After CEO Li Xiang said last year that Li Auto needed to return to startup mode, the automaker has continued to streamline its organization and processes.

36Kr has learned from several people familiar with the matter that Li Auto is preparing a new round of organizational restructuring around its product decision-making process. The plan is to split off some key functions from the product department and fold them into R&D.

Based on its group structure, Li Auto’s product department is led by Fan Haoyu and includes electric vehicle platform, spatial intelligence, autonomous driving products for end users, interaction design, platform operations, the app, and the official website.

People familiar with the adjustment told 36Kr that the current changes mainly target the core product definition teams for the EV platform and autonomous driving. The EV platform definition team under the product department will be split off and merged into the vehicle R&D department led by Liu Liguo. The autonomous driving end-user product team will be merged into the foundation model R&D team led by Kun Zhan.

“Some product staff related to user experience may also move into Liu Jie’s product line,” one person familiar with the matter said.

A person close to Li Auto’s senior management told 36Kr that the decision-making process for vehicle product definition has long been criticized inside the company for being cumbersome. The main reason is that three departments are involved in product decisions: Liu Jie’s product line department at the front end, Fan’s platform product definition department in the middle, and the R&D team.

The purpose of this adjustment is to return management of product definitions to the R&D side, with the goal of shortening the product decision-making process and improving the efficiency and quality of product definition. After the adjustment is completed, vehicle product decisions will be reduced from three departments to two, namely the product line and R&D, removing the intermediate layer of platform product definition.

Driving assistance has weaker product attributes than vehicle platforms, and whether it needs to sit alongside R&D as a parallel department has long been debated. At many companies, assisted driving R&D and product functions sit within the same business department. Li Auto’s move to put product functions under R&D is also broadly in line with that direction.

Removing a middle layer from product decisions

36Kr has learned that news of the adjustment has already spread inside Li Auto, and employees have been notified verbally.

Under the previous organizational structure, Li Auto’s product direction was determined by a broad network covering the software and hardware experience of its vehicles.

“After the adjustment, the product department will continue to be responsible for businesses such as spatial intelligence and smart glasses,” a person close to Li Auto said.

The core change is to shorten the product definition process. Under the previous structure, a function or vehicle model had to pass through three layers, with the final layer responsible for engineering implementation. After the adjustment, the product department will be removed from that process, allowing the product line and R&D department to interface directly.

Li Auto’s decision to set up both a product line and a product department came from its broader effort, beginning in 2022, to learn from Huawei.

In 2024, Li Auto adjusted its product organization. It separated the vehicle model product development team from the product department led by Fan and merged it with the product commercialization team to form the product line, led by Liu Jie.

Since then, the product line has focused on decision-making and execution around commercial product success. It is responsible for collecting and consolidating information from users, experts, partners, and the industry, then using that information to coordinate product definition and cross-department work.

The thinking at the time was that the product line’s mission and goal were to “achieve commercial success, satisfy users, and make money for the company.” During the product development stage, it focused on working with the R&D and supply teams. During the launch period, it focused on working with sales and marketing.

After that adjustment, Fan’s product department took a step back and continued to focus on platform products such as smart driving and electric platforms.

At the time, Li Auto was trying to transform itself from a product company dependent on hit models into a process-driven company with a stronger organizational system and long-term competitiveness. In product R&D, this meant using rigorous processes to ensure that every product had precise market positioning and to avoid unfocused development. But in practice, the organization became bloated and decision-making slowed.

During the earnings call for the third quarter of last year, Li said the company needed to return to a startup-style management model. One part of that was improving efficiency, rather than occupying more resources. Li added that only when value increases, efficiency improves, and key problems are solved can the company continue to meet users’ needs in a highly competitive and rapidly changing environment.

Li Auto searches for its rhythm

As automotive technology changes slow and products become increasingly similar, a small group of top product managers may no longer be enough to help Li Auto maintain its edge. Li has therefore been trying to give the company a broader capability moat beyond product definition.

Since the start of this year, Li Auto has made several organizational adjustments. In January, to prepare for a new round of competition driven by artificial intelligence, Li Auto integrated its R&D teams into groups including the foundation model team, software embodiment team, and hardware embodiment team. In late May, Li Auto added three second-level departments related to embodied intelligence under the foundation model department: embodied engineering, embodied interaction, and embodied behavior.

Since last year, Li Auto has faced pressure from sluggish sales growth. The latest disclosed delivery data show that Li Auto delivered 30,895 new vehicles in June. By comparison, Nio delivered 40,597 vehicles, while Xpeng delivered 40,126.

The pressure on Li Auto’s sales is widely seen as stemming first from the conservative refresh of its L series in 2025, which allowed rivals to narrow the product gap that had been central to the company’s strength. At the same time, the Li Mega and Li i8, which were meant to help the brand break through in the battery EV market, stumbled one after another. As a result, Li Auto’s momentum in the battery EV segment now depends mainly on a single model, the Li i6.

For a company whose identity has long centered on product, losing ground in product competitiveness is a serious risk. That is especially true this year and in the years ahead, as industry competition intensifies. Upscale players such as Nio, Zeekr, and brands under the Harmony Mobility Intelligent Alliance have moved quickly, while Xiaomi is also expanding in the market. The pressure on Li Auto to defend its product strength has grown accordingly.

“The first priority is to improve decision-making efficiency and eliminate those endless, ineffective meetings,” a person familiar with the matter said, adding that the goal is to let product and R&D speak directly to each other.

Every adjustment comes with pain. Power structures are dispersed and reorganized, affected departments may see some employees leave, and the original rhythm is disrupted before a new one takes shape.

A Li Auto engineer told 36Kr that after Chen Wei left, progress on Li Auto’s foundation model R&D slowed, and some departments went several months without receiving new requirements.

In the long run, Li Auto is using organizational adjustments to move closer to a startup-style operating model. In the short term, it will have to absorb the disruption that comes with that reset.

KrASIA features translated and adapted content that was originally published by 36Kr. This article was written by Xiao Man for 36Kr.