China’s EV Self-Driving Is a TOTAL Scam: BYD, XPeng, Huawei, Li Auto, Xiaomi, and NIO All Failed

0:00 On July 23rd, Chinese car website Don't Show a D teamed up with CCTV to create a highly realistic highway smart driving test, which quickly went viral with its 90-minute video. The main highlight of this test was that it was the first time a 15 km stretch of highway and city roads was fully rented out to evaluate the driver assist capabilities of mainstream car models under extreme conditions. The test involved 36 cars including Tesla's Model 3 and Model X, Xpunk G6, WJ M9, Jujuia R7, Avita 12, Lee Auto L6, and other models from Tangu, Wayblue Mountain, Xiaomi, NIO, Leap Motor, Zeer, Leo, and BYD. The test simulated 15 types of high-risisk scenarios, which are rare, but very likely to cause accidents. Hundreds of

0:46 simulated collision tests were performed, fully replicating how the assistive driving systems would react in complex environments. The results showed that the Tesla Model X and other Tesla models ranked highest, while 15 Chinese domestic brands such as Huawei, Xiaomi, and BYD had pass rates as low as zero. Here are the specific test results. One, city scenarios. Nine different city scenes were tested, including entering a large roundabout, merging in the roundabout, four children crossing the road, avoiding a broken down car, making a standard U-turn, dealing with a parked electric bike, parking problems, and hidden cars in the blind spot. A total of 26 cars participated in the city tests with 233 tests conducted. Of

1:32 these, 103 tests passed, giving a pass rate of 44.2%. Tesla's Model X performed the best with eight successful tests achieving a pass rate of 88.9%. The second tier models including the Judia R seven, Avita 12, and Hawi 3X each passed seven times. Highway scenarios. Six types of highway scenarios were tested, such as encountering an accident, meeting a truck on a construction road, sudden construction, a disappearing car, aggressive merging at a highway intersection, and a wild animal crossing the road. 183 tests involved 36 cars. Only 44 tests passed, giving a pass rate of 24%. This result is quite surprising and

2:18 showed a stark contrast from the typical perception of smart driving systems. In one test, the car had to avoid a model of a wild boar crossing the road. Almost all 36 cars failed with only five passing. The test did not only assess whether the system could see the obstacles, but also if the system had multi-layered perception, dynamic recognition, and a complete emergency response chain. Many cars failed this test because the system didn't recognize the wild boar as a high priority object, mistakenly treating it as a plastic bag, ground shadow, or another low priority object. On July 23rd, a car blogger analyzed the results from the smart driving test, which quickly resonated with online users and garnered 47,000 likes in under 24 hours. The test was an embarrassment for the

3:04 car industry. If this test hadn't been done in collaboration with CCTV, the video might have been pulled down by now. Dongched has upset every car company involved by renting an entire highway to evaluate the smart driving system of 36 mainstream car models. The outcome was disastrous even for the so-called top tier brands like Huawei, Tesla, Xiaomi, and Neo. The results were a wake-up call for those who believed in the hype surrounding Chinese car brands. The test included six different projects and not a single car passed all of them. If the test was meant to assess their driving assistance systems, it ended up being a harsh reality check about their collision avoidance capabilities. For an industry professional like me, it was not surprising that cars crashed. The surprise was when some of them managed

3:50 to pass the test. Some cars couldn't even detect the most common obstacles like speed bumps. Instead of slowing down, they crashed straight into them, damaging mirrors, windshields, and bumpers. Some even wrecked the radiators. If the obstacle had been a real car, the consequences would have been unthinkable. With speeds so high, many vehicles didn't even attempt to break when approaching construction zones. They just plowed right through. The best performing cars were the Tesla Model 3 and Model X, which passed five out of six tests. The first test simulated a car suddenly changing lanes when there was a broken down car on the road ahead. Tesla's Model 3 and Model X easily passed, while Xiaomi's SU7 Max detected the car but couldn't stop in time, crashing into it. Huawei's WJ M8

4:37 was blocked by traffic and also failed to avoid the collision. The second test involved driving in pitch black conditions on a highway with no street lights where an accident vehicle has stopped in the middle of the road. This was a test of nighttime recognition. Tesla, which uses a purely visual system with no LAR, performed well and was able to recognize and stop in time. Xiaomi's SU7 Max Elite Auto L6 detected the obstacle but couldn't break in time, which was quite embarrassing. Despite having both visual and LAR systems, they couldn't match Tesla's performance. The overall results were shocking. Many cars had a survival rate of only 17% when tested in high-speed accidents. How can they still claim that their systems are L3 level, comparable to L4? This is

5:22 nothing but a dangerous lie. Many vehicles barely pass one of the six tests with some failing even that. Some cars had a score of zero. How can they still boast about the performance? Tesla 2 didn't pass every test. Model X avoided a sudden animal crossing, but Model 3 didn't. This reinforces a point that real world driving conditions are far more complex and controlled tests. No cars smart driving system can guarantee your safety. The results of this test left users stunned, especially since over the past 2 or 3 years, they had heard numerous claims about the advanced nature of driver assistance systems and witnessed many brands at launch events proclaiming that their smart driving systems had completely surpassed Teslas. For example, in April 2023, the CEO of the terminal BG and CEO of

6:09 the intelligent automotive solutions BU Y chang directly stated at a launch event that Huawei's highle smart driving system had far surpassed new entrance and Tesla. He said the Huawei ADS 2.0 excelled in areas such as irregular obstacle recognition and the frequency of driving assistance takeovers. Yuch Chong Dong compared Huawei system to a certain foreign T model, claiming that Huawei smart driving system could recognize irregular obstacles much better. While the foreign T model would crash at 40 km/h, Huawei system could stop at 80 km/h. He went on to state Tesla does much better than the new entrance and other brands whose systems only perform at 20 to 40% of the capability. Huawei system performing at 80% is considered excellent, but we're at over 98% making us even better. At

6:57 the launch of the WJA M5, Yuch Chong Dong even described Huawei's smart driving system as a never-ending lube. He added, "What's holding us back is that the laws and regulations are not yet fully developed, which prevent us from officially calling it L3, but in reality, we've already achieved it. L3 advanced smart driving refers to a level of autonomous driving where under specific conditions, the vehicle can independently complete driving tasks. for the 2024 WJ M7 Ultra 4WD Smart Driving Edition, Yuchong Dong proudly stated that this car is so hard to crash that I can do whatever I want in the car. However, just a month earlier in April 2024, a WJ M7 collided with a water truck on the highway in Shine, leading to a fire that resulted in three fatalities. Following the incident, the

7:43 victim's families and many users realized that not all 1J models were equipped with a Huawei automatic emergency braking system that you had boasted about. The same unfortunate scenario also occurred with Xiaomi's vehicles. In March 2024, at the launch of the Xiaomi SU7, Xiaomi CEO Le Jun revealed that Xiaomi had tested its smart driving system on real roads for 10 million km. He claimed that during the test, the SU7 would only require driver intervention once every 300 km for highway navigation and once every 100 km for emergency acceleration or braking. He also praised the braking function, claiming that it could successfully detect a stationary broken down car and stop the vehicle at speeds of up to 135 km/h during the day and 120 km/h at night. But reality proved

8:30 otherwise. On March 29th, 2024, a Xiaomi SU7 crashed into a highway guardrail, causing the car to catch fire and killing three people. Xiaomi's data showed that the last recorded speed before the crash was 97 km/h within the range where the AE should have been triggered. According to their claims, in February 2024, BYYD even equipped its lowcost BYD Seagull with a threeey vision system, marketing it as highle smart driving. The systems A, B, and C were all labeled as highle smart driving, leaving the public confused about what exactly constituted true advanced technology. The Dongcha D test has made the public realize an important truth. The current mainstream smart driving systems still cannot replace human drivers, let alone achieve autonomous driving. Regardless of the

9:16 marketing hype, these systems still have many gaps in complex scenarios. Therefore, even when using driving assistance features, drivers must remain vigilant. The fantasy of driving with your eyes closed needs to be thoroughly debunked. The devastating results of this test have sparked heated discussions online. Smart driving at the mobile chip level, just getting a reaction is good enough. In China, they're using open- source versions. They copy car shells at a big screen and a fridge, and only a fool would buy a Chinese electric car. The reason Chinese car companies can't enter the US market isn't because of American pressure. It's because they know if they make cars like this in the US, they'd be sued and go bankrupt in two days. Old Teslas can beat all of these knockoff Chinese electric cars. This is what they call catching up in the curve. Chinese electric cars are pure industrial trash. Not a single one is an exception. After

10:02 the Dongcha smart driving test gained attention online, several car companies responded. Tesla CEO Elon Musk took notice of Dongch's test results. On July 25th, Musk shared a post from the electric vehicle analyst Sawyer Merritt. Musk pointed out that Tesla achieved the highest score in China without local training data. He added that Tesla is now incorporating data from its world simulators and test tracks to achieve a perfect score. Voya Motors Xiaoming Fun stated that Dongu's closed highway test was like a mirror common technical bottlenecks in extreme scenarios, the need to improve high-speed emergency avoidance and the ability of perception systems to identify non-standard obstacles. On July 25th, Harmony Intelligent Mobility Alliance released a Waybo post saying, "We have seen the so-called test on a certain platform,

10:48 but we won't comment on it." Additionally, Idaho Motors and Hemotors also posted on Waybo that they wouldn't comment. This test clearly isn't about ranking the cars. It's about proving whether smart driving is safe and can meet safety standards for public roads. It could also be said that this wasn't a test of who has the smartest driving technology, but rather a life and death exam of which technology is most reliable. The more hardcore tests there are, the less room car companies have for storytelling, and true innovations that prioritize safety will be noticed. In addition to the Dongcha D smart driving test, another shocking incident in the car world made headlines. On July 24th, a video of an electric vehicle driving repeatedly in circles inside the Chungdu airport parking lot went viral. The car had been spinning for hundreds

11:33 of circles which was quite terrifying. The traffic police and police cars arrived and the driver was already dizzy from spinning. The car couldn't stop. It just wouldn't break. In the first video, the footage showed a large crowd gathered around a black BYD Hansown in the parking lot, which was continuously driving in circles. The car's tire screeched loudly as it turned, causing sparks. A video posted by a witness showed that the car would not break properly and had been spinning in circles for about 30 to 40 laps. Another video shows the arrival of fire rescue teams and police. An eyewitness reported that the driver urgently contacted the traffic police and the manufacturer. The car was stopped remotely and when the door was opened, the driver appeared to be in

12:20 shock. The driver was taken to a hospital for emergency treatment. Fortunately, no one was injured at the scene. One witness commented, "I saw the BYYD car spinning out of control with my own eyes. The driver was forced to run around the airport. It was terrifying. Luckily, no one was injured. I hope BYD can solve this problem quickly and give consumers an explanation." One Edison commented, "The driver must have been terrified. The key is that the car wasn't going slow. Fortunately, there was space for him to keep turning in circles." On July 24th, the Satron Airport Police released a statement. Upon investigation, the driver, a 40-year-old man named Jang, had been unable to reach his passenger near Chungdu airport, which made him anxious, and he began driving in circles in the parking lot. After testing, it was confirmed that Jang was not under the influence of alcohol or drugs. Some

13:07 nitizens questioned the official explanation, saying that the reason given was that the ride hailing driver had been unable to contact his passenger, which made him frustrated and caused him to spin in place. But the fact that Tanfu airport is about 2 km away from the parking lot makes this explanation seem ridiculous. Others commented, "If the driver lost control, would he just spin in circles, not hurt anyone, or crash into other cars?" In previous states of drivers losing control due to frustration, none of them handled the vehicles this rationally. This seems more like the vehicle malfunctioned, and the driver tried to control it to prevent it from hitting anyone or anything. Another Nison said, "Please watch the on-site videos before criticizing the driver. You can clearly see that when turning, the car slows down, indicating that the ESP is working. The brake lights don't come on.

13:53 Initially, the brake lights did come on, but after so many circles, the brake lights didn't turn on again. From other videos, you can see the driver pressing the brakes, then accelerating to continue. It's possible that the car system crashed or malfunctioned, causing the car to enter a loop, and despite the driver breaking, it kept spinning. In recent years, domestic electric cars in mainland China have been involved in frequent accidents. However, due to the Chinese government's heavy subsidies and the heavy push of EVs in the international market, negative news about Chinese electric vehicles is often tightly suppressed by both the car companies and the Chinese authorities. In May of this year, a BYD tongue owner posted a video that when driving back home from Shanghai, his car still had 17% of battery left. But the car stopped responding, leaving him stranded just

14:39 500 meters from home. He said sarcastically, "Domestic cars are a disaster." However, the video led to trouble for him as the police detained him for 7 days. More nisonens are concerned that whenever someone complains about quality issues with domestic cars, the car companies are quick to suppress the comments. When it comes to matters involving people's lives, who will stand up for the consumers? [Music]