NIO, XPeng, Li Auto, Leapmotor Begin Selling Technology: How Big Can This Business Be?

Overview

Chinese electric vehicle (EV) startups, including Li Auto, NIO, XPeng, and Leapmotor, are increasingly turning to technology licensing and external supply of core components as a new revenue stream. While Li Auto has yet to officially comment, its CTO Xie Yan confirmed in June that the company does not rule out supplying self-developed chips externally.

At the recent Leapmotor Technology Day, the company positioned itself as a "leading supplier of technology to multiple automakers." NIO founder William Li stated in early June that its self-developed "Shenji" chip could generate "several billion yuan in external revenue this year." XPeng's technical collaboration with Volkswagen has already begun to show meaningful financial impact.

What Are the "Four Major New EV Players" Selling?

According to reports, Li Auto is preparing to externally supply three core technologies: the Mahe chip, silicon carbide components, and intelligent driving chips. Although unconfirmed officially, chip development has been a key focus for all three major players—Li Auto, NIO, and XPeng—with results materializing in May.

On May 15, the new-generation Li L9 debuted with the self-developed Mahe M100 chip, while NIO's Onvo L80 launched with its self-developed Shenji chip. On May 20, XPeng's GX hit the market.

However, each company's external supply strategy differs:

NIO

In June 2025, NIO spun off its chip business into Anhui Shenji Technology Co., Ltd. The company now offers a full product lineup including NX9031X, NX9031U, and NX9031C chips, covering intelligent driving, embodied intelligence, and AI agent reasoning. Shenji has completed nearly 3 billion yuan in financing.

William Li confirmed on September 1 that Shenji would generate "several billion yuan in external revenue this year," though detailed figures were not disclosed. Beyond chips, NIO also supplies battery-swapping technology. In mid-August, NIO Energy completed the asset transfer of 36 charging and swapping stations with Wuhan's Optics Valley Transportation Group, with all existing stations in Wuhan now state-owned. A similar deal with Guangzhou Science City Group followed.

XPeng

XPeng has already benefited significantly from technology sales. Its collaboration with Volkswagen covers vehicle platforms, software, electronic architecture, intelligent driving systems, and the dual Turing AI chip. According to XPeng's financial reports, "service and other income" reached 8.34 billion yuan in FY2025 (68.2% margin), and 2.70 billion yuan in Q2 2026 (75.1% margin). Service revenue gross profit in Q2 2026 was approximately 2.03 billion yuan, nearly matching its car sales gross profit of 2.06 billion yuan.

Leapmotor

Leapmotor took a different path. Founder Zhu Jiangming, an engineer by training, was the first among the new players to develop a self-driving chip—launching the "Lingxin 01" in October 2020. However, he was also the first to abandon it. In June, Zhu stated that AI chips for intelligent driving are now "oversupplied."

Leapmotor instead focuses on its core strengths in three-electric technology (battery, motor, electronic control). It positions itself as "not just a carmaker, but a leading technology supplier to multiple automakers." Its technology offerings include LEAP vehicle architectures, electronic/electrical architectures, and batteries. At its September Technology Day, Leapmotor unveiled the LEAP 5.0 architecture, LWM world model intelligent driving, MM-i multi-mode hybrid electric drive, and CTC 3.0 battery technology. The MM-i system will first be supplied overseas, while CTC 3.0 patents will be open-sourced.

Leapmotor has secured two major partners: Stellantis (via the joint venture "Leapmotor International," upgrading to Tier 1 supplier status) and FAW (which invested approximately 3.744 billion yuan for a 5% stake, adopting the LEAP 3.5 architecture).

The Urgent Need to Share Costs

New EV players face a dual financial challenge: low margins on vehicle sales and relentless R&D spending.

According to Chen Shihua, Deputy Secretary-General of the China Association of Automobile Manufacturers, the average profit margin for domestic vehicle manufacturing in the first half of 2026 fell to 1.5%, a decade low. For a 200,000-yuan new car, net profit is only about 3,000 yuan.

XPeng's financials illustrate the impact of technology sales. In the first half of 2026, XPeng reported a net loss of 3.12 billion yuan (vs. 1.14 billion yuan in H1 2025). Overall gross margin was 20.6%, while automotive gross margin was only 12.1%. The improvement in overall margin was driven by technology R&D service revenue recognized in Q2 (H1 2026: 4.73 billion yuan, up 67.1% YoY, with a 71.4% margin).

However, XPeng's technology revenue remains heavily dependent on a single customer (Volkswagen) and a single project. Starting in 2027, this revenue will shift from fixed technical service fees to variable royalties tied to sales volumes.

Cost Advantages of Self-Development

William Li has repeatedly stated that self-developed chips can achieve "one chip replacing four," saving approximately 10,000 yuan per vehicle. Based on NIO's 2025 sales, mass production of the Shenji chip could save over 1.8 billion yuan annually.

Xie Yan of Li Auto noted in May: "The industry needs greater computing power—thousands or tens of thousands of TOPS—and the cost is high. If a supplier's chip costs half as much with three times the performance, we wouldn't need to develop our own." He added that self-developed chips become cost-effective only at production volumes of hundreds of thousands of units or more.

To maximize the benefits of self-development, automakers also need external supply. A CITIC Securities report argues that self-developed chips represent not just a substitution strategy but a shift toward integrated hardware-software development. The report estimates that a lifetime volume of one million chips is sufficient for economic viability.

XPeng's Turing chip has set a target of shipping 1 million units this year. NIO's Shenji chip has already shipped over 300,000 units.

Why Can These Technologies Be Sold?

For automakers to successfully supply technology externally, they must offer clear advantages in cost or performance.

Xie Yan stated: "Chips are evaluated on only two dimensions: cost and performance. While many chipmakers exist, only limited production capacity can actually be sold. Competition is fierce." He believes only leading automakers can sustain self-developed chips and maintain competitive advantages.

NIO's Shenji unit also plans cross-sector applications, transferring vehicle perception, real-time control, and functional safety technologies to humanoid robots, logistics vehicles, and high-performance computing terminals.

Zhu Jiangming of Leapmotor noted that platform-level cooperation with FAW and Stellantis gives Leapmotor a cost advantage of at least 1,000–2,000 yuan per vehicle compared to partners sourcing from multiple suppliers, along with faster integration and lower risk.

Cui Dongshu, head of the China Passenger Car Association, told Yicai that rising R&D costs for intelligent EVs make it difficult for single brands to fully amortize investments. He sees technology external supply as both a short-term catalyst and a long-term industry trend.

Ji Xuehong, director of the Automotive Industry Innovation Research Center at North China University of Technology, noted that technology transfer from automakers is not new. Previously, joint ventures imported foreign technology via equity investments; now, Chinese brands like XPeng and Leapmotor are exporting technology to international companies. He believes this aligns with specialization trends, as companies make different strategic choices based on their technological endowments.

Challenges to Scaling

Despite the potential, technology cooperation faces significant commercial hurdles.

Ji Xuehong argues that technology cooperation is unlikely to become a primary revenue source for automakers. It remains supplementary income with high uncertainty: partners may switch suppliers, and cooperation duration is unpredictable. Long-term competitiveness ultimately depends on core business execution.

Cui Dongshu predicts that industry profit pressure and high R&D spending in 2026 will accelerate technology openness. Even if price competition eases and vehicle margins recover to 4% or above, automakers will likely retain external supply but narrow its scope. When industry conditions are favorable, companies will prioritize their own models and only supply mature technologies externally.

Ji also cautioned against overemphasizing competitive tensions. While market participants compete, the benefits of cooperation are highly certain, while conflict risks are relatively low. The roles of "shareholder, competitor, and technology supplier" should be viewed with the cooperative and transactional aspects taking precedence.

Cui Dongshu added that for buyers, the core concern in purchasing a competitor's underlying architecture is not technical feasibility but control, differentiation, and security risk. If a competitor controls the electronic/electrical architecture—the vehicle's hardware-software foundation—the buyer risks version lag, reduced differentiation, and data leakage. Escalating competition could also lead to abrupt termination of cooperation, creating supply chain risks for future model updates. Buyers therefore typically seek secondary development rights to maintain upper-layer self-development capabilities.

(Source: Yicai)