When will Tesla FSD enter China? And how will it land after entering? It has always been a hot topic in the industry, and there is almost always new news coming out every once in a while.
Recently, there has been new progress on this topic.
On June 14th, it was reported that an official from the Shanghai Lingang Commission revealed that 10 Tesla FSDs are being tested in the area. However, the day after, the screenshot of the official's statement was spread online, stating that he did not understand the relevant information.
The reporter contacted Tesla China-related personnel but did not receive a response as of the time of publication.
However, according to a report by Caixin on June 16th, quoting insider information, the Shanghai Autonomous Driving Demonstration Zone has issued a road test license to Tesla, and FSD may be testing.
An industry insider speculated when communicating with Titanium Media App, 'This news may be released on August 8 when Tesla officially announces Robotaxi landing.'
Two key issues for the entry of FSD into China
FSD, short for Full-Self Driving, is an automatic driving system developed by Tesla.
In 2021, Tesla officially released the FSD test version (Beta version). Although FSD is launched globally including China, Tesla does not allow its vehicles to enter FSD Mode in all places around the world.
Currently, Tesla FSD is only available in the United States, Canada, Australia, and some European countries. In March of this year, Tesla pushed the latest FSD V12.3.1 version to North American car owners. In June, it continues to push the latest FSD V12.4.1 to its employees and plans to open the test to 'a small number of external users.'
In China, Tesla cars are equipped with a driving assistance system called Autopilot, but the FSD function is not complete. According to Titanium Media App, the FSD function in China is priced at 64,000 yuan, while the already activated FSD package in the United States is priced at $8,000, and the monthly subscription fee is priced at $99.
Since this year, Musk has released information that FSD will soon enter China on multiple occasions. On April 20th, Musk revealed on a social platform, 'FSD may soon enter China.' A few days later, when Musk 'flash' visited China, there was news that Musk's visit to China was aimed at promoting Tesla's FSD system to land in China and seeking Chinese approval.
In fact, there are two core issues that must be resolved for FSD to land in China: first, how to obtain basic geographic information data? Second, how to collect new data and train it, and how to ensure data security compliance?
Compared with the traditional ACC adaptive cruise control, automatic driving not only needs to use navigation maps but also collects related road information. Therefore, according to Chinese regulations, all automatic driving systems must obtain a surveying and mapping qualification before driving on public roads, and foreign car companies need to cooperate with Chinese companies that have obtained qualifications to obtain surveying and mapping qualifications.
To solve this problem, Tesla has recently reached a cooperation agreement with Baidu. The latter will allow Tesla to obtain a map license for collecting data on China's public roads. At the same time, as part of the agreement, Baidu will also provide Tesla with its lane-level navigation system.
What about the second problem? From the current information, it is still in progress.
During the Beijing auto show, when Musk visited China, the China Association of Automobile Manufacturers and the National Computer Network Emergency Technical Handling Coordination Center released the 'Notification on the Inspection Results of Four Safety Requirements for Automobile Data Processing (Batch One)', which mentioned that all models produced by Tesla's Shanghai Super Factory are fully compliant: anonymous processing of external facial information, default non-collection of cabin data, in-car processing of cabin data, and significant notification of personal information processing.
Previously, Tesla's data security in China has been questioned, and the content of this notification has released positive information to a certain extent.
Immediately afterwards, Reuters mentioned in a report in May that multiple insiders revealed that in order to promote Tesla's global deployment plan for FSD, Tesla plans to collect data in China and establish a data center in China to process data and train algorithms for automatic driving technology. Insiders also said that this is part of Musk's strategic shift.
Will there be a 'catfish' effect happening again?
Although there is still no precise information on when Tesla's FSD will enter China, the industry knows that this "big fish" is really coming. But can this "big fish" bring huge industry impact? How does the industry view the entry of Tesla FSD into China?
At the 16th China Automotive Blue Book Forum, Momenta CEO Cao Xudong still said that the effect of Tesla FSD's entry into China will be to drive the development of the Chinese auto market and the autonomous driving industry.
Zhou Shengyan, CEO of ZhiJia Technology, also said in communication with Titanium Media App that Tesla's FSD will soon enter China, and there is not much time left for local automakers.
However, Huawei's executive director, terminal BG director, and intelligent automotive solutions BU director Yu Chengdong said that Huawei has enhanced its perception capabilities through the use of lidar, which makes Huawei's intelligent driving system surpass Tesla's performance in the Chinese market.
At the same time, he also welcomed the entry of Tesla FSD into China's market. He believes that competition is conducive to promoting the development of the industry, and can also let consumers understand the performance of various companies' products more clearly.
Looking back on Tesla's FSD from its initial launch to now, it has iterated to version V12. The uniqueness of this version is that it not only removes the "Beta" label, but also introduces "end-to-end" technology.
Tesla claims that FSD has replaced 300,000 lines of code with large-scale AI models, realizing the workflow of "input data-processing data directly-output control instructions".
This technology quickly pointed out a direction for the domestic intelligent driving industry and gradually became an industry consensus. Host manufacturers and intelligent driving technology companies represented by Xiaopeng Automobile, Li Auto Inc, and Yuanrongs Startup have successively entered this field.
Xiaopeng Automobile is the first domestic automaker to release an end-to-end model for mass production. It is reported that Xiaopeng's end-to-end model includes the neural network XNet+, the rule control large model XPlanner, and the large language model XBrain.
At the 2024 China Automotive Chongqing Forum, Li Auto Inc Chairman Li Xiang also revealed that it will push a 3 million clip-trained end-to-end+VLM autonomous driving system to test users. It is expected that Li Auto Inc will launch a more sophisticated autonomous driving system trained on more than 10 million clips from the end of this year to early next year.
Elemento has also showcased the upcoming mass-produced high-end intelligent driving platform DeepRoute IO and the end-to-end solution based on DeepRoute IO at the Beijing Auto Show in April.
The reason why Tesla's FSD is called "big fish" is that on the one hand, it has started early, and on the other hand, compared with the end-to-end that is now commonly referred to in the industry, Tesla's end-to-end still has its advantages.
For example, the database is large enough, and it has already run 1 billion miles worldwide. There are already more than 10 million real driver decision-making video clips, which are fed as "fodder" to the large models of FSD neural networks for training.
In addition, after Dojo matures, Tesla also joined the computing power "arms race." By February 2024, Tesla's computing power scale will enter the top five in the world, and Tesla's total computing power scale will reach 100 EFlops by October 2024.
Wu Gansha, Chairman and CEO of Idriverplus, said at the 16th China Automotive Blue Book Forum that in the past two years, Tesla has invested in the high billions of dollars to improve FSD's computing power level, training data, and on-board computing power. If it fails, Tesla may face enormous pressure from the stock market. If successful, it may leave behind small and large companies on the autonomous driving track.
It is not difficult to see that end-to-end is not just a concept, but also requires real data, computing power, and algorithm investment, and none of them can be missing.
Although Chinese automakers have made remarkable achievements in the era of electrification, the ability to continue to advance vigorously in the era of intelligence, especially high-end intelligence represented by autonomous driving, is still a big unknown. Intelligence, especially high-end intelligence represented by autonomous driving, is not only about manufacturing capabilities, but also requires stronger cross-border integration capabilities, and is closely related to the integrated circuit, big data, and artificial intelligence industries. Chinese automakers really need a "big fish" like Tesla FSD V12 to "drive" and promote.
At present, the elimination game among automakers is becoming increasingly fierce. Many automakers have placed intelligent driving technology in a prominent position in their publicity. Whether it is from real accidents on the road or complaints about intelligent driving technology on social media, such as inability to recognize parking spaces, poor road recognition capabilities, and car black screen crashes, etc., the problems that exist can be felt.
We also look forward to the arrival of Tesla FSD, which can bring positive inspiration and guidance to domestic car companies and third-party intelligent driving suppliers on the one hand, and to force domestic autonomous driving technology to truly advance on the other hand, exposing the true nature of car companies that sell cars by riding the wave of intelligent driving hype.
Edited by Jeffrey