Deji Osas
US electric vehicle leader Tesla has started operating a China-oriented artificial intelligence training centre aimed at improving its assisted and self-driving technologies, as competition intensifies in the world’s largest EV market.
Confirming the development, Tesla vice-president Grace Tao said the new facility has enough computing capacity to support the advancement of the company’s assisted-driving systems. Speaking to Shanghai-based financial outlet Cailian on Friday, Tao declined to disclose the centre’s location or the scale of Tesla’s investment.
The move is widely seen as a critical step toward expanding and commercialising Tesla’s Full Self-Driving (FSD) navigation software in mainland China, following recent regulatory easing by Chinese authorities. It also comes at a time when domestic automakers are accelerating plans to roll out vehicles with level 3 (L3) autonomous driving capabilities.
Industry observers say the new hub positions Tesla more strongly in a market where local manufacturers are moving quickly. Yin Ran, a Shanghai-based angel investor, noted that growing rivalry between Tesla and Chinese brands would benefit consumers.
“With thousands of L3-capable vehicles expected on Chinese roads by 2026, competition will intensify as automakers race to deliver reliable and cost-effective autonomous driving solutions,” Yin said.
Tesla’s FSD system is built on neural network technology trained using real-world driving footage, allowing vehicles to make decisions that mimic human driving behaviour. According to David Zhang, secretary general of the International Intelligent Vehicle Engineering Association, the system has long been regarded as one of the most advanced globally.
Tesla introduced its FSD software in mainland China in February last year, pricing it at 64,000 yuan (about US$9,200). However, the company faces regulatory constraints, including restrictions that prevent driving data collected in China from being transferred to the United States, while US regulations also limit Tesla’s ability to train its AI systems within China.
The new AI centre is expected to help Tesla fine-tune its software to better recognise local road conditions, such as lane markings, traffic signals and driving patterns unique to Chinese cities, Zhang added.
Under international standards set by SAE International, L3 autonomous driving allows a vehicle to handle most driving tasks independently, though drivers must remain alert and ready to take control when required.
China took a major step toward legalising L3 technology in December, when the Ministry of Industry and Information Technology approved state-owned automakers Changan Automobile and BAIC to develop L3-capable vehicles. Since then, dozens of Chinese carmakers—including industry leader BYD and tech firm Xiaomi—have been conducting extensive road tests in preparation for mass production.
Many of these companies already offer advanced driver-assistance systems comparable to Tesla’s FSD, which have proven popular with tech-savvy Chinese consumers.
Despite its technological edge, Tesla’s market position in China has weakened. Data from the China Passenger Car Association shows the company delivered 625,698 vehicles in China in 2025, a 4.8 per cent decline from the previous year. Tesla ranked fifth among EV makers in the country, trailing BYD, Geely, Changan Automobile and SAIC-GM-Wuling.
China remains the dominant force in the global EV industry, accounting for roughly 70 per cent of worldwide new electric vehicle sales, according to CPCA figures—making Tesla’s push to localise its self-driving development increasingly crucial to its long-term strategy.
