August 21, 2024 – Tesla is facing challenges in the Chinese market, where its most advanced driver assistance features have encountered difficulties in gaining approval, putting the company at a slight disadvantage compared to local electric vehicle manufacturers. These domestic producers have already begun selling vehicles capable of autonomous navigation on city streets.
China represents Tesla’s largest market outside of the United States, and the company has been eager to see widespread adoption of its “Full Self-Driving” (FSD) software among local drivers. This software promises fully autonomous driving from start to destination, although it still requires the driver to be prepared to intervene at any time. CEO Elon Musk predicted last month that Chinese regulatory bodies would approve these features, already permitted in the US, by the end of this year.
However, according to sources familiar with the official thinking, the Chinese government has not yet approved Tesla’s latest driving technology due to concerns about accidents caused by the software in the US and data security issues. Additionally, Tesla has not obtained permission to transmit Chinese vehicle owner data to US servers for training its AI models. The uncertainty surrounding approval has prompted Tesla to consider further investing in computing infrastructure in China.
Meanwhile, some Chinese automakers have obtained government approval for their driving assistance features, which are not only more advanced than what Tesla can currently offer in China but also priced lower. For instance, Xiaopeng (Xpeng) launched a nationwide driving assistance system last month, capable of navigating city streets and responding to traffic signals. The system is driven by artificial intelligence, according to Xpeng.
Other Chinese companies, including Huawei and Li Auto, also provide similar driving assistance features in hundreds of Chinese cities.
If Tesla gains approval for its FSD service in China, it may initially be limited to a pilot program in a few cities like Shanghai, following the common practice of Chinese regulators who often require local companies to first test their services in limited areas.
Currently, Tesla’s assisted driving software available in China costs around $4,500, while Chinese competitors typically charge lower fees or offer the software free of charge to car buyers. Analysts suggest that this price difference makes Tesla’s software less appealing to Chinese drivers.
“Tesla has already shown a backward trend in China!” said Zhang Yu, general manager of Automotive Foresight, a consulting company in Shanghai, when comparing the assisted driving functions currently available to Chinese drivers.
Tesla did not respond to requests for comment but noted on its Chinese official website that it needs to accumulate billions of kilometers of driving data and obtain relevant regulatory approvals before providing advanced features available in the US market to Chinese consumers.
The development of autonomous vehicles is crucial for Musk and Tesla. “Anyone who doesn’t believe that Tesla will solve the self-driving problem should not hold Tesla stock,” he stated during the latest earnings call.
Some Chinese automakers acknowledge that if Tesla’s advanced software gains approval, it could surpass their own offerings. Gu Junli, a former technical executive at both Tesla and Xpeng, indicated at an AI conference in July that Tesla leads Chinese companies in driving assistance technology by up to two years.
Tesla’s sales in China, which accounted for 20% of its total revenue in the first half of this year, are declining, along with those of other American and European automakers. According to the China Passenger Car Association, Tesla’s share in China’s new energy vehicle market dropped from 9.5% in the same period last year to 6.8% in the first half of this year.
Tesla’s FSD driving software, version 12, employs artificial intelligence and learning tools that mimic the neural network of the human brain, with frequent updates of new sub-versions. Since the beginning of this year, American drivers have been able to use this version, allowing vehicles to drive independently in most cases, including on city streets, without the need for manual steering or acceleration. However, Chinese regulators have not yet permitted vehicle companies to provide hands-free driving systems.
Tesla has adopted a new software architecture based on artificial intelligence, abandoning complex rule-based programming that attempts to preset all possible scenarios a car may encounter. The company announced in March that its latest system had undergone learning and training through millions of video clips, replacing hundreds of thousands of lines of computer code.
Meanwhile, Chinese automakers are actively developing and training their own AI models, with the competition in electric vehicles helping to push Chinese manufacturers to the forefront of global innovation.
“The beauty of this market is that due to its diversity of participants and fierce competition, it encourages more trial and error, which accelerates the progress of the entire industry,” said Gong Min, an analyst at UBS.
During his visit to Beijing in April, Musk gained initial support for Tesla’s FSD promotion plan, significantly boosting the company’s stock price and paving the way for Tesla to begin testing vehicles and collecting data necessary for regulatory approval. In May, Shanghai authorities allowed Tesla to transmit data from test vehicles in China to the US in a limited manner, provided that the video data did not include faces or license plate numbers.
According to informed sources, Tesla is currently testing a small fleet of vehicles equipped with the latest driving software in Shanghai and preparing for trials in Hangzhou and Beijing. Additionally, Tesla is actively seeking permission from Chinese authorities to transmit data from regular road vehicles (not just test cars) to the US and has discussed methods for filtering sensitive data. However, the future remains uncertain.
Facing challenges, Tesla has considered an alternative plan – establishing a data center in China to train driving software under Chinese road conditions. However, this plan is not simple. Tesla is exploring the possibility of procuring state-of-the-art NVIDIA chips to power this AI effort, but the US currently restricts the export of these chips to China.

In the US, federal regulators have investigated hundreds of accidents involving Tesla vehicles using earlier driving assistance software, including some fatal incidents. Last December, to increase driving alertness, Tesla recalled over 2 million cars, adding more alarms and controls. Informed sources indicate that these accident reports have significantly impacted Chinese regulators’ review of Tesla’s technology.
Tesla’s AI-driven driving system has raised additional concerns among Chinese regulators, primarily due to worries that the system may not effectively handle China’s complex traffic environment without adequate local road training. China’s unique driving conditions, such as varying road conditions in small cities and dense bicycle and pedestrian traffic, differ from those in the US, exacerbating regulatory concerns.
After experiencing Tesla’s latest features in California in June, He Xiaopeng, founder of Xpeng, expressed impression but also noted, “The road conditions in Chinese cities are much more complex than those in the US, probably ten times more so.”