FSD, ushering in the second half of intelligent electric vehicles?

Although there is no clear timetable yet, the entry of Tesla’s most powerful intelligent driving system FSD into the Chinese market is just around the corner. According to news from The Paper in mid-May, the leaders of Shanghai Municipal Commission of Economy and Information Technology said during their visit to Tesla's Shanghai factory that Shanghai will further deepen cooperation with Tesla, promote the layout of automated driving and other functional modules in Shanghai, which is considered the first signal released by domestic relevant departments to accept FSD.

As a leader in global autonomous driving technology, what does the entry of FSD into China mean for the Chinese electric vehicle market?

CITIC Securities Lianxi team stated in a research report released last week that the overall entry of FSD into China is expected to accelerate the intelligence process of China's electric vehicles. It will promote the elimination of weaker competitors while tightening consumer education and recognition, and expanding the market. If Tesla's "extreme cost-effective hardware + high-gross margin FSD software" model can be widely implemented in China, it may be similar to the "ChatGPT moment" of smart cars. It means new possibilities for current domestic automakers and supply chains, whether in terms of incremental demand or business models.

Compared with the domestic host manufacturers, CITIC Securities found that the main difference between the two lies in the engineering capability of handling large-scale data to establish the "data flywheel". We believe that Tesla's core advantage lies in building a complete industrialized data training process, including data collection, training, labelling, black-white testing and other process loops, as well as basic facilities such as the supercomputer center dojo. Although domestic host manufacturers are also building similar "data loops," it is not easy to build such a highly automated and procedural system, requiring not only an understanding of the framework but also strong engineering practice capabilities and years of polishing. Even leaders in China's intelligent driving self-developed ability such as XPENG-W and Huawei may lag behind Tesla by 2-3 years in terms of "data loops." In addition, only from the perspective of data volume, China's domestic automakers with relatively prominent intelligent driving self-developed ability are not currently comparable to Tesla in terms of automobile sales and fleet size. Therefore, once FSD lands in China and the "flywheel" starts running, the difference between Tesla and domestic manufacturers in terms of "data loops" will become more apparent, and the time window left for domestic manufacturers will quickly narrow.

As background, Tesla's automatic driving products are currently divided into three categories: AP, EAP and FSD. The basic version of AP and the enhanced ADAS system EAP can be used in China. FSD is the highest-level, whose primary features include NOA, automatic lane changing, automatic parking, intelligent summoning, traffic signal recognition, and (navigation-based) automatic turning on city roads, and it is currently only available in the beta version in North America. Tesla CEO Musk said at the shareholder meeting last month that FSD will definitely be 10 times safer than human drivers in the future, but he also admitted that it will take some time to achieve true completely autonomous driving.

CITIC Securities believes that FSD's entry into China will generally improve the intelligence process of China's electric vehicles and accelerate the elimination of weaker competitors while tightening consumer education and recognition and expanding the market. In terms of consumers, the consideration of intelligent driving is relatively low when buying a car. Tesla, as an industry leader, has a major role in market education, as evidenced by its push into China's pure electric vehicle market. Therefore, CITIC Securities believes that FSD’s entry into China may significantly strengthen Chinese consumer awareness of vehicle intelligence, while fostering consumer's daily demand and usage habits for intelligent driving, and China's pre-installed smart driving market is expected to expand significantly. From the perspective of China's host manufacturers, we believe that they will benefit from the increase in consumer awareness of intelligent driving for high-end models that emphasize intelligent driving capabilities. The high selling price of FSD will further broaden the imagination of Chinese automakers. However, considering that there are significant differences in road environments and driving habits between China and the US, FSD may not be immediately successful in China, and it may take some time to establish a localized "data loop" and adapt the algorithm model. Based on this reason, CITIC Securities believes that domestic Chinese host manufacturers still have a certain first-mover advantage, and it is crucial to accelerate the development of intelligent driving products and position their meaning for consumers in the remaining time window.

For lower-end models that have been hit hard by price wars, if Tesla introduces FSD with its extremely high gross profit into China, it will not rule out the further reduction of the prices of whole cars, including the release of lower-priced Model2/ Q models (priced or in the range of RMB 150,000 to 200,000). If the "extreme cost-effective hardware + high-gross margin FSD software" model can be widely implemented in China, the cost pressure on lower-end models will be further increased, and offering a low-priced but qualified intelligent driving product through third-party suppliers may become the mainstream choice for host manufacturers with weak self-developed capabilities. However, this may also weaken their chances of obtaining premiums in the new intelligentization race. Earlier this month, Tesla opened up FSD for use by other automakers. Musk believes that once FSD's model scale and data accumulation exceed a certain threshold, it will also begin to erupt like ChatGPT. Almost instantly, 3 million, 5 million, or even 10 million vehicles will achieve full autonomous driving.

If Tesla achieves complete autonomous driving, vehicles can be shared among many users, greatly improving the efficiency and utilization of vehicles. Tesla estimates that this can raise the value of a car by five times.