Every gift from fate comes with a hidden price. This summary of Marie Antoinette's dramatic life, by Stefan Zweig, aptly applies to China's "Wei Xiao Li" trio—especially to Li Auto today.

The year 2025 witnessed an unexpected reversal in China's new energy vehicle market. XPeng and NIO, two brands long considered on the brink of survival, managed to leave intensive care. With year-on-year sales growth of 125.9% and 46.9% respectively, they drove revenues to 76.72 billion yuan and 87.49 billion yuan. Both companies achieved profitability in the fourth quarter, marking their first profits after years of losses. In contrast, Li Auto, once the sales champion among new automakers, became the only brand in the sector to report negative sales growth. The company achieved just over 60% of its annual sales target, with revenue falling 22.3% year-on-year to 112.3 billion yuan. Net profit plummeted 85.8% to 1.14 billion yuan. Market concerns are growing that it may fall back into losses in 2026.

On the surface, this appears to be a tactical failure. While XPeng successfully moved upmarket with models like the G6 and X9, and NIO narrowed its losses through the synergistic "Ledao + Firefly + NIO" three-brand strategy, Li Auto found itself mired in internal product conflicts and negative publicity. A deeper reason may be that Li Auto needs to catch up. The extended-range vehicle window allowed for rapid rise, but it also meant the company missed the tough battles fought by its pure-electric peers. This absence deprived it of opportunities to hone its competitive edge and led to an over-reliance on a single path. Given this performance divergence, a comprehensive reorganization has naturally become Li Auto's most urgent task. For several consecutive months, the company has implemented a series of adjustments in personnel, organizational structure, and corporate strategy, demonstrating a determination for radical reform. However, in the fiercely competitive new energy vehicle battlefield, determination alone is insufficient. To secure its place at the table, Li Auto must rediscover its original identity.

During an earnings call last November, Li Xiang admitted to making a mistake that persisted for three years. He stated that over the past three years, Li Auto attempted to manage itself using a professional manager model, which was ill-suited to the volatile market conditions and the company's specific realities. Consequently, starting in the fourth quarter of 2025, the company fully returned to a startup-style operating model. In recent years, Li Auto had indeed lost some of its original character. When the Li ONE became a surprise hit in 2021, despite market skepticism, the company stuck to its product formula of "extended-range + spacious interior + refrigerator, TV, and large sofa." It later leveraged the similar "cookie-cutter" designs of the L9, L8, and L7 to claim the top sales spot among new automakers and become the first to achieve profitability.

This success stemmed from Li Auto's "zero-to-one" innovation in products and systems, supported by an internal debate mechanism conducive to innovation—Li Xiang admitted he once wanted to cancel the L7 project but lost an argument with senior executives. However, this success was later misinterpreted as finding a replicable business model. The failure of the MEGA model was the first domino to fall. This highly anticipated flagship pure-electric model, launched in March 2024, was immediately embroiled in controversy over its design, pricing, and target use case. It sold just over 3,000 units in its first month, far short of the 8,000-unit monthly target. Li Xiang's subsequent comments revealed that the internal debate mechanism had already faltered. He stated that one reason for the MEGA's rocky start was that "younger employees" on the project were afraid to challenge him, leading to insufficient attention to user experience and pain points upon launch. Instead of using the MEGA's failure as an opportunity to reassess consumer expectations for pure-electric vehicles, Li Auto hastily integrated its pure-electric products into the established "cookie-cutter" model from its extended-range era. In April 2024, Li Auto restructured its product lines into three independent divisions: vehicles over 400,000 yuan (MEGA/L9), 300,000-400,000 yuan (L8/L7/i8), and under 300,000 yuan (L6/i6). While intended to better target different price segments, this adjustment triggered widespread confusion across product planning, supply chain, and sales systems.

To artificially create tier distinctions, the i8's Pro and Max versions omitted the signature refrigerator and second-row entertainment screens respectively, leaving consumers feeling shortchanged. The i6 initially garnered strong orders but missed its launch window due to battery supply issues. Its low-price strategy also cannibalized sales from the L series and the i8. This "product line experiment," which lasted a year and a half, only concluded in December 2025 when Li Auto merged its first and second product divisions. By then, significant damage had been done to the brand's reputation, and its gross margin had fallen from a historical high of 22.2% to 16.8% in Q4 2025, perilously close to the 15% break-even line. As Li Xiang acknowledged, he misjudged both the market and his own company. Transitioning prematurely to a professional manager model and rigidifying experience into internal processes not only failed to support success in the pure-electric segment but also constrained the company's most valuable capabilities: product definition and innovation. Fortunately, after paying a steep price, Li Auto has begun to correct its course. Unfortunately, the market—whether for pure-electric or extended-range vehicles—has become a brutally competitive red ocean.

Recent months have seen reports of senior executives departing from Li Auto's autonomous driving department. Li Xiang revealed that their replacements are largely from the post-90s and post-95s generations, claiming that "everyone has recently noticed that work efficiency has increased significantly, with the AI driving model iterating from once every two weeks to once per day." While everyone agrees Li Auto must accelerate, the direction may not be the AI focus that Li Xiang firmly believes in. On January 26 this year, Li Xiang abruptly convened a two-hour all-hands meeting focused almost exclusively on his views regarding AI trends and Li Auto's AI strategy, with little mention of the core automotive business. This left many employees confused and dissatisfied, as it failed to address the pressing issue of sluggish sales. They were hoping for reflections on sales performance and the company's near-term strategy. In 2025, the growth rate of the domestic extended-range vehicle market plummeted to 6.0% from 70.9% in 2024. As market growth stagnated, price wars erupted around nearly every Li Auto model. Its signature "refrigerator, TV, and large sofa" have become industry standards, with almost all SUVs now promoting "spacious interiors, six/seven seats, and no range anxiety." Although the proportion of pure-electric vehicle deliveries increased from 2.5% in 2025 to 66.6% in February this year, restructuring the company's revenue mix, Li Auto's position in the pure-electric market remains challenging. Consider charging infrastructure. Competitors like Tesla, XPeng, Xiaomi, AITO, and BYD all feature 800V systems; Li Auto's 500kW 5C supercharging is not unique. While Li Auto aggressively built over 4,000 supercharging stations from 2024 to 2025, leading in total numbers, its coverage spans just over 280 cities, lagging behind XPeng's 430+ cities and Tesla's 370+ cities.

This deficit is a legacy of Li Auto's previous focus on the extended-range niche. Some once believed this debt need not be repaid, confident that Li Auto's profitability and massive cash reserves would outlast competitors. However, with NIO and XPeng reporting profits in Q4 last year, this advantage has diminished. Surviving rivals have also become profitable. As competitors grow stronger and Li Auto remains mired in difficulties, employee anxiety is understandable. Two months later, they received an answer, albeit a somewhat disappointing one. During the March earnings call, management announced a 2026 sales growth target of 20%, aiming for approximately 480,000 to 500,000 vehicles. This is a relatively conservative target. Li Xiang commented, "Due to intense competition and market conditions, there are more new models priced above 200,000 yuan this year than in previous years, but overall market growth is very limited." One detail is noteworthy. For the first quarter of 2026, Li Auto's sales guidance is only 85,000 to 90,000 vehicles. To meet the full-year target of 490,000 units, the company must achieve a significant V-shaped recovery in the second half of the year. The next-generation Li L9, set for release in the second quarter, is seen as the linchpin for this rebound. Based on available information, this new product represents a major test for Li Auto under its revived startup model.

On the product front, Li Auto is demonstrating a fight-to-win attitude, focusing efforts on the 400,000+ yuan extended-range market, significantly simplifying SKUs, and aiming to recapture market share in the premium segment. Concurrently, the company is revitalizing its sales team by delegating operational authority and profit sharing. Starting in March, Li Auto fully implemented a store partner program. Li Xiang stated plainly that the goal is to "develop store managers capable of earning over 1 million yuan annually."

Li Auto is returning to a startup model, but whether it is on the right path to rediscovering its core identity is debatable. Several signs suggest it is leaning towards technology, rather than its true strength: product excellence. On the March earnings call, Li Xiang told investors, "Facing an increasingly competitive new energy vehicle market, we are committed to raising our technological moat and completing the transition from a smart electric vehicle company to an embodied AI enterprise." This is part of a broader technology narrative. In Li Xiang's view, 2026 is the final year for leading AI companies to "enter the vehicle," and L4 autonomous driving systems will be implemented by 2028 at the latest. By then, he believes no more than three companies globally will be simultaneously developing foundation models, chips, embodied AI (AI + hardware), and operating systems—and Li Auto aims to be one of them. The upcoming L9 is seen as the first battle in this transformation into an AI tech company; internally, it is defined not as a car, but as an "embodied AI robot."

Certainly, technological ambition and product focus are not mutually exclusive. However, the two-hour meeting that confused employees also revealed a degree of misalignment within Li Auto today. Previously, Li Xiang criticized competitors for "having a group of people with no user mindset, who完全不关心用户的这帮人,天天研究技术路线" (having a group of people with no user mindset, who完全不关心用户的这帮人,天天研究技术路线). Now, Li Xiang seems to be doing the opposite. He extensively discusses AI, while the concerns of employees closer to customers go unanswered. A comment from Li Auto shareholder and Meituan founder Wang Xing somewhat reveals the disconnect between this shift and product focus. He said, "The MEGA seems like a car built for the CEO of Li Auto himself, while the L6 is a car that Li Auto employees built for themselves." Extending this analogy, for whom is the L9 being built? Li Auto's consumers are generally not hardcore technologists; the concept of an "embodied AI robot" is unlikely to resonate strongly with them. This suggests that the most valuable capabilities Li Auto lost in recent years—the deep understanding of user needs and the internal debate mechanism that fostered innovation—have not fully returned. Beyond faster execution and more aggressive profit-sharing—hallmarks of a startup mentality—what Li Auto truly needs is to rediscover the product-centric spirit of its early days. Even if embodied AI is indeed the future, the path forward is not to mimic the technology narratives of Tesla, XPeng, and others, but to find the "refrigerator, TV, and large sofa" of autonomous driving—delivering an embodied AI experience that genuinely delights users. One promising attempt is Li Auto's decision to display the reasoning process of its V-L-A (Vision-Language-Action) autonomous driving system on the vehicle's screen.

This intuitively shows drivers why the system decides to change lanes or whether it has detected a red light. This transparent design enhances user trust and echoes the user-centric product philosophy of Li Auto's early years. Whether Li Auto can apply this same standard to develop a compelling product logic for embodied AI will likely determine the success of the L9 and subsequent models. With the extended-range segment stagnating and the pure-electric arena fiercely contested, the crowded new energy vehicle market has less and less room for mediocrity. The ability to firmly replant the flag of product excellence may ultimately determine Li Auto's final outcome.