Tesla users are about to receive a significant FSD update. Musk was recently asked on social media when he could try FSD V12 autonomous driving, and he replied, "About two weeks.
Perhaps influenced by this news, Tesla's US stock market rose by over 4% last night.
It is worth noting that CITIC Securities' October research report predicted that FSD V12 is expected to be launched early next year. Therefore, if FSD V12 can achieve user trial in the next two weeks as mentioned by Musk, the landing progress may exceed expectations.
Tesla FSD V12 is its most important upgrade, achieving full AI end-to-end driving control, inputting sensor signals such as vision, and outputting vehicle control signals such as steering, braking, and acceleration, improving the performance of autonomous driving algorithms. On November 10th, Huatai Securities reported that end-to-end architecture has a higher upper limit, and when the model training reaches a critical point, it will continue to improve its level and surpass existing technological paradigms.
In August, Musk used a Model S equipped with HW3 to showcase the Tesla FSD V12 test version live to the outside world. In this 45 minute live broadcast, Musk, sitting behind the steering wheel and holding his phone, only intervened once in vehicle behavior. When the vehicle is on a non predetermined road, it can yield to pedestrians, avoid roadblocks, turn at intersections, and choose one of the two straight lanes with fewer vehicles.
Since 2020, Tesla has shifted Autopilot decision-making from programming logic to neural networks and AI. After three years of development, it can be seen from Musk's FSD V12 live broadcast that almost the entire decision-making and scene processing has shifted to Tesla neural networks and AI.
There are over 300000 lines of C++code in the FSD V11 exclusive control stack, while there is very little code in V12. Previously, Musk also pointed out that vehicle control is the last piece of puzzle on the Tesla FSD AI puzzle, which will reduce the amount of C++code over 300000 lines by approximately 2 orders of magnitude.
Nowadays, Tesla continues to increase its investment in AI, and the scale of training computing has more than doubled compared to Q2. It can continuously improve training speed and FSD experience. On November 12th, Guojin Securities reported in a daily report that the FSD Beta has traveled over 500 million miles, and the FSD V12 continues to advance. It will adopt a pure vision radar free solution to achieve end-to-end autonomous driving. Under the support of the innovation cycle, the company is expected to hit a bottom and rebound. At the same time, Cyberrack will start delivering at the end of November, and the company's new product cycle is approaching, with both growth and explosiveness. It is recommended to continue to pay attention to Tesla and its supply chain.
In addition, Tesla will launch an entry-level electric vehicle with a cost of approximately 150000 RMB. Kaiyuan Securities added that all Tesla models are equipped with FSD software and hardware as standard, which is expected to further sink FSD functions and drive the penetration rate of self driving functions to further increase.
Overall, Huatai Securities believes that intelligence will be the main investment focus for 2024, with intelligent driving being the key. In the current new stage of intelligent driving development, OEMs are increasing their investment in intelligent driving, focusing on the development of software algorithms, and strengthening their overall discourse power. They are most directly benefiting from the rapid increase in sales of intelligent driving models and the resulting economies of scale; In terms of components, companies that can play the role of Tier0.5 are more optimistic, as well as manufacturers with localization capabilities in the field of chassis segmentation.
On October 25th, Zhejiang Securities proposed to focus on two main investment themes: 1. Intelligent driving incremental market: including domain controllers, cameras, AR-HUD, LiDAR, chips and other fields; 2. Intelligent driving leads the entire vehicle supply chain: core suppliers such as Tesla, Huawei, and Xiaopeng.