software, Tesla may now be actively testing the vehicle. The next place to watch will be code in the Tesla app, which should hopefully provide us with some previews of the seat layout.
Captain’s Chairs
So, what would a six-seat Model Y actually look like? Well, the configuration all but guarantees a 2x2x2 seat layout. That configuration would replace the traditional second-row bench seat with two more spacious captain’s chairs, which will have an aisle between them. That’s a big bump in comfort for second-row passengers, who would have seats more similar to the ones up front. Tesla didn’t include armrests on the six-seat Model X, so it’s likely the Model Y won’t have them either.
For the Model X, the seven-seat configuration is a $3,500 bump in price, while the six-seater costs an additional $6,500, so expect the Model Y to have a similar price structure.
We can also expect Tesla to add some convenience features into this package, similar to the power-reclining seats introduced in the new Model Y, which can automatically fold and unfold with the tap of a button. In a six-seat Model Y, the middle seats would likely be able to automatically move forward and backward, while the rear two seats would be the folding ones.
Front seats learned some manners
If rear row needs more space to fold flat, front seats in new Model Y will automatically adjust pic.twitter.com/O8ik9QuPwA
However, most importantly, this 2x2x2 layout creates a center aisle. This provides easier access to the third row, eliminating the need to fold down the center bench seat to get in and out of the vehicle. That’s a big improvement for families who use the third-row regularly.
Confirming Rumors
This firmware update actually confirms rumors we heard back in December 2024. China’s Latepost and Reuters had already indicated that a six-seat variant was in the works back then, but many agreed that the variant would be an exclusive option for the Chinese market. That particular market prefers having spacious second-row seating, and the six-seater just makes sense there, given the offerings of other manufacturers.
Now, with Green’s assessment that a global release is more likely helps to confirm those reports, while also providing some hope for customers in North America and Europe who have been holding out.
Utility vs Comfort
The plot thickens when we look at the fact that Tesla just teased another recent development - the Model Y seven-seat variant. Some customers in North America and Europe received emails teasing the return of the seven-seat Model Y.
But why would Tesla release both? Likely for similar reasons that they have a six-seat and seven-seat Model X. The answer is that some will value the utility of the seven-seat option, while others will value the comfort and premium feeling of the six-seat option.
By offering both, Tesla gives its customers a bit more customization and preference, especially those who often have to look outside the brand for larger family vehicles.
Longer Wheelbase
While the firmware finding confirms the seating configuration, there have also been some rumors from Chinese bloggers about the possibility of a longer wheelbase, known as version E80. If this proves to be true, the six-seater would be more than just a new interior layout; it would be a fundamentally larger vehicle. Recent rumors suggest the vehicle may have a wheelbase that’s 5.9” longer than the regular Model Y, which would significantly increase the legroom available in the third row. While the third row may have ample legroom in this longer wheelbase, it would likely still be limited for smaller individuals due to the available headroom.
The Tesla Model X is Tesla’s only current offering for larger families, but it costs significantly more than the Model Y. A six or seven-seater Model Y could help open up a whole new category for Tesla.
Price
We expect that Tesla will likely price the seven-seat variant for an additional $2,000 USD ($3,000 CAD), and the six-seat variant for an additional $4,000 USD ($5,000 CAD) in the North American market, based on the pricing of the old seven-seat variant and the Model X’s six-seat premium configuration.
Below is a video of the seven-seater Model Y:
What Next?
The addition of the six-seat Model Y to Tesla’s public software that’s being delivered to vehicles today practically confirms that Tesla is looking to release a six-seat version of the Model Y. While the longer wheelbase remains speculative, we could expect Tesla to unveil a six-seat variant in the coming months. Alongside the seven-seat variant, customers shopping for a larger vehicle will have some additional choices soon.
While most automakers are adding basic and un-integrated AI assistants to their vehicles, Tesla appears to be working hand-in-hand with xAI to build out an entire cast of characters for in-vehicle Grok. In the latest dive into Tesla’s firmware, Tesla hacker @greentheonly uncovered a few new details about the upcoming Grok feature.
This discovery, part of a larger batch of changes found in Tesla update 2025.20, also offers a hint at the hardware required for Grok.
A Dozen Groks
While we knew from previous firmware teardowns that Tesla was adding various Grok personalities to the vehicle UI, this update introduces a new one called Language Tutor, which may allow users to learn and practice a new language.
These are the different personas that will be included in Grok:
Argumentative
Assistant
Conspiracy
Doctor
Kids’ Stories
Kids’ Trivia Games
Meditation
Motivation
Romantic
Sexy
Storyteller
Therapist
Language Tutor
and the Grok Classic - Unhinged Mode
Release Date
Icons for these personalities have been added to the firmware for the first time, leading us to think Tesla is getting close to releasing the feature. However, even after the feature is developed and included in vehicle software, Tesla may only enable it selectively — such as for select employees for testing.
Grok is expected to be one of the largest vehicle features added via a software update, so it’ll be a big deal when it’s finally released. While we know more or less what to expect from Grok, what we don’t know is about vehicle requirements or whether it’ll require a subscription beyond Premium Connectivity.
Grok Requirements
While subscriptions required are anyone’s guess, vehicle requirements may be a little clearer now, thanks to Green’s discovery that Grok is only being added to AMD-based vehicle software builds.
Unfortunately, this likely means that older, Intel-based vehicles will not receive Grok support, at least for the time being. When Tesla introduced a voice assistant in China, they also started off with AMD-only support but later added Intel support, so it’s possible that the same could happen with Grok.
Energy Saving
Green also found a new undocumented navigation feature in 2025.20.3. This one focuses on leveraging Tesla’s vast amounts of data in an interesting way and offers drivers proactive advice to save energy on a well-traveled route.
In the Energy App, your vehicle will display, “This route is typically driven at X mph. Slow down by Y mph to save Z% for the rest of this trip.’
As you can see in the photo below, Tesla recognizes that vehicles typically drive at 66 mph on the route being driven and offers the driver the option to go five mph slower to save 1% of range. While this example doesn’t provide much incentive to slow down, it could be useful for someone with a low state of charge or if the savings are more significant.
In the future, Tesla could even use this, alongside the road surface data they’re gathering, to help plan routes for Robotaxi and Unsupervised FSD.
![Tesla Launches 'TeslaVision' Contest With Big Prizes — See Last Year’s Winner [VIDEO]](https://www.notateslaapp.com/img/containers/article_images/2025/tesla-vision.webp/0458df1c6ed085c427608f0cf762de64/tesla-vision.jpg)
