Autonomous driving may be the future, but the present still has a lot of explaining to do. Especially when cars with so-called “Full Self-Driving” capabilities start careening off the road for no obvious reason.

That said, it’s rare to see what we just have in a newly released set of videos involving a Tesla. According to the title, it shows a crash while running what Tesla calls its autonomous system, Full Self-Driving (Supervised). What’s worse, though, is that it seems to do so without rhyme or reason in broad daylight with no traffic on a straight road.

More: Tesla Stiffs Cybertruck Owners On Another Promised Feature

Tesla famously uses vision-based software and hardware to run its semi-autonomous Autopilot and Full Self-Driving (Supervised) software. In theory, it makes complete sense since we humans also drive almost entirely via vision-based mechanics. In practice, though, there are some major concerns, and this video highlights them. We’ll circle back to that.

The Incident: Straight Road, Sharp Left Turn

A YouTube channel recently uploaded four videos showing each side of a car during a crash. They say this is a Model 3 and that it’s running FSD 13.2.8, which is almost the latest available version. On May 11, Tesla released 13.2.8, but this crash happened on February 26 so indeed, it was up to date given that information.

That said, what the video shows is the most shocking part of this entire situation. Across three of the four clips, we see the car moving for 45 seconds. In all of them, everything appears totally normal for the first 31 seconds as the car trundles down a two-lane road. Then, just as a car passes going the opposite way, all hell seems to break loose.

The car turns hard to the left, goes across the opposing traffic lane, goes off the road, and hits a tree before rolling over. From the moment it begins to turn to the moment it impacts the tree is less than three seconds. While that’s tough to swallow, it’s the conditions that really make this a bad deal for Tesla.

The road was perfectly straight. This appears to be at some time in the relatively early or later part of the day as the shadows cast on the ground are long. Despite that, the sun is bright and seemingly unobstructed by clouds, so there’s no lack of lighting in the scene. Finally, there’s no complex traffic situation here with markings, other cars, or road signs.

Still, for whatever reason, it appears as though this car allegedly on FSD just decided that it needed to leave the roadway and did exactly that. Adding even more confusion to this crash are videos of YouTubers testing FSD against inanimate objects on the road. In almost every case, the technology focuses on slowing itself down, stopping even, to avoid an obstacle. Very rarely does it try this sort of hard steering input at speed.

The Lidar Elephant in the Room

And this brings us back to vision-based autonomous driving systems. Again, we humans use vision to determine how to control our cars. Tesla is trying to do that too, but it’s caught flack, and I suspect is about to catch far more, over its choice to skip using lidar and radar tech.

While vision can work, and obviously does for most people on most days, Lidar and radar offer the ability to easily see through bad weather conditions like fog or haze. They could simply be used as a redundancy to confirm what a vision-based system thinks it sees too. Nevertheless, Tesla ditched it years ago and its CEO Elon Musk appears committed to never bringing it back.

Reports From The Driver

According to the person who posted the videos on Reddit, he was going around 55 mph when the crash happened. He says of the experience, “I loved the FSD until this happened. I was a full believer in autonomous vehicles until this happened to me. Lesson learned.” Thankfully, the only injuries he suffered included a cut on his chin, some lower-back discomfort, and “emotional damage,” as he calls it.

It’s worth pointing out that there are many unknowns here. While there appears to be no reason to suspect these videos and their description are inaccurate, there could be more to the story that we’re not being told. If that doesn’t end up being the case, though, Tesla is likely in a lot of hot water over this. The owner has submitted requests for all of the data relating to the crash so hopefully more of that sees the light of day.

Previous crashes involving the software typically offered some sort of purchase for Tesla defenders to cling to. Based on everything available in the four videos here, it appears as though FSD just made its most blatant mistake in the public sphere.

If this is possible with the hardware and software running Tesla’s planned Robotaxi service, it might have to be even more careful than it’s already planning on being. When asked if he’d ever buy another Tesla, the owner of this car’s words were damning. “I want another but would NEVER use FSD again.” Yeah, I think we can all understand why.

Tesla has been promising fully autonomous robotaxis for what feels like a lifetime, and now they’re telling us we can finally expect them next month. At least, that’s the plan, according to Tesla itself, with a “bullish” analyst now offering up some fresh details.

More: Tesla’s Robotaxi And Cybercab Might Need New Names

However, it’s not going to be the massive fleet of robo-cabs that everyone was hoping for. Instead, we’re looking at fewer than two dozen vehicles, and they’ll only be available to a select group of people. So, get in line (if you’re invited).

That news comes from Morgan Stanley Analyst Adam Jonas. According to a report from Investors, he recently met with Tesla to review how things are going. After all, Tesla has said repeatedly that it’ll launch its Robotaxi service in June of this year.

The Fleet’s Not So Grand After All

Jonas writes that “Austin’s a ‘go’ but fleet size will be low. Think 10 to 20 cars. Public roads. Invite only. Plenty of tele-ops to ensure safety levels (‘we can’t screw up’). Still waiting for a date.” That’s certainly not the huge splash many might have expected from Tesla. At the same time, some of that confirms what we already knew. Let’s break it down a little.

First, the fact that everything is still a ‘go’ sounds like Tesla is still on schedule. That’s encouraging. The small fleet size is a bit surprising, though. Even at the top end of 20 cars, one would imagine it’ll take a while to gather the data necessary to expand the program quickly. Tesla has long hinted that the service would include public roads, so that’s not shocking, but limiting it to invite-only status is interesting.

Perhaps that’s just a preliminary stage of the project, but it could help the brand avoid bad press should something go wrong. Of course, Jonas also confirms once again that Tesla will have teleoperators in place to monitor cars and take over if need be. How that’ll work in practice is up for debate, though.

A Reality Check

If we’re being serious here, this doesn’t sound all that different from what Tesla is already doing in Austin and San Francisco. Both cities offer a ridesharing app for some employees. In each city, users can hail a Tesla that will show up and take them to their destination with one big safety net: a human in the driver’s seat in case something goes wrong.

The new Robotaxi service in Austin sounds like it’s the same idea, but with a human driver dialed in remotely. The real standout here is Tesla’s Ground Truth Machine technology, which is busy mapping Austin using lidar and radar sensors.

Optimus News

Jonas says that we can expect additional news on Optimus, Tesla’s humanoid robot, later this year as well. If all goes according to plan, investors will have the opportunity to tour the robot manufacturing site during the final quarter of 2025. The automaker/tech company is evidently targeting a $20,000 price tag for Optimus. That’s a lot of cash, but no doubt some one-percenters won’t think twice about it. 

As more automakers revise their EV plans in response to the realities of the market, Honda is also adjusting its electrification strategy. President and CEO Toshihiro Mibe recently announced significant changes to the company’s approach, including a sharp reduction in both its ambitious sales targets and, more importantly, its investment in electric vehicles.

More: Honda’s Future EVs Will Let You Pretend You’re Driving An NSX Or S2000 With Simulated Sounds And Shifts

Slower-than-expected EV adoption has been a key factor behind this shift. While Honda still sees electric vehicles as the best long-term path to carbon neutrality by 2050, the pace of adoption has not kept up with predictions. A combination of evolving environmental regulations and shifting trade policies has kept EVs from breaking through at the rate many hoped for.

Adjusted Expectations

Honda now predicts that EVs will make up less than 30% of its global sales by 2030. In response, the company is slashing its planned investment in electrification from ¥10 trillion ($69 billion) to ¥7 trillion ($48 billion) by 2031. Part of this reduction stems from the postponement of a major EV investment project in Canada.

More importantly, Honda is introducing a new mixed production system that can handle both EVs and HEVs, with the added flexibility to shift between different factories. This will be paired with a “resilient supply chain strategy” designed to make adjustments as needed, depending on market fluctuations in different regions.

Next-Generation Hybrids

With demand for hybrids expected to continue growing toward the end of the decade, Honda plans to introduce 13 next-generation HEV models globally between 2027 and 2031. These hybrids will sport the redesigned “H” emblem, which was previously reserved for EVs. The company aims to reach 2.2 million annual HEV sales by 2030, contributing to a broader sales increase beyond the projected 3.6 million units in 2025.

More: 2026 Honda Prelude Coupe Interior Revealed With Civic Vibes

Honda’s two-motor e:HEV hybrid system will also see improvements, offering enhanced efficiency and better packaging. A new all-wheel-drive (AWD) unit will further elevate performance. The next-generation hybrid system will be 30% cheaper to produce than the current version, making it a more cost-effective option for the company.

Big Plans for North America

Specifically for North America, Honda is developing a new hybrid system tailored for larger vehicles, with an emphasis on high performance and towing capabilities. This powertrain will debut in models set to launch in the next few years, including a large SUV. While the model isn’t named, we suspect that it could very well be a replacement for the Pilot.

Sophisticated ADAS

The company is also investing heavily in intelligent technologies like advanced driver-assistance systems (ADAS). Aiming to enhance the competitiveness of both EVs and HEVs, the next generation of Honda’s ADAS will offer a higher level of autonomy in both city and highway driving. These systems are expected to be launched around 2027 across a wide range of EVs and HEVs in North America and Japan.

Besides its own-developed next-gen ADAS, Honda will also work with Chinese startup Momenta Global Limited to develop systems tailored for all future Hondas that will launch in China.

Honda

The dream of a driverless Tesla fleet shuttling people around while their owners kick back at home has been around for years, always just around the corner, but never quite here. Now, as the company continues to promise that reality is almost within reach, the U.S. Patent Office may have just introduced another delay.

As it turns out, names like “Robotaxi” and “Cybercab” might be too generic to trademark, and that could complicate Tesla’s rollout plans.

Read: Tesla Stiffs Cybertruck Owners On Another Promised Feature

It’s worth noting that Cybercab and Robotaxi refer to different things in the Tesla world. The former is the two-door prototype the automaker unveiled last year. The latter is the software that could enable everyday Tesla owners to let their car go around picking people up and moving them around while the owner is busy working or doing just about anything else.

According to TechCrunch, the USPTO just issued a non-final office action on the trademark application for the name Robotaxi. Specifically, the office said that name “describes a feature, ingredient, characteristic, purpose, function, intended audience of applicant’s goods and/or services.” In layman’s terms, it’s too general. The office even cited Wikipedia, Zoox, and The Verge to prove it.

“This term is used to describe similar goods and services by other companies,” the agency wrote. That mirrors a similar decision it made in April regarding Tesla’s attempt to trademark “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

In that motion, the USPTO pointed to multiple concerns, including the potential for consumer confusion. In fact, it even mentioned other companies that use the word Cyber, including ones that specifically build aftermarket products for the Cybertruck. In a way, Tesla did this to itself. For now, though, the names aren’t dead and gone.

In both cases, Tesla can argue its case with whatever evidence it thinks is relevant. No doubt, the two words do seem tied to the automaker a little more all the time. Tesla must respond within three months or the USPTO will abandon the application. That runs well past Tesla’s planned June rollout, so expect more news on this topic sooner rather than later.

Tesla is rolling out a pair of what it hopes will be high-impact incentives. In an effort to drive sales before the end of Q2, it’s offering 0% APR financing on Model 3 and the return of Full Self-Driving (Supervised) transfers. Both offers are appealing on paper, but like most things Tesla, there are some strings attached.

The headliner here is the 0% APR deal, which is only available on the Model 3. Not the Model Y, not the S, not the X, just the entry-level sedan. Qualifying depends on options and tax incentives, but it appears to be available with $0 down in some states.

Read: Tesla Quietly Launches Robotaxi Rideshare App For Employees

It’s also capped at 60 months. That’s about $6,000 down (over the loan term) on the cheapest rear-wheel-drive Model 3, and you’ll also need top-tier credit to lock it in. Still, it’s a rare move in today’s high-interest environment, and one that could tip the scales for shoppers on the fence.

Also back, for a limited time: FSD transfers. If you previously bought Tesla’s Full Self-Driving beta package for anywhere from $7,500 to $12,000 and are thinking about upgrading to a new vehicle, you can bring that feature with you, no need to repurchase. There’s no word on when the deal will end, but we expect it won’t live into the third quarter, so interested parties will want to move sooner rather than later.

0% APR available for financing (60 months, US only) https://t.co/eKuwpC1UKI

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

These moves, of course, come at a time when Tesla is navigating a tougher sales climate. Competition is getting better, opposition is getting fiercer, and tariffs could drive prices up more, too. Obviously, Tesla wants to clear inventory and keep customers inside the family with these deals.

While these two deals are good for those interested, they’re not the only ones we’ve seen in recent months. The new Model Y is available in China with 0% APR, and the old Model Y is the subject of massive discounts too. Even the opinion-splitting slab-sided Cybertruck is available for thousands off of MSRP right now. Clearly, Tesla knows it needs to move metal, and it’s doing all it can to achieve that goal.

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

Tesla just took a big step forward toward bringing robotaxis to the public. Strangely, it didn’t come with a bunch of fanfare from its CEO either. Instead, the company quietly confirmed on X that it has made its ride-hailing service available to some employees.

“FSD Supervised ride-hailing service is live for an early set of employees in Austin & San Francisco Bay Area,” read the post. Those cities are both very important for Tesla as the brand has facilities in each. A short video posted along with the announcement tells us a bit more about how it works.

More: Canadian Dealers Want To Blow Up Border Rules And Import Cool Cars Americans Can’t Buy

A rideshare app allows users to click a big button that simply says “Pick Me Up.” Once the car arrives, users get a notification and a reminder to buckle up in the car. Cybertruck-styled font shows a new script that calls the service “Robotaxi.” The video seems to indicate that, for now, a human safety driver is in the driver’s seat during the testing.

Once in the car, it appears riders get a message on the rear-seat infotainment system to confirm their name, address, and arrival time. Riders click a button in the car that says “Start Ride,” and off goes the car. According to Tesla, it’s already completed 1.5k trips and 15k miles of driving.

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

Interestingly, the Robotaxi nomenclature helps us understand an important point. The Cybercab refers to the vehicle, and Robotaxi is a service that can leverage more of Tesla’s models. No doubt, they’ll share software and functionality to a degree.

Perhaps the biggest surprise here isn’t that this is happening but that it’s seemingly on time. Elon Musk has famously been overly optimistic about true Level 5 driving tech. This time, it seems like his promise that a paid version of this service will go live in June could end up coming true. Even then, though, we expect a full-scale rollout to take a good long while. 

After a lackluster first quarter, where Tesla’s global sales took a nosedive, the EV giant has now reported its earnings for the same period. And, unsurprisingly, the results aren’t impressive. Revenue fell short of where it was expected to be by analysts, not just because of sluggish sales, but also due to uncertainty across industries and challenges in global supply chains.

But not everything is doom and gloom for Tesla, there’s some good news tucked in there, too. The company insists it’s on the cusp of launching more affordable models.

More: VW Finally Beats Tesla In EV Sales Across Europe

Let’s start with the figures. In Q1, Tesla’s total revenue fell to $19.3 billion, a 9.4% decline from $21.3 billion in Q1 2024. When compared to Q2, Q3, and Q4 of last year, the drop is even more significant, with revenue sitting at $25.5 billion, $25.1 billion, and $25.7 billion, respectively. Even worse was Tesla’s automotive revenue, which dropped from $17.3 billion in Q1 2024 to $13.9 billion this quarter, a 19.6% decline, and peaked at $20 billion in Q3 2024, a 30.5% decrease from that high.

Declining revenue isn’t the only thing that could concern shareholders of the EV brand.T otal income plummeted by 71%, dropping to $409 million, down from $1.39 billion in Q1 2024 and $2.1 billion in Q4 2024.

Why Did This Happen?

Tesla has blamed its decline in revenue on the obvious drop-off in deliveries during the first quarter. As we reported earlier this month, Tesla delivered 336,681 vehicles, down 13 percent from the year prior. It says deliveries fell in part because it was updating its four factories to start building the new Model Y. Tesla also says average vehicle selling prices dropped last quarter.

“Uncertainty in the automotive and energy markets continues to increase as rapidly evolving trade policy adversely impacts the global supply chain and cost structure of Tesla and our peers,” Tesla said. “This dynamic, along with changing political sentiment, could have a meaningful impact on demand for our products in the near-term. We remain committed to expanding our business model to include delivering autonomous robots across multiple form factors and use cases – powered by our real-world AI expertise – to our customers and for use in our factories, as we navigate these headwinds.”

New, More Affordable Models Are Coming

Falling revenue and deliveries don’t make for pretty reading, but Tesla does have some interesting things around the corner. New production vehicles, including more affordable models, will start in the first half of 2025, echoing a statement Tesla made in January and contrary to recent reports of these new models being delayed.

Read: Tesla Reportedly Delays Cheaper Smaller Model Y, Plans Stripped-Down Model 3

Details about these affordable versions are limited, with some suggesting they’ll be stripped down versions of the Model 3 and Model Y. Tesla says they will “utilize aspects of the next generation platform as well as aspects of our current platforms and will be produced on the same manufacturing lines as our current vehicle lineup.”

Elsewhere, Tesla announced that production of the Cybercab is scheduled to begin in 2026. The company also shared that the next-generation Roadster has moved from the “development” phase to “design development,” though it has yet to provide any details on when it will actually launch. Just a reminder: it’s been 7.5 years since the Roadster was first revealed.

Musk Winding Down DOGE Work

Finally, in news that may actually make some shareholders breathe a sigh of relief, Elon Musk will start scaling back his involvement with the Department of Government Efficiency (DOGE). Starting next month, Musk will devote more of his attention to Tesla, where he’s needed most. Perhaps this will bring more stability, though, knowing Musk, expect a few distractions along the way.

Mercedes is working on a new steer-by-wire system set to arrive in production vehicles by 2026, accompanied by a yoke-shaped steering wheel. Joining the likes of Tesla, General Motors, Geely, Toyota, and Lexus, the German automaker is jumping on the steer-by-wire bandwagon, which replaces the traditional mechanical connection between the steering wheel and wheels with an electrical cable.

More: Mercedes Vision V Concept Is A Billionaire’s Shuttle With A 65-Inch TV

While we don’t know exactly which model will get this upgrade first, a camouflaged prototype of the facelifted EQS recently popped up, and it’s rumored to be the car leading the charge. The electric sedan is expected to debut next year, possibly sporting the yoke-shaped wheel. The new S-Class might follow suit, along with the next generation of their ICE-powered models.

According to Mercedes, the new steer-by-wire system offers “superior driving characteristics,” which they say translates into a more responsive feel on the open road and greater agility in urban environments. The system is also designed to make parking easier, eliminating the need for the driver to re-grip the wheel. As an added bonus, it’s adaptable to various vehicle types and architectures, providing a wider range of settings for comfort or sportiness.

Yoke: A Solution Searching For A Problem

Mercedes-Benz

The company claims the yoke-shaped steering wheel frees up space in the cockpit, making it easier to get in and out of the car, while also providing an unobstructed view of the digital instrument cluster. With no mechanical connection to the wheels, the yoke can even be used for gaming and is more suited for autonomous vehicles.

However, Mercedes’ decision to adopt it is questionable. While it may open up space, it makes basic tasks like turning at low speeds or parking awkwardly difficult, leaving drivers missing the comfort and control of a traditional steering wheel. It seems more like a solution in search of a problem, prioritizing style and novelty over practicality and driving ease.

More: Tesla Tries, And Fails, To Reinvent The Steering Wheel With Its Yoke

Mercedes’ steer-by-wire system has clocked over one million kilometers on test benches, with a similar distance covered on proving grounds and in regular road traffic conditions. For safety, the system includes redundant architecture, onboard data, and a backup power supply, ensuring the driver can always steer the vehicle. Additionally, in the unlikely event of a complete failure, lateral guidance remains possible thanks to rear-axle steering and targeted braking interventions on individual wheels.

Markus Schafer, Mercedes-Benz’s CTO, sees steer-by-wire as a major step toward the future of mobility. He says, “The technology enables a unique customer experience that goes far beyond steering alone. In combination with SAE Level 3 conditionally automated driving, it will enable an even more immersive entertainment experience in the medium term. The flat steering wheel will provide a better view of the display when streaming your favorite show, for example.”

Steering Wheels, Joysticks, And Yokes

Mercedes also likes to remind us that Carl Benz’s 1886 Patent Motorwagen, the first automobile, had a steering crank, and that the first traditional steering wheel appeared in the 1894 Panhard 4HP. Since then, the company has dabbled with alternative steering systems in concept cars, including the futuristic joysticks found in the 1996 F200 Imagination and the F-Cell Roadster from 2009.

Currently, other production vehicles with steer-by-wire systems include the Tesla Cybertruck, GMC Hummer EV, Rolls-Royce Spectre, Lotus Eletre, and the Lexus RZ 405e. The latter will be the first to introduce this technology to Europe later this year.

Tesla has a lot riding on its Cybercab robotaxi, but the program has hit a major snag, and CEO Elon Musk’s sometimes best buddy, President Trump is to blame. The automaker has been forced to drop its plans to ship essential Cybercab components from China as a result of Trump slapping enormous tariffs on Chinese imports, a report claims.

More: Tesla’s CyberCab Promises 300-Mile Range with Surprisingly Small Battery

The automaker expected to start sending parts to the US from China in the next few months, necessary to begin trial production of both the Cybercab and the Semi truck, the latter of which has so far only been produced in tiny quantities. The plan was to scale up to full production of both models in 2026.

Tariffs Throw a Wrench in the Works

Trump’s announcement of a 34 percent tariff wasn’t great news, but Reuters sources say Tesla was prepared to absorb the financial pain. But when Trump’s tit-for-tat tariffs battle escalated to the point where import duty on Chinese imports had reached 145 percent, Tesla slammed on the brakes.

Exactly how long Tesla will keep its foot on the brakes is unclear, since no one, perhaps not even Donald Trump himself, knows the duration of the massive tariffs. But the US President revealed earlier this week that he was considering making changes to the 25 percent tariff on imported auto parts built in Canada, Mexico, and other regions, and has recently announced an exemption on electronic devices such as iPhones which are made in China.

Tesla’s US Sourcing Strategy

Reuters says Tesla has, for the past couple of years, been increasing the amount of parts it gets from within US borders because it sensed that tariffs might one day come into effect. How quickly Tesla can switch suppliers and get Cybercab and Semi components from America isn’t clear – we’ve asked the question, but don’t expect Tesla to reply.

Tesla unveiled its long-awaited Cybercab last fall, a Honda CRX-shaped pod with scissor doors and no steering wheel, and is currently working on getting approvals to begin testing and operating driverless cars in the US and beyond.

Images: Telsa

News about recent firings by Elon Musk’s Department of Government Efficiency (DOGE) has raised concerns about potential conflicts of interest for the Tesla CEO. During DOGE’s fiery critique of the federal workforce in February, it was revealed that several staff members from the National Highway Traffic Safety Administration’s Office of Vehicle Automation Safety had been let go.

The tension between Musk and the NHTSA is hardly a secret. After all, the safety agency is currently investigating Tesla vehicles in eight active cases, and it is responsible for regulating self-driving technology. Musk has long promised that Tesla will deploy millions of robotaxis across the United States, a plan that would be more achievable with fewer regulatory hurdles standing in the way.

Read: Tesla Threatens Lawsuit After Canada Freezes $30M EV Rebates Over Shady Sales Claims

Speaking with The Financial Times, insiders revealed that approximately 30 people were fired from the NHTSA in February. It’s believed that those working in the NHTSA’s vehicle automation safety team were “disproportionately affected.” While it’s unclear precisely how many were fired from the specific department, DOGE cited poor performance in axing the jobs of these employees.

Despite the controversy surrounding the layoffs, it’s important to note that the NHTSA’s vehicle automation safety division was only formed in 2023. As a result, many employees were still in their probationary period, making it easier for the agency to terminate their employment. The total layoffs at NHTSA accounted for 4% of its workforce, which may seem small, but it raises concerns from some corners of the industry.

Potential Conflict of Interest

One unnamed Tesla manager expressed concern that the layoffs could weaken the NHTSA at a time when it needs more resources to tackle the challenges of autonomous vehicles. “Letting DOGE fire those in the autonomous division is sheer madness—we should be lobbying to add people to NHTSA,” the manager told the FT. “They need to be developing a national framework for autonomous vehicles, otherwise Tesla doesn’t have a prayer for scale in FSD or robotaxis.”

A former NHTSA employee also weighed in, noting, “There is a clear conflict of interest in allowing someone with a business interest influence over appointments and policy at the agency regulating them.”

Xiaomi’s team has been showered with praise in recent months and has probably become used to reading nothing but positive headlines. However, the Chinese carmaker is now hitting the headlines for all the wrong reasons, after the first major accident involving the SU7 sedan resulted in three fatalities. The incident has cast a shadow over the tech giant’s automotive ambitions and raised tough questions about the safety and reliability of advanced driver-assistance systems.

On March 29, a Xiaomi SU7 was traveling in China with three passengers onboard. The vehicle was driving along the Dezhou–Shangrao Expressway in Anhui Province with its Navigate on Autopilot (NOA) system engaged, maintaining a steady speed of 116 km/h (72 mph). As it approached a construction zone and a lane closed off with barriers, the SU7 struck a concrete divider at 97 km/h (60 mph), sparking a fire that ultimately engulfed the car.

Read: Xiaomi Aims To Sell More EVs In China Than All German Brands Combined

Shortly after the crash, Xiaomi released a timeline detailing the moments leading up to the collision. At 10:44:24 p.m., the SU7’s Navigate on Autopilot system issued a warning—“Obstacle ahead”—and began braking. One second later, the driver took control, steering left by 22.06 degrees and applying 31% brake pressure. Another second later, they adjusted slightly to the right by 1.06 degrees and increased braking to 38%. Despite these apparent efforts to avoid the crash, the vehicle struck the concrete barrier.

The impact immediately triggered the eCall emergency system. Within 20 seconds, the registered owner of the SU7 was contacted and confirmed he was not the driver. An ambulance was dispatched and arrived around 11 p.m., but tragically, all three occupants had already died.

The Xiaomi SU7 vehicle in China was driving at 116km/h in NOA on the highway, with the last speed reported to be 97km/h just before the collision with the guardrail.

A fire broke out after the collision, killing all three female college students in the vehicle.

Xiaomi issued a… pic.twitter.com/nxHGGYXhOR

— Tsla Chan (@Tslachan) April 1, 2025

Xiaomi says it is fully cooperating with local investigators to determine the cause of the crash. According to Reuters, which cited reporting from the Chinese newspaper Economic Observer, the father of one of the victims was told by local police that the car key failed to unlock the doors after the impact. In a statement, Xiaomi said it has not yet been able to confirm whether the doors could be opened at the time of the accident.

Shares in the Chinese technology giant fell 5% after word of the crash became public. Company boss Lei Jun has vowed to “respond to the concerns of families and society,” while investigating the crash.

Photos Weibo / Shine

Earlier this week, Mark Rober sparked off a giant online battle by testing autonomous driving tech. In a long video, titled Can You Fool A Self Driving Car?, he set lidar against optical systems like the ones Tesla uses. The result? Instant backlash, praise, confusion—basically, the whole internet lost its mind. Now, someone else has re-run the same test, and, unsurprisingly, the results are both familiar and a little bit different.

In short, lidar tends to see more clearly, and more accurately, in certain situations compared to optical systems. That shouldn’t be much of a surprise, as after all, it’s a high-definition radar system that can sense objects in complete darkness.

Nevertheless, when Rober’s video highlighted Tesla’s failure to detect a wall that looked like a real road, fans of the brand came out with their pitchforks, so to speak. To their credit, Rober’s test didn’t use Full Self-Driving (Supervised) but rather Autopilot.

More: Tesla Autopilot Smashes Through Fake Road Wall While LiDar Lexus Stops Like A Pro

That’s where Kyle Paul, a Tesla owner himself, comes into play. He decided to rerun the same test with the same general parameters, but this time using FSD rather than Autopilot. He printed out his own wall that looked just like the real road it sat on and drove his Model Y up to it multiple times.

In every test, the Tesla failed to see the wall until he was literally inches away from it. As Rober suggested in an interview, it’s plausible that the ultrasonic parking sensors noticed the wall rather than the autonomous driving tech.

That said, Paul then switched things up by bringing in a Cybertruck to run the same test. Interestingly, it passed the test with flying colors by stopping on its own every time it began to approach the wall. What’s the difference? Other than the obvious fact that FSD is more advanced than Autopilot, the Cybertruck was on Tesla’s latest FSD hardware called HW4. The Model Y, a 2022 year model edition, wasn’t as it was running HW3.

The Missing Pieces

Notably, some in the comment section pointed out the tests that Paul didn’t do. For instance, he didn’t test FSD with a mannequin or in the rain—two factors that could offer a more realistic sense of how the system performs in everyday conditions.

Nevertheless, this should at least help calm the noise around Rober’s video. There’s clearly some truth to the criticisms, and those who continue to challenge Tesla’s approach to autonomous driving aren’t entirely off the mark.

Sceenshot Kyle Paul

Just a month after leading Chinese EV maker BYD announced its new ‘God’s Eye’ suite of driver-assistance systems, the Geely-owned Zeekr has launched its intelligent driving program. Zeekr’s solution, available in several different tiers, will be added to existing models via an over-the-air update over the coming months.

All versions of the system, dubbed G-Pilot, include a General Automated Evasion System (G-AES) and Full-Capacity Vehicle-to-Parking (V2P) intelligent drive capabilities. The G-AES system uses a large AI training model and allows for “continuous evasive maneuvers for unexpected obstacles” at speeds of up to 81 mph (130 km/h).

Read: Zeekr 9X Grand PHEV Is Geely’s $140k BMW And Range Rover Rival

Meanwhile, Zeekr’s V2P intelligent drive system allows their EVs to automatically search for empty parking spaces and park autonomously. The upcoming Zeekr 9X Grand will build on these systems with even greater capabilities.

The G-Pilot system will come standard in the company’s upcoming flagship luxury SUV and will be Level 3-capable self-driving and is powered by dual Nvidia Drive AGX Thor chips. Much like Tesla’s assisted-driving systems, Zeekr says its G-Pilot suite will evolve and improve throughout the lifespan of its EVs. This Level 3 system will also allow the car to drive itself from one destination to another, without driver input.

Zeekr 9X

While many details about the 9X remain a mystery, recent teasers show it will be a large SUV similar in size to the Range Rover and Rolls-Royce Cullinan. Not only will it be the brand’s largest and most luxurious model, but it will also be the first to use a plug-in hybrid powertrain. Little is known about this powertrain, but it’ll likely include a spritely gasoline engine and at least two electric motors and could be identical to the mechanically related Lynk & Co 900.

In the base Lynk & Co 900, a 1.5-liter turbocharged engine is paired with a 215 hp front motor and a 308 hp rear motor, making for a combined 710 hp. A pair of 2.0-liter turbocharged four-cylinder PHEVs are also offered, with one offering a combined 724 hp and the second an even more impressive 845 hp.