Tesla Stumbles in China: Self-Driving Tech Lags, Data Security Hinders Growth

Tesla Stumbles in China: Self-Driving Tech Lags, Data Security Hinders GrowthAugust 20 - Tesla is facing headwinds in China. Its advanced driver-assistance feature, Full Self-Driving (FSD), has yet to be approved, putting it at a disadvantage against Chinese electric vehicle (EV) manufacturers

Tesla Stumbles in China: Self-Driving Tech Lags, Data Security Hinders Growth

August 20 - Tesla is facing headwinds in China. Its advanced driver-assistance feature, Full Self-Driving (FSD), has yet to be approved, putting it at a disadvantage against Chinese electric vehicle (EV) manufacturers. Meanwhile, homegrown brands are already selling vehicles that can autonomously navigate city streets, leveraging more advanced technology and lower prices.

China, Tesla's largest market outside of the U.S., was seen as a key area for FSD adoption and user growth. However, Chinese regulators remain concerned about FSD-related accidents in the U.S. and data security issues, impeding its approval. Tesla's struggle to gain permission to send Chinese user data to its U.S. servers for AI training further hampers its development.

FSD is already operational in the U.S., and Tesla CEO Elon Musk predicted Chinese regulators would approve it by year-end. However, sources familiar with the matter say Chinese authorities remain cautious about accidents involving Teslas software in the U.S. and data security concerns, and no clear timeline for approval has been given.

Facing an uncertain approval outlook, Tesla is considering further investments in infrastructure, including computing power, in China. Meanwhile, Chinese automakers have secured government approval and launched more advanced and affordable driver-assistance features than those currently offered by Tesla in China.

XPeng Motors last month rolled out its nationwide driver-assistance system, which features AI-powered city driving capabilities and traffic signal recognition. Other Chinese companies, including Huawei and Li Auto, have also introduced similar functionalities in hundreds of Chinese cities.

Sources say that if Tesla receives approval for FSD services in China, it might initially be limited to a few pilot cities like Shanghai. This aligns with Chinas regulatory approach, often requiring local companies to pilot their services in limited areas first.

Tesla's current driver-assistance software in China costs around $4,500, while Chinese competitors often charge lower fees or even offer the software for free. Analysts note that this price difference makes Tesla's software less appealing to Chinese drivers.

Tesla is already lagging behind in China, said Zhang Yu, general manager of Shanghai-based consulting firm AutomotiveForesight, while comparing driver-assistance features currently available to Chinese drivers.

Tesla did not respond to a request for comment. The company states on its Chinese website that it needs to accumulate billions of miles of driving data and obtain regulatory approvals before offering the advanced features available in the U.S. market to Chinese consumers.

The development of self-driving cars is crucial for Musk and Tesla. Anyone who does not believe that Tesla will solve autonomous driving should not own Tesla stock, Musk said during the recent earnings call.

Some Chinese automakers admit that they may be outmatched by Tesla's advanced software if it gains approval. "Tesla is as much as two years ahead of Chinese companies in driver-assistance technology," said Gu Junli, a former technical executive at both Tesla and XPeng Motors, at an AI conference in July.

However, Tesla's struggles in China are not solely due to lagging technology, as its sales are also declining. China accounted for 20% of Tesla's total revenue in the first half of this year, but, like other American and European automakers, Tesla's sales in China are declining. According to the China Passenger Car Association, Tesla's share of the Chinese new energy vehicle market declined to 6.8% in the first half of the year, down from 9.5% in the same period last year.

Tesla says its FSD version 12 features AI and learning tools that mimic the human brain's neural network and are regularly updated with new sub-versions. This version has been available to U.S. drivers since early this year, allowing vehicles to drive themselves in most scenarios, including on city streets, without requiring human intervention for steering or acceleration. However, Tesla still warns drivers to remain ready to take over control.

In contrast, Chinese regulators have not yet permitted vehicle companies to offer systems that allow hands-free driving. Tesla has adopted an entirely new AI-based software architecture, abandoning its previous approach that used complex rules-based programming to try and anticipate all possible scenarios a car might encounter. The company announced in March that its latest system has been trained on 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. This similar competition in the EV space is helping Chinese automakers advance to the global forefront. The beauty of this market is that because it's so diverse and competitive, it encourages more experimentation and iteration, accelerating the whole industry's progress, notes UBS analyst Gong Min.

Musks visit to Beijing in April, where he garnered initial support for Tesla's FSD rollout plans, significantly boosted the companys stock price and paved the way for Tesla to begin testing vehicles and collecting data needed for regulatory approval. In May, authorities in Shanghai allowed Tesla to limitedly transfer data from its Chinese test vehicles to the U.S., although video data could not contain faces or license plates.

Sources say that Tesla is currently testing a small fleet of cars equipped with the latest driver software in Shanghai and is preparing for trials in Hangzhou and Beijing. Additionally, Tesla is actively pursuing permission from Chinese authorities to transfer data from regular road vehicles (not just test vehicles) to the U.S. and has discussed methods to filter sensitive data.

However, the future remains uncertain. Faced with the challenge, Tesla is exploring a backup plan establishing a data center in China to train its driving software on Chinese road conditions. But this plan is not straightforward. Tesla is investigating whether it can procure the most advanced Nvidia chips to power this AI effort, but the U.S. currently restricts exports of these chips to China.

In the U.S., federal regulators have investigated hundreds of accidents involving Tesla vehicles using earlier driver-assistance software, including some that were fatal. Last December, Tesla recalled over 2 million vehicles to add more alerts and controls to increase driver awareness. Sources say these accident reports have significantly influenced Chinese regulators scrutiny of Teslas technology.

Tesla's AI-powered driving system has raised additional concerns among Chinese regulators, primarily due to fears that the system might not be effective in handling China's complex traffic environment without sufficient local road training. China's unique driving conditions, such as unpredictable road conditions in smaller cities, dense bicycle traffic, and pedestrian activity, all differ from training conditions in the U.S., fueling regulatory concerns.

XPeng Motors founder He Xiaopeng, after experiencing Tesla's latest features in California in June, said he was impressed but noted that the road conditions in Chinese cities are ten times more complex than in the U.S.

Teslas struggles in China reflect its global challenges. Despite its technological leadership, Tesla still needs to overcome numerous hurdles, including data security, regulatory approval, and local market competition, to ultimately achieve its self-driving ambitions.