In the AI era, global automakers are starting to look for their own "Model Y".
In the past two years, Elon Musk's dwindling enthusiasm for car - making has been widely criticized.
Specifically, as Tesla transforms into a technology company, the Model Y shoulders half of the car - selling business. In 2023, the global sales volume of the Model Y reached 1.22 million units; in 2024, its sales volume still stood at 1.185 million units. In the first half of 2025, the sales volume of the Model Y in the Chinese market reached as high as 171,491 units, ranking first in the SUV sales list.
This remarkable achievement seems to make Musk, who is more interested in the vast universe, not intend to put more effort into car - making. Instead, he has created a series of "new cars" based on the Model Y. For example, Tesla officially announced the 6 - seat Model Y L, and the upcoming affordable car is a new version of the Model Y.
In fact, after a wave of car - making frenzy, the global technology circle has gradually calmed down. At the beginning of last year, Apple announced the cancellation of all development plans for its autonomous electric vehicle project. After burning through $500 million, Niutron has not achieved mass production. The car - making projects of real - estate companies such as Baoneng, Evergrande, and Agile basically have no further progress.
On the other hand, domestic car companies are still accelerating the launch of new models. Foreign media have compiled a list of the number of new cars launched globally in 2024. The results show that nearly half of the new cars come from China, with the number reaching 79. In April 2025, the number of new car iterations reached the highest peak in the past five years, totaling 167 models.
However, dozens of cars may not have the commercial viability of a single classic car. Compared with blindly making and launching new cars, global car companies urgently need to find their own "Model Y".
To this day, why do car companies strive to keep their "classic models"?
To this day, the "flood - of - models" strategy of car companies has not been abandoned.
Incomplete statistics show that Changan Automobile alone plans to launch 50 new energy products in the next five years, including more than seven global best - sellers. Avita expects to launch 17 cars by 2030, covering a product matrix including sedans, SUVs, and MPVs.
With the "Golden September and Silver October" approaching, domestic car companies are gearing up. For example, Hongmeng Zhixing will launch at least five new cars, including the Xiangjie S9T, Shangjie H5, the new AITO M7, the Zhijie R7, and the facelifted Zhijie S7. Leapmotor still has two aces up its sleeve: the flagship D - series and the Leapmotor B05. The NIO ES8 and the Ledao L80 are about to be launched. Li Auto also has another pure - electric model, the Li i6, yet to be released.
But is the "flood - of - models" strategy really meaningful?
According to data research by Autohome, among the new energy vehicles launched between 2022 and 2023, more than 80% of the models have monthly sales that never exceed 5,000 units, and 40% of the models even fail to reach 1,000 units. Some models are directly discontinued, completing their life cycles in less than two years.
In 2023, 126 new models (excluding facelifts) were launched in the Chinese market. Among them, 44% of the new cars had monthly sales that always remained below 1,000 units after their launch. Only 23 models reached monthly sales of 5,000 units, and 7 models reached 10,000 units. The cruelty of the new energy vehicle industry probably lies in the fact that most models cannot recover their R & D costs during their life cycles, and only classic models can continuously create value.
Looking at global new energy car companies, only Tesla and BYD can be said to have classic models. The representative model of Tesla is the Model Y, and BYD has the Qin PLUS. Since these two models were mass - produced in 2021, their current monthly sales are basically stable at around 30,000 to 40,000 units.
What can a classic car bring to a car company?
First, in terms of sales, the scale effect created by classic models cannot be underestimated. In the first half of 2025, Tesla's total sales in the Chinese market reached 263,400 units, a year - on - year decrease of about 5.4% compared with 278,300 units in the same period of the previous year. In the US market, the total sales were 271,600 units, a year - on - year decrease of 10.8%. The sales decline in the European market was the most severe, with a year - on - year decline of 33%.
However, against the trend, the Model Y holds up half of the sky.
Taking the Chinese market as an example, in June 2025, the sales of the Model Y increased by 9.09% year - on - year and 81.06% month - on - month. The sales data released by the China Passenger Car Association shows that in May 2025, the sales of the Model Y were 24,770 units, accounting for as high as 64.19% of Tesla's total sales in China.
Second, in terms of profit. Undoubtedly, classic models have always been the "cash cows" of car companies, which was proven in the fuel - powered vehicle era. It is reported that the first - generation Volkswagen Polo had global sales of over 500,000 units in four years, contributing up to 15% of the profit to the Volkswagen Group.
Similarly, the Model Y was once regarded as the most mediocre product in Tesla's history, but in terms of commercial value, no one can deny that it is Tesla's most successful product. Data shows that two years after the mass production of the Model Y, it directly drove up Tesla's gross profit margin in the automotive business from 25.5% to 32.9%.
This is a huge temptation in the automotive industry with mediocre profit margins, especially when the profit margins in the automotive sector have been declining in the past two years.
Data from the China Automobile Dealers Association shows that in 2024, the total revenue of the Chinese automotive industry reached 10.65 trillion yuan, a year - on - year increase of 4%. However, the profit decreased by 8% year - on - year to 462.3 billion yuan, and the industry profit margin was only 4.3%, lower than the industrial average (6%). In November 2024, the profit of the automotive industry decreased by 35% year - on - year, and the industry profit margin was only 3.3%.
In addition, classic models make a subtle and long - lasting contribution to the brand of car companies.
The National Bureau of Economic Research in the United States pointed out in the research report "Intergenerational Transmission of Automobile Brand Preferences" that there is a significant correlation between parents and children in automobile brand selection. By analyzing data from 1968 - 2013, it was found that the brand of the first car in a family affects the car - buying decisions of children. The intergenerational brand transmission effect makes the brand repurchase rate of families of classic - model users higher than that of ordinary families.
For a long time, the "flood - of - models" strategy has become a lifesaver for many car companies. They believe in the logic of "more models, more success" and try to cover all market segments with dozens of models. However, as the car - making competition intensifies, this strategy has become a burden, and the Model Y is the best example.
In the AI era, has the thinking of car - making changed?
There was a time when the car - making industry believed in the concept of "hardware supremacy". Between 2015 and 2020, the focus of competition in the new energy vehicle market was mainly on quantifiable indicators such as cruising range, motor power, and screen size. Some car companies even directly transplanted the parameters of household appliances such as refrigerators and TVs into cars.
This kind of thinking once made car companies lose their innovation ability.
However, the era of hardware - defined cars has gradually passed, and "software" has become a hot word in the automotive industry. As the competition further intensifies, AI has replaced software as the main term for defining cars in the industry. It is reported that at this year's Shanghai Auto Show, Li Bin of NIO proposed the concept of the new three major components of intelligent vehicles: intelligent driving chips, full - domain operating systems, and intelligent chassis.
While the large - scale AI models are rapidly promoting the transformation of the automotive industry, they also turn cars from mobile household appliances into "super intelligent agents". However, it should be noted that has the thinking of car - making really changed from the era of piling up household - appliance features to the AI era? Not necessarily. During the two key transformation periods of the automotive industry, the evolution trajectory of the industry shows amazing similarities.
First, there is the same problem of homogenization.
A research report by Avnet Insights shows that AI is rapidly revolutionizing product design, and 42% of global engineers have applied AI to product design. In the automotive field, from vehicle - end perception, control logic, to human - vehicle interaction and the overall vehicle ecosystem, due to the high similarity of solutions, enterprises have fallen into the pursuit of "explicit parameters".
It is reported that in the current sales channels of intelligent electric vehicles, "computing power" has become one of the most common sales pitches.
Second, there is the same price war.
In the era when sofas, color TVs, and large refrigerators dominated the automotive transformation, price war was a common tactic used by car companies, and it remains the same in the AI era. Taking intelligent driving as an example, since 2025, more and more car companies have been pursuing equal access to high - level intelligent driving. Chery, Leapmotor, BYD, and XPeng have successively announced that high - level intelligent driving assistance systems will be standard on all their models, totaling more than a hundred models.
Moreover, the number of companies offering free services is increasing. It is reported that in April this year, Leapmotor announced that the intelligent driving software for models with intelligent driving assistance functions would be completely free, and even refunded the fees for the driving assistance software to users who had already paid. Both the XPeng G6 and the XPeng X9 emphasize that they come standard with Turing AI intelligent driving, with no option to install additional features, no subscription, and no charge.
This has a great impact on companies that adopt the one - time purchase or subscription - based payment models.
The representative car companies in this regard are Tesla and some car companies that hope to profit from software and services in the future, such as Volkswagen, which once planned to mainly make money from software by 2023. In 2024, Tesla once lowered the subscription price of its Full Self - Driving (FSD) software overseas. Once consumers get used to free services, the situation for these companies will become more embarrassing.
As of now, there are already voices opposing free high - level intelligent driving. The president of Bosch's intelligent driving and control division in China once said at the "2025 Artificial Intelligence Conference" that free services will bring disaster to the entire industry. But if not free, how can these companies survive? A report from the China Passenger Car Association shows that the penetration rate of L2+ level intelligent driving systems has reached 55.7%, and for every 1 - percentage - point increase, it means that 3 - 5 companies will be eliminated.
However, the premise of free intelligent driving is the reduction of costs in domain controllers, computing power, and sensors. In recent years, with the continuous development of artificial intelligence, the cost of high - speed Navigation on Autopilot (NOA) has begun to decline significantly. According to a research report by Minmetals Securities, the hardware cost of BYD's Tian Shen Zhi Yan C pure - vision solution has dropped below 4,000 yuan.
However, after the reduction of hardware costs, car companies still have to face higher implicit costs.
For example, algorithms need to continuously adapt to new cities, new scenarios, and new regulations, and the cost of model training is high. In terms of data annotation, the total investment of leading companies often reaches tens of millions or even hundreds of millions of yuan. According to data from Gasgoo, last year, R & D investment in intelligent driving accounted for 38% of the total expenditure of car companies, second only to battery costs and becoming the second - largest expenditure item.
Taking three leading car companies as examples, BYD plans to invest up to 100 billion yuan in intelligent R & D; by the end of 2024, NIO's R & D investment in the field of intelligent driving reached 46 billion yuan; Xiaomi's Lei Jun revealed that the company's annual R & D investment in intelligent driving exceeds 2 billion yuan.
Car - making is still a money - burning business. Even in the AI era, the thinking of the entire industry remains unchanged.
In 2025, is car - making really just for making cars?
Although the enthusiasm of industry giants for car - making has declined significantly, in 2025, new companies are still pouring into the car - making industry. Typical examples include Huolala, Jinyu, and Chueneng New Energy... Among them, Jinyu has launched a pure - electric convertible sports car that looks similar to the Porsche 911 and claims a cruising range of 500 kilometers.
After a period of heavy investment, hardware - based car - making is indeed easier than before.
BYD's financial report data shows that during the price war in the automotive market in 2024, BYD's profit per vehicle still increased by 8% year - on - year, mainly due to the optimization of the cost of the three - electric system. Among them, the mass - production cost of the blade battery has decreased by 41% compared with three years ago, and the yield rate of silicon carbide power modules has reached 98%.
Not only BYD, but the cost of the entire battery industry chain is also decreasing. According to the latest research data from the International Energy Agency (IEA), in 2024, the price of lithium - ion battery packs decreased by 20% year - on - year, the largest decline since 2017. But is car - making really that simple now?
Not to mention the still - money - burning AI - based intelligent transformation. As the industry chain has continuously lowered the difficulty of car - making, the automotive industry has long lost its respect for the manufacturing process. This has been an established fact for several years and is even more obvious now. Research data shows that in 2025, the overall number of quality problems in the industry is 226 PP100 (new - car quality is measured by the average number of problems per 100 vehicles), an increase of 16 PP100 compared with 2024.
Second, future car - makers will also face a new problem.
According to data from the National Bureau of Statistics, in 2024, the capacity utilization rate of the automotive manufacturing industry was 72.2%, a decrease of 2.4% compared with the previous year, and slightly lower than the 75.0% utilization rate of industries above a designated size. Looking back, in 2017, the capacity utilization rate of China's automotive manufacturing industry reached a high of 82.2%.
In this context, the government will only tighten the thresholds for car - making, including R & D, patents, production standards, and even entry qualifications. It is reported that it will be difficult to replicate the situation where Xiaomi obtained the qualification of Baowo through a special approval from Beijing. Currently, Huolala can only produce cars through Changan due to qualification issues.
In addition, it is difficult for new players to enter the car - making industry.