Profit Margin Drops to 2.9%, Automakers in AI Era Eager to Rebuild Brand Premium Logic

Author |Ben Yi

Editor |De Xin

In April 2026, there were 660 car models available in the Chinese market, but only 29 had monthly sales exceeding 10,000.

In late May, Shenzhen, 13 leaders from the automotive industry gathered at the 4th Future Car Pioneer Conference. Morning talks focused on product reconstruction, afternoon discussions on vast horizons, but almost everyone was saying the same thing: The competition of specifications is unsustainable; the next card to play must be the brand.

Industry profit margins fell to 2.9%, domestic passenger car retail dropped nearly 19% year-on-year, and more than 150 new models crammed into the market this year but most couldn't even enter user choice circles. The endpoint of the "specification racing" road should be visible to everyone.

I. Industry-wide "Brand Awakening"

Li Bin quoted a set of data from McKinsey research: the ranking of brands in consumer car purchasing decisions ranked fifth two years ago, but has now risen to second. "Believe it will soon enter the first place."

Source: Future Car Pioneer Conference

This judgment was repeatedly verified in the mouths of different speakers.

Avatr Chairman Wang Hui said consumers no longer need you to tell them how fast the acceleration is or how far the range is. "They want to know more rationally what this brand believes in, pursues, and whether it is worthy of entrusting."

Leapmotor Senior Vice President Xu Jun was even more direct. "Specifications have become the most basic entry ticket, no longer the deciding factor."

Chery Vice President Wang Lang believed that cost-performance ratio solves "worth buying", brand value solves "worth believing, worth long-term choosing, worth recommending to others".

Every press conference talks about 800V platforms, large AI models, end-to-end intelligent driving. "If it continues like this, there will be no product differentiation."

If important decision-makers in the whole industry begin to collectively use "brand" instead of "specifications" for external expression, this itself conveys a clear signal: the competition of China's intelligent electric vehicles is shifting from "hardware racing" to competition at the brand dimension.

II. Profit Margin Drops to 2.9%, A Consensus Forced Out

Avatr's Wang Hui shared a set of data from CPCA: domestic car sales exceeded 34.4 million vehicles in 2025, but profit margins were only 4.1%.

By January-February 2026, this number fell further to 2.9%.

He spoke sharply on site, "Sales without profit are fake sales, scale won by price wars is even more false prosperity."

CPCA Secretary-General Cui Dongshu added a supplement to this data from the supply chain perspective.

Whole vehicle enterprises selling cars don't make money, "basically becoming a state of working for upstream", while mining industry net profit margins have risen to about 40%. Car companies expanded the industry cake, but profits were taken by upstream.

Source: Future Car Pioneer Conference

The market side is also seriously challenging. Li Bin revealed that domestic passenger car retail sales dropped 18.9% year-on-year from January to April 2026, and even dropped more than 24% in the first few weeks of May. CPCA's annual forecast adjusted from -1% at the beginning of the year to -11% all the way. Cui Dongshu himself said, "This "drastic forecast adjustment is historically rare".

The backlash caused by price wars has also fully manifested.

Wang Hui calculated an account, "Price can only hold for two months, after two months there is not much effect, and there is a pile of old users being backstabbed." Since April, nearly 20 products in the market have announced price hikes. Avatr's own 06T and new Avatr 12 average price was 20,000 to 30,000 yuan higher than the previous generation. He said, "After the price increase, we can finally sleep soundly. Because we told customers we wouldn't stab them in the back, this counts as fulfilling the promise to users."

In a sense, it is not that car manufacturers chose to wake up proactively to carry out "brand upgrading", this is more of a passive turn under profit squeeze. When prices can't be driven down, specifications can't stack differences, "brand" almost became the only competition dimension that hasn't been fully penetrated.

III. Path Divergence, Creating AI Era Automotive Brands

Consensus is consensus, but specifically "how to build a brand", each gives a completely different answer.

Li Bin stated he would not make extended range, plug-in hybrid, and MPV. "Cars that are obviously not profitable will not be developed." Basic R&D (chips, operating systems, batteries) is firmly invested, application layer looks more at ROI. NIO brand average transaction price is about 400,000 yuan, 50,000 yuan higher than BMW, 130,000 yuan higher than Audi, achieving operating profit for two consecutive quarters. ES8 launched 215 days, delivered over 100,000 units, consecutive 5 months over 400,000 yuan regardless of energy form sales champion, so the current NIO is "save where you should, spend where you should".

Changan Group just streamlined its future five-year products from 63 models to 36 models. Wang Hui's led Avatr has struck a combination of punches in the overseas market: entering more than 40 countries, Avatr 11 starts at 290,000 yuan in China, selling near 450,000 yuan overseas, Thailand luxury electric SUV sales first, achieving stable profit after only one and a half years entering overseas.

"50:50 market, but looking globally, may need to put 80% of resources to gain the recognition of another 50% of customers." Wang Hui's underlying meaning is obvious, Domestic competition is exhausted, globalization is real incremental.

Xu Jun did not reveal details of the second brand, but said a thought-provoking sentence, "The end game of China's automotive industry development must be brand". When Leapmotor starts talking about brand upgrading, this itself is a clear signal. System synergy capability (CTC+ Central Integrated Architecture overall optimization), scenario definition capability (C16's large six seats solves second-child family travel, not just two extra seats), and long-term compound interest accumulated from over a decade of full-domain self-research, are the three layers of "hidden kung fu" supporting Leapmotor's turn.

Zhang Zhengping revealed, AITO M9's intelligent driving mileage share has exceeded 52% (overall data of 280,000 vehicles). Users choose AITO Top 3 reasons are: Intelligent driving, Safety, Brand, traditional "high performance" (acceleration, range) even did not enter the top three.

AITO also did one thing in the manufacturing end: "One Car One File", every screw of each car, each welding time, voltage, current, and operator have complete records. This is adding user trust in the brand from the underlying infrastructure.

New Generation AITO M9｜Source: HarmonyOS Intelligent Official Website

If a car company wants to build "Intelligent Brand Competitiveness" in the AI era, may need to learn from Geely.

Geely Research Institute Director Li Chuanhai admitted facing severe AI talent shortage, "Those who understand AI don't understand products, those who understand mechanics don't understand training models", and traditional recruitment can't attract top AI talents, those people are in large model companies.

Geely's solution is not hard recruitment, but building an ecosystem, and cooperating with partners like StepFun, Qianli Technology, XinQing Technology, etc., pulling AI talents into their own ecosystem. Product end, ZEEKR 4 years cumulative delivery nearly 750,000 units, average price over 300,000 yuan, ZEEKR 9X average transaction price over 530,000 yuan, consecutive months over 500,000 yuan SUV sales champion, proved the output ability of this system.

The intelligent driving supply chain side also has resonance on this.

DeepRoute CEO Zhou Guang judged, current small model intelligent driving has reached the ceiling, "Investment is increasing more and more, improvement is slower and slower".

He used "See-Saw Effect" to describe: small models solved Shanghai's problem, Shenzhen might get worse, the repeated see-saw between versions is the core reason users find it difficult to trust intelligent driving systems long-term.

DeepRoute bet on large model paradigm transition, deployed 40B base model, goal is to increase city safety takeover mileage from tens of kilometers to over 1,000 kilometers. If realized, intelligent driving will change from "usable" to "commonly used", in reverse will also become the core barrier of brand differentiation.

Source: 2026 Future Car Pioneer Conference

Additionally, worth mentioning is that the topic of FSD entering China was mentioned multiple times that day.

Zhang Zhengping stated "New M9 equipped with 6 laser radars, 40 sensors, hardware configuration definitely not lose to FSD". Zhou Guang was much more open-minded, he believes "This is a positive signal, Tesla FSD entering will make intelligent driving safety higher", but also pointed out, if still insisting on small model paradigm, "it will be quite troublesome".

Competition has upgraded, every company must bring out real stuff.

IV. Disintegration and Reconstruction of "Luxury"

"Future ultra-luxury cars must collapse. There is no luxury car concept in the world, only premium car concept. Luxury will be replaced, just like electronic consumption, no century accumulation, only technological innovation." Cui Dongshu gave a very radical judgment.

But the big shots doing premium products on stage may not fully agree with this.

Li Bin holds 400,000 yuan average price not letting go, Wang Hui's Avatr overseas wants to sell 450,000 yuan, Xiang Xingchu's Luxeed S800 consecutive 8 months holding luxury sedan sales champion over 700,000 yuan, Zhang Zhengping's AITO M9 consecutive 2 years 500,000 yuan level champion.

Interesting is, This "disagreement" on luxury concept may not be true disagreement.

Cui Dongshu speaks of old luxury disintegration, that is luxury supported by century accumulation, handcrafting, Logo premium. What Chinese brands are doing is exactly using "Technology + Experience" to redefine luxury, which is "Real luxury is not who you are like, but who you are."

As the only foreign car company guest that day, Mercedes-Benz China R&D Head Drummond Jacoy being present on site itself makes the discussion interesting.

He mentioned "Oil and Electric Equipped with Intelligence" and "Chinese Speed" in speech, for example Chinese R&D team has begun to lead multiple global projects, including new generation S-Class rear entertainment system. When asked if Chinese luxury new brands pose a threat to Mercedes, his answer was relatively restrained, but significant: "China has many very excellent brands and products, I highly respect them."

Source: 2026 Future Car Pioneer Conference

Cui Dongshu and manufacturer representatives actually speaking of "Luxury" one body two sides: Old luxury is dissolving, new luxury is building. Disagreement also fundamentally not at "whether to do premium", but "what supports premium".

When Luxeed S800 sells to 800,000 yuan with technology, when NIO average price exceeds BMW, market data has already given the answer. The foundation of brand premium, is shifting from historical accumulation to technology experience.

This is probably the most worth paying attention to topic of this year's Future Car Pioneer Conference: when industry profit margin drops to 2.9%, when specification competition reaches the end, each car brand must answer is, what exactly are we selling?

Obviously, a function configuration table piled with various specifications is already not attractive enough.