Author: Xu Shanshan, Editor: Li Motian
On September 7, Li Auto announced that for the new-generation MEGA, orders locked from 3:00 PM that afternoon would switch from CATL's 5C batteries to Li Auto's in-house 5C batteries. The following day, CATL's market capitalization dropped by 3.65%, which was interpreted as automakers "de-CATL-izing."
That is not all. The Li Auto i9, launched on September 16, initially adopted CATL batteries. However, the head of Li Auto's product line later stated that the i9 Home newly ordered from September 17 onwards would be equipped with in-house 5C batteries.
According to Li Auto, the material system and structure of the cells for its in-house batteries are developed independently, while Sunwoda and CALB are responsible for production; the battery pack is developed and produced by Li Auto. In September this year, Li Auto announced a planned capital increase of CNY 2.65 billion in Sunwoda.
Coincidentally, in July this year, Xiaomi released the "Longjia Battery": Xiaomi is responsible for product definition, leading battery pack design, and participating in cell design and development, as well as overseeing full-process quality management.
By September, Xiaomi announced its two strategic partners for the Longjia Battery, CALB and Sunwoda. CALB develops the electrochemical system and materials according to requirements; Sunwoda provides cell manufacturing capabilities, and both parties are building dedicated production lines.
Simply put, Xiaomi sets the vehicle requirements, leads the battery pack definition, and gets involved in cell development and supplier quality management, while battery manufacturers provide cell technology, engineering, and manufacturing capabilities. Subsequently, the Pengcheng N70 and N90 series equipped with Longjia batteries were launched, with CATL not listed among the battery suppliers.
In addition, XPeng is also preparing to take action. During the interview session at the XPeng G9L launch event in September, He Xiaopeng stated that starting this year, "we will handle all battery production ourselves," but immediately added, "not the cells."
In other words, all three companies call their products "in-house batteries," but their actual scopes of work differ: XPeng is currently only preparing to manufacture battery packs, Xiaomi is participating in cell development, and Li Auto has pushed its R&D into cell materials and structures.
However, the timing of these three automakers' defection coincides exactly with CATL's "Global Quality Open Month" event, making it hard not to make a statement. Ni Jun, CATL's Chief Manufacturing Officer, stated at the event:
Knowing how to build cars does not necessarily mean knowing how to build batteries; professionals should still do professional work (BYD: ?).
This covert competition is not limited to simply "reducing reliance on CATL." The common question they raise is: Who should control the design and manufacturing of a battery?
Seeds Planted Long Ago
When it comes to developing in-house batteries, Li Auto moved the earliest and acted the fastest.
The first-generation MEGA launched in 2024 used the Qilin 5C, jointly developed by Li Auto and CATL. According to the statements at the time, Li Auto proposed the vehicle requirements, technical standards, and validation methods; CATL was responsible for capabilities in materials, cells, structures, and manufacturing processes.
From underlying materials to battery pack architecture, the two companies worked together on solutions and experiments. Li Auto invested over 200 personnel, while CATL invested about 800. The scale was 5 to 8 times that of a standard battery project [1]. R&D started in 2020, taking three years to reach mass production.
This year's change goes a level deeper. According to Li Auto, it starts with positive and negative electrode materials, electrolytes, and formulations, validates cell prototypes in its own laboratories, and then works with partner battery manufacturers to translate the solutions into mass-producible engineering designs.
Meanwhile, the battery pack, management system, and vehicle control are also led by Li Auto. Sunwoda and CALB are only responsible for the production of cells and battery packs.
Compared with the Qilin 5C back then, the R&D division of labor and cooperation methods between upstream and downstream have changed. However, regarding the specific ownership of intellectual property for each formulation and process, as well as the cost differences between the two factories producing the same solution, there are no answers yet.
Given that the timing of these new forces' defection to in-house R&D coincides exactly with the prevailing narrative that CATL's profitability dwarfs that of all automakers, it inevitably sparks market speculation about automakers "reducing reliance on CATL."
According to previous supply chain research, before Li Auto switched to in-house batteries, CATL's share in Li Auto's product line was close to 70%, and it is expected to drop to about 50% for the full year of 2026.
Considering Li Auto's support for its second supplier, Sunwoda, this share may decline further in the future.
Previously, Xiaomi only had two models, the SU7 and YU7. The standard version's LFP batteries also included orders from CATL. Combined with ternary lithium batteries, CATL's share within the Xiaomi system is close to 70%.
However, during the supplier designation for extended-range models, CATL was excluded because CALB and Sunwoda's products had already been validated on a large number of extended-range models. Nevertheless, for Xiaomi's subsequent pure electric models, CATL still holds the majority share.
CATL's bargaining power in the industry chain is obvious to all. Therefore, the diverse in-house battery initiatives are ultimately interpreted as automakers further supporting second-tier suppliers and seizing power from the upstream.
In November last year, Ma Donghui, President of Li Auto, stated at the earnings call that the delay in the delivery of the Li Auto i6 was mainly affected by "core component supply and capacity ramp-up." Although he did not explicitly name anyone, it was interpreted as dropping hints in advance for introducing a second supplier.
By December, Li Auto i6 owners received a survey on battery selection preferences. Choosing Sunwoda batteries would allow for faster vehicle delivery and also include a free extended warranty.
One Li Auto pursuing in-house R&D might not be a big deal, but if many "Li Autos" suddenly emerge, the situation becomes unpredictable.
The Power Struggle
CATL's weight in the industry chain can be illustrated by several sets of data.
By the end of 2025, CATL had CNY 263.6 billion in accounts payable and notes payable on its books, which can be roughly understood as the money it still needs to pay to suppliers; on the other hand, the accounts receivable and notes that CATL's customers need to pay to it amount to CNY 121 billion. The former is 2.18 times the latter, and in the first half of this year, the ratio was 2.52.
In the first half of this year, approximately CNY 57.8 billion of CATL's accounts payable entered supplier financing arrangements. Essentially, this means suppliers can use CATL's payment vouchers to secure financing, with CATL making the actual payment on the agreed-upon date.
In the first half of this year, CATL accounted for 46.7% of the domestic passenger car battery installation volume. If looking solely at ternary lithium batteries, its market share is 75.2%. For automakers using ternary lithium batteries, finding alternative suppliers with sufficient scale is particularly difficult.
A company that can both extend upstream payment terms and collect downstream advance payments can hardly be said to lack bargaining power. At CATL's scale, the description might shift to "dominant."
In the second half of last year, as battery raw material prices rose, CATL's capacity utilization rate once reached as high as 102.6%. Automakers not only had to wait for production schedules but also had to grit their teeth and accept slight price increases.
Consequently, automakers either tacitly introduce second and third suppliers or roll up their sleeves and dive in to do it themselves. The goal is to reduce CATL's order quota, and the trigger for this movement lies in CATL's dominance.
In 2023, Li Auto demonstrated its intention to support Sunwoda by dispatching a large number of technical personnel to the company. During the same period, CATL and Li Auto signed a supply guarantee and strategic agreement, but it was only valid until the end of 2025.
In 2025, the two parties renewed a five-year strategic agreement. However, concurrently, Li Auto and Sunwoda established Shandong Li Auto Battery Co., Ltd., with each holding a 50% stake, pushing the in-house battery initiative a major step forward.
Ultimately, CATL's dominance and the automakers' resistance both stem from their respective business instincts. The power struggle between upstream and downstream in the industry chain is an objective reality in any industry. It is actually a necessary process for the industry to mature and should not be over-interpreted at the moral or value level.
In the early stages of the new energy industry's development, high-quality batteries were a scarce resource; whoever controlled battery capacity controlled the bargaining power. More importantly, since 2021, CATL has elevated its R&D investment to the tens of billions level, which inevitably requires recovery through scale and profit.
Now, however, the industry has entered a mature phase, and batteries are no longer a scarce resource. Coupled with a brutal price war in the terminal market, automakers' gross margins are being extremely compressed. Under these circumstances, if a single supplier captures 40% of a vehicle's BOM cost, automakers are no different from mere wage earners.
CATL's issue, perhaps, lies in its attempt to extend the dominance accumulated during a specific historical period into today's mature industry, but automakers are no longer willing to accept this and are ready to flip the table.
Introducing second-tier suppliers can indeed enhance automakers' bargaining power to a certain extent, but there are some things that even second-tier suppliers cannot provide.
The Truth About In-House Development
First, it must be clarified that a perfect battery does not exist, and different vehicles have different definitions of perfection.
For instance, performance cars require high discharge power, so energy density can be compromised; ride-hailing cars need long cycle life, so energy density can also be compromised; high-end cars demand comprehensive excellence without shortcomings, so costs can be negotiated. Moreover, unlike plug-in components, batteries are becoming increasingly deeply coupled with the chassis and thermal management systems.
However, battery manufacturers mostly deliver general solutions to automakers, which must cater to the demands of the mainstream market. Even with customization, it is impossible to completely overturn the formulation and start from scratch. Therefore, a primary objective for automakers to develop in-house batteries is to reclaim the right of definition. This can be gleaned from the evolution of automakers' battery procurement models.
In the early stages, automakers procured a complete battery pack assembly (PACK) from battery manufacturers, encompassing all components such as cells, BMS, thermal management, and the casing, known as the black-box model. During this period, automakers had to awkwardly adapt their chassis designs to fit the batteries, and most of the profits were captured by upstream suppliers.
Later, automakers split the assembly in half, procuring only cells from battery manufacturers, then sourcing BMS from third-party companies and customizing the casing. They handled the final assembly themselves, saving the costs previously incurred for system integration.
Around 2022, CTC/CTB (Cell to Chassis / Cell to Body) technology gained popularity, turning batteries into chassis structural components, which made continuing to procure third-party casings meaningless. At this point, the number of automakers building their own PACK production lines and developing in-house BMS significantly increased.
In other words, over a long period, automakers' level of involvement in battery R&D and production has been steadily deepening.
The current wave of new forces launching in-house R&D is because this involvement has reached a critical point: the definition of the battery.
What is the definition of a battery? It refers to how various specifications of batteries and battery packs are designed. In the past, battery manufacturers were responsible for both definition and production. Today, automakers want to seize complete product definition rights from battery manufacturers.
Ma Donghui, President of Li Auto, previously noted that automakers used to be unable to obtain complete material formulations, process parameters, and individual cell testing data. Therefore, Li Auto requires its partners to share this information, linking manufacturing data with post-installation driving, charging, and battery aging data to identify issues.
For example, battery manufacturers provide standards for allowable output power under various conditions, and the vehicle controls power accordingly. With in-house R&D, Li Auto can leverage its understanding of cell states to adjust control algorithms independently, reducing unnecessary power restrictions while still protecting the battery.
Another example is fast charging. A cell's ability to withstand high C-rates is merely a prerequisite; the vehicle's thermal dissipation capacity, the cabin's air conditioning load, and the charging pile's power supply will all affect how long it can sustain charging in high-temperature weather.
This round of changes will redraw the division of labor for certain battery orders: automakers will propose more specific cell solutions, demand access to manufacturing data, and assume greater R&D and quality responsibilities; battery manufacturers will be responsible for mass-producing these solutions and can also continue to sell their independently developed products to other customers.
The context behind CATL's remark that "knowing how to build cars does not necessarily mean knowing how to build batteries" refers to whether hundreds of millions of cells can achieve sufficient consistency in the production and manufacturing process. This statement hits the nail on the head regarding why automakers have not easily built their own factories.
However, the automakers' strategy is to decouple the design and definition phases from battery manufacturers while leveraging the know-how accumulated by these manufacturers in production and manufacturing, thereby indirectly enhancing their own bargaining power in the industry chain.
This definition phase is absolutely unattainable from CATL and BYD, but for second-tier battery manufacturers lacking orders and eager to increase their market share, it is a timely blessing.
Just as Apple's initial in-house development of A-series chips was not to become a chip company, but to deeply integrate the chips with the iOS system, develop required specific features, and keep the right to define the core experience in its own hands, the automakers' purpose is exactly the same.
However, this automaker-led initiative to reclaim definition rights will inevitably relegate battery manufacturers back to the role of contract manufacturers.
Once the right of definition is lost, the premium margin becomes difficult to sustain. For battery manufacturers accustomed to holding pricing power, this is perhaps the most painful outcome.
References
[1] Twice the development cycle, eight times the manpower investment: Decoding the R&D journey of the Li Auto MEGA Qilin 5C battery, Gasgoo
[2] Dialogue with technical heads of Li Auto and CATL: How extended-range automakers develop 5C batteries, 36Kr