Discussing Li Auto's AI strategy

Li Auto is set to increase its investment in AI, and the scope has expanded compared to the NOA autonomous driving. The complete AI landscape will only be revealed after this cross-day press conference concludes.

At this point, the two important directions in this stock's hosting position have converged. Today, the Hong Kong stock market has not opened, and the US stock market's night trading rose by 2 points, which doesn't really indicate much; this is my understanding.

The most direct question is whether there is a possibility of stimulating and driving the stock price in the short term? The short-term bottom for Hong Kong stocks is around 80+, which should be relatively solid. A week before the press conference, the stock price has already partially reflected this, reaching 90+. Is it possible to push it up further, say to 120? I think we shouldn't have too high expectations. Simply put, while the long-term prospects for AI are indeed high, the competition is equally fierce, and the capital investment is substantial, with high risks. The product concept is good, and the cognitive ability is strong; they have made a remarkable turnaround in NOA AI technology over the past year, which is commendable, but it is still not enough for the capital market (including myself) to immediately have a certain expectation for Li Auto in the AI sector.
If we must talk about stimulating the stock price to rise, I think we need to see the content of the next two events, whether they can better stir market expectations and guide the market's perception to establish a stronger connection between Li Auto and Tesla, with valuation benchmarks against Tesla being the only logic that can significantly stimulate the stock price in the short term. If they can take advantage of the situation and strengthen AI awareness among potential users, it would also be better to enhance the traffic for electric vehicles.
How likely is Li Auto to succeed in AI? I am cautiously optimistic. The optimism mainly stems from the fact that the entire AI sector presents a certain opportunity at the level of the internet and mobile internet over the next 10 years, which is irrefutable. The cautious aspect includes: compared to B2B, the clear industrial logic for B2C has yet to emerge. Even with current domestic ceiling products like Doubao, among my potential seed user group, there are very few users who use it for more than 10 minutes a day. The stickiness of usage and market indicators do not even come close to being phenomenal; whether the determination to invest is sufficient is crucial, as this is a path requiring heavy capital investment. In the short term, it is indeed impossible to match the funding input and talent density of ByteDance, but it is sufficient to benchmark against Kimi and Deepseek. There is no need to blindly trust background theories; despite the criticism of Baidu, let’s not forget that there is also Baichuan, which has a similar starting line halo.
Specifically regarding intelligent driving, how is AI? Li Auto is daring to think and act, and shareholders consistently give a thumbs up. In 2024, this local battle for NOA is a complete victory. End-to-end + VLM, the former allows for overtaking in curves, while the latter positions itself as completely leading in the technological path. This press conference, or the surrounding interviews, should release some technical information in this regard (strengthening AI awareness). If I’m not mistaken, in 2025, there should be a significant upgrade in the architecture of these two large models, transitioning from the current glue dual-core to an organic dual-core Just like the dual-core upgrade process of Intel back in the day. This reflects not only innovation in algorithms but also innovation in engineering. The team behind it has expertise across the supply chain, manufacturing, vehicle systems, and AI algorithms.
Don't forget, although Tesla has become increasingly secretive about the details of its technology roadmap after V12, based on various expert interviews, Tesla's V13 also introduces an LLM dual-core architecture on an end-to-end model. Where is Huawei? Where is XP?
Finally, why do this? Building cars is essential, but without carving out a differentiated path, this year it's surrounded by Seres, and next year it will be encircled by a circle of Xiaomi.

Conflicts of interest: This stock holds 10% of $Li Auto(LI.US)$LI AUTO-W(02015.HK) good position, and there may be trading plans in the next week