On September 19, Xiaomi Group (01810. HK) has staged a "roller coaster" market. At the opening of the day, because its Xiaomi Auto announced the recall of nearly 117,000 SU7 standard electric vehicles, the stock price fell nearly 2% at one point. But it didn't take long for the new energy sector to strengthen as a whole, market sentiment picked up, and Xiaomi's stock price slowly rebounded, and finally there was no danger.
This wave of shocks reflects investors' complex mentality towards "technology companies building cars": on the one hand, everyone is optimistic about Xiaomi's "halo" in intelligent driving and technological innovation; On the other hand, in the event of a safety issue or recall, there will be concerns about real risks.
When the technical halo of intelligent driving encounters real safety hazards, how can investors weigh short-term risks and long-term value?
A recall is not just about fixing a piece of software
The recall of Xiaomi cars is not a small problem and has a great impact, which is directly related to driving safety and user trust.
1. Which cars have been recalled? How serious is it?
Model model: Xiaomi SU7 Standard Edition.
Time Frame: Vehicles produced from February 2024 to August 2025.
Quantity: Up to 116,900 units, accounting for about 45% of Xiaomi's total production over the past year and a half.
What's the problem? It mainly focuses on the L2 high-speed pilot assisted driving function (similar to "automatic following + lane change") may fail in extreme special scenarios - such as failure to recognize obstacles, untimely reminders, and improper handling. If the driver does not take over in time, it may cause a collision.
This problem is highly related to the fatal accident of the Anhui Expressway in March this year: a Xiaomi SU7 with assisted driving turned on failed to identify the isolation zone in the construction diversion section and crashed at a high speed of 116km/h, killing three people.
This is already the second large-scale recall of Xiaomi cars since they were launched: the first (November 2024): More than 30,000 vehicles were recalled due to occasional errors in the cloud service of the automatic parking function, causing vehicles to crash into stationary obstacles. In just 10 months, the two recalls were due to the core and most sensitive function of intelligent driving, which inevitably makes people question: Is Xiaomi's R&D and quality control reliable, as a "new car-making force"?
This problem exposes Xiaomi's shortcomings in autonomous driving algorithms: although L2 assisted driving can "help drive", it cannot be completely relied upon; In "extreme situations" such as sharp curves, construction sections, bad weather, and unclear lane lines, the system must be "smart" and "stable" enough; This requires training on a large amount of real road condition data, as well as multiple security backups.
However, as a "cross-border player", Xiaomi's autonomous driving team has only been established for three years, while Tesla has been working for more than ten years, and the data accumulation and experience are far behind. More importantly, the L2 level system is easy to make people "mistakenly think it is safe", and as a result, the driver relaxes his vigilance, and once the system has a problem, it is too late to react. This places extremely high requirements on the design of human-machine reminder and takeover mechanisms.
The market is "quite tolerant" of Xiaomi
Although the recall was quite serious, Xiaomi's stock price did not "crash", but quickly rebounded after falling, and there was no cliff-like plummet. There are several key reasons behind this.
The general environment is good, and new energy vehicles are rising as a whole. On the same day, the entire new energy vehicle sector performed well: the ChiNext New Energy ETF (Huaxia 159368) rose 0.78%; There have been inflows of funds for 10 consecutive days, with a total of more than 600 million yuan.
Industry fundamentals are also strong: in August 2025, China's new energy vehicle exports doubled year-on-year to 224,000 units; In the first eight months of this year, a total of 1.532 million units were exported, an increase of 87.3% year-on-year. This "hot industry" background is like a "buffer", diluting the negative news of individual stocks and preventing market sentiment from getting out of control.
Judging from the trading data, when Xiaomi's stock price fell, there were funds "buying the bottom": Hong Kong Stock Connect data shows that mainland funds bought a net of HK$230 million in Xiaomi on September 19, 40% more than the average daily purchase volume in the past five days. This shows that many mainland investors are still optimistic about Xiaomi's long-term value, believing that this time it is only a short-term fluctuation.
Why can you bear it? Because cars aren't that important yet. At present, Xiaomi cars account for a small proportion of the group's overall revenue (estimated to be less than 5%), and Xiaomi's real main source of money is mobile phones and smart home devices (IoT).
Therefore, even if there is a problem in the automobile business, it will not shake the fundamentals of the entire company, and investors will not be bearish because of this.
The big risk is that users don't trust it
In addition to short-term stock price fluctuations, this recall may make the market rethink: how much are Xiaomi cars worth? In the past, investors were usually willing to give three more "pluses" to the valuation of "technology companies that build cars" like Xiaomi.
Brand bonus: Xiaomi has hundreds of millions of mobile phone users, and everyone thinks that these users may "buy" Xiaomi cars.
Ecological bonus: mobile phones, home appliances, and cars can be connected to achieve "whole-house intelligence".
Technology bonus: Emphasize "software-defined vehicles", such as autonomous driving and intelligent cockpits.
Among them, the technical bonus is the most vulnerable - once there is a security problem, everyone will wonder: can you really do the software well? Can the story of "software-defined cars" still be believed?
Direct cost: loss of real money. It costs money: on average, the recall cost of each car is 500 to 2,000 yuan; Xiaomi involved 116,000 units this time, and the direct expenditure could be between 580 million and 2.34 billion yuan.
What burns more money is the follow-up remediation: to strengthen the algorithm, you have to build more data centers; To test extreme scenarios, you have to buy more test cars and expand your fleet; It is expected that the proportion of R&D investment will increase from the current 18% to 22% in 2025, further compressing the company's profits.
To put it simply: not only do you have to lose money to repair the car, but you also have to burn additional money for research and development.
The bigger risk is that users don't trust it. For Xiaomi, the most feared thing is not losing money, but the user's heart is cold. The survey shows that for every 1% decrease in trust in autonomous driving among smart car users, the number of people willing to buy a car decreases by 1.8%. The main buyers of Xiaomi SU7 are young technology enthusiasts aged 30-35, accounting for 62%. Although they love to try new technologies and are willing to try new technologies, they also pay special attention to whether the brand is reliable.
If the problem is not dealt with this time, or if a similar accident occurs in the future, it may trigger a chain reaction. Money can be earned again, but once user trust is damaged, it is difficult to recover.
Jingtai Observation|Focus on the key points in stages
The recall of Xiaomi cars exposes a fundamental problem: the style of Internet companies "quick trial and error, rapid upgrades" and the automotive industry's "safety first and foolproof" requirements are inherently in conflict. Technology companies are accustomed to "going online first and then optimizing", but cars are a matter of human life and cannot make big mistakes. This cultural "collision" is the biggest challenge faced by cross-border players like Xiaomi.
The performance of the capital market reflects the attitude of "walking and watching": knowing that intelligent driving is risky and cannot be taken lightly; But it also recognizes Xiaomi's long-term potential in ecology, brand and technology, and does not completely deny it because of a single problem.
If you want to see whether Xiaomi cars are good or not, you can look at them in three stages:
✅ In the short term (next few months)
to focus on: What is the real feedback from car owners after the software upgrade in October? Is the system more stable? Do users dare to use assisted driving?
✅ In the medium term (half a year to a year),
look at two key data. SU7 sales trend: will it continue to grow, or will it start to decline? R&D investment: More money is spent, has it been exchanged for technological progress?
✅ In the long term (more than two or three years),
it will be seen whether Xiaomi can truly achieve "ecological linkage" - such as the seamless connection of mobile phones, cars, and home appliances, bringing real convenience, not just a concept.
At present, Xiaomi's stock price is around HK$51, with a price-earnings ratio of about 28 times, which is about 15% higher than the industry average. Whether this "premium" can be supported depends on two points: whether Xiaomi can quickly fix technical problems; Do you believe that it has a reliable risk control system?
If it can be done, the premium is worth it; if something happens repeatedly, the market will reprice and the stock price may also pull back.
