Data regulation is a key issue facing the
US electric vehicle giant in China, according to experts
·
Tesla’s
approval for higher-level autonomous driving in China remains uncertain, with analysts pointing to regulatory and
data-related obstacles.
·
Data
regulation is the key challenge.
Data collected by Tesla vehicles in mainland China cannot currently be sent
overseas, potentially limiting Tesla’s ability to train and improve its
autonomous-driving models.
·
China
has introduced its first mandatory standards for L3 and L4 autonomous driving, which will take effect on July 1, 2027.
·
Under
the new standards, automakers
would be responsible for accidents when autonomous systems are active.
·
The
rules also require autonomous-driving systems to provide drivers with at least 10 seconds to take control
before a potential accident.
·
Tesla’s
current FSD is classified as L2-plus,
meaning it does not yet meet China’s requirements for L3 or higher autonomous
driving.
·
L3
allows “hands-off” driving
under specified conditions, while L4
enables “minds-off” driving, with the system taking much
greater responsibility.
·
Analysts
believe Tesla may initially prioritise
obtaining full FSD entry into China rather than immediately seeking L3/L4
certification.
·
Tesla
announced in May that “FSD
Supervised” had become available in China, representing a step
towards broader market entry.
·
China
has already granted L3 licences
to two models from state-owned Changan Automobile and BAIC.
·
China’s
L2 autonomous-driving
market is highly competitive, with companies such as Xpeng and technology suppliers such as Momenta investing
heavily in R&D.
·
Tesla
faces a data-training disadvantage in China: its driving-assistance system has limited
familiarity with mainland Chinese roads because of insufficient locally trained
data.
·
Tesla’s
autonomous-driving technology currently depends heavily on its advanced computing and training
infrastructure in the US, creating difficulties because Chinese
vehicle data cannot simply be transferred overseas.
·
Analysts
suggest Tesla’s licensing prospects will depend heavily on whether it is
willing to build
additional computing and AI-training capacity within mainland China.
Tesla’s
challenge in China is not simply technological—it is fundamentally regulatory
and data-related. To
secure approval for higher-level autonomous driving, Tesla may need to localise
data processing, computing
and AI training infrastructure, while adapting its FSD
technology to China’s stricter autonomous-driving standards.
The
likelihood that US electric vehicle (EV) giant Tesla will obtain Beijing’s approval
for higher-level autonomous driving remains a distant prospect, according to analysts.
Though
Tesla recently hired people to test its Full Self-Driving (FSD) system in China,
many obstacles remained, with data regulation a core issue, they said.
Earlier
this month, China outlined its first mandatory standards for autonomous driving
systems, setting requirements for vehicles that feature level 3 (L3) and level 4
(L4) self-driving functions.
The
standards, set to come into effect on July 1, 2027, would hold carmakers accountable
for accidents that happen when the system is active and require systems to leave
at least 10 seconds for the drivers to take control of a car before a potential
accident, which was stricter than the global standards set by the United Nations
in June, according to China’s Ministry of Industry and Information Technology.
“The
foundation [of the approval] would be the regulations over data,” said Xing Lei,
an independent analyst and founder of automotive platform AutoXing.
L3
capability refers to “hands-off” driving in a conditionally automated vehicle, while
L4 is defined as technology where drivers can be “minds-off”, according to a set
of standards published by SAE International.
Tesla’s
FSD is currently at the L2-plus level, which means the US giant would not yet meet
China’s latest requirements, Xing said. However, Tesla may prioritise its FSD full
entry into China over L3 or higher approvals, according to Xing.
In
May, Tesla announced that “FSD Supervised” was now available in China, one step
closer to a full entry.
Beijing
granted L3 licences to two models by the state-owned Changan Automobile and BAIC
in December.
Currently,
the L2 market in China is highly competitive as carmakers like Xpeng, alongside third-party suppliers like Momenta, double
down on research and development.
The
L2-plus level handles driving on both highways and city roads, Navigation on Autopilot
(NOA), automatic on- and off-ramp manoeuvres and memory parking, while still being
classified as L2, where drivers are held fully responsible in a partially automated
vehicle.
Billionaire
investor Duan Yongping, whose high-profile portfolio includes shares in PDD Holdings
and Pop Mart, described his Tesla driving experience in California, where autonomous
driving has been permitted.
“It
works great”, he wrote on his Chinese social media account. Duan had been a long-time
sceptic of Tesla CEO Elon Musk.
Tesla’s
driving assistance function was not familiar with the roads on mainland China due
to insufficient data training, said Yale Zhang, managing director at the consultancy
Automotive Foresight in Shanghai.
Data
collected in mainland China is not yet allowed to be sent overseas, Zhang explained,
and Tesla’s technological dominance currently relies on its advanced model that
runs in the US.
“The
key to Tesla’s licences in China will be subject to whether it has the determination
to add compute and training to its data centre in the mainland,” he noted.