On October 9, CATL (SZSE: 300750) finally saw a long-awaited rally, with its A-shares surging more than 5% at one point during intraday trading. By the close, the stock was still up approximately 3.84%.

Over the preceding two months, CATL's share price had been steadily declining, accumulating a drop of more than 20%. Debates surrounding automakers' in-house battery development kept intensifying, with the market worried that as OEMs pushed their R&D deeper into battery cells, CATL's orders and pricing power would come under pressure.

Li Auto has disclosed that the first batch of its new MEGA model will be delivered with CATL's 5C ternary lithium batteries, and once its self-developed battery production is ready, it will fully switch to its own 5C ternary lithium batteries. Xiaomi also announced it would jointly develop batteries with Sunwoda and CALB.

On September 28, Li Xiang posted a response to the debate over in-house battery development, opposing the narrative of "de-[brand]-ification" and recounting CATL's involvement in developing batteries for the Li ONE and MEGA. Li Auto wants to continue deepening materials research, connecting data across cells, battery packs, and vehicles, while also welcoming partners to participate.

SNE Research data shows that in the first seven months of this year, CATL's global power battery usage grew 26.6% year-on-year, with its market share rising from 38% in the same period last year to 39.9%.

The cells still come off professional manufacturers' production lines, but more OEMs are beginning to participate in formulation, structure, and manufacturing data. Automakers want to make batteries cheaper and better suited to their vehicles, while CATL must renegotiate business with these increasingly battery-savvy customers.

Automakers Want to Renegotiate Battery Prices

On September 17, He Xiaopeng said after the launch of the XPeng G9L that starting this year, batteries would be made by XPeng itself. He then added: "XPeng doesn't make cells."

What XPeng has taken on is battery pack structure, battery management systems (BMS), thermal management, and vehicle integration, while continuing to purchase cells externally. Li Auto has already gone deeper into material formulations and cell structures, but it hasn't moved all production into its own factories either. Li Auto's management told Wall Street CN that battery packs are self-developed and self-manufactured by Li Auto, while cells are left to professional partners to produce; the 5C cells for the new i6 were developed by Li Auto and produced by Sunwoda and CALB.

Xiaomi's Dragon Armor battery was jointly developed by both sides. Xiaomi is responsible for product definition, leading battery pack design, and participating in cell material, formulation, and process development, while Sunwoda and CALB provide cell technology and manufacturing capabilities. Xiaomi has also dispatched quality personnel to participate in production management through a "shadow factory" approach.

Li Auto also wants to unlock the information inside the battery. Knowing only external performance without visibility into materials and processes can be compared to a "black box." Ma Donghui, at a previous media briefing, called batteries lacking complete formulations, production parameters, and per-cell inspection data "gray boxes."

Once a battery is delivered to an automaker, it still has years or even longer of use ahead. Charging, driving, temperature, and degradation records come from the vehicle, while material and process records come from the factory. Li Auto hopes to connect both ends, enabling traceability when anomalies occur and feeding usage performance back into next-generation development. How much manufacturing data to open up requires specific agreements with partners.

When developing a vehicle, automakers must also coordinate battery pack shape, cabin space, crash protection, and heat dissipation. After increasing fast-charging current, how to balance temperature, lifespan, and sustained charging time must also be considered within the whole vehicle. Once deeper into cells, these trade-offs also enter materials and structure development.

The conditions for automakers to make batteries are also different from a few years ago. Song Yining, head of Leapmotor's battery product line, links the current opportunity to the maturity of the lithium iron phosphate supply chain. In his view, over the past three to five years, materials, equipment, and production environment control for this chemistry have all reached more mature conditions. If automakers have stable supporting demand, they can learn and invest around their own models.

Leapmotor has already reached cell production. At its September technology day, the company disclosed that the first production line at its Zhongling cell factory, a joint venture with CALB, has begun production, with capacity sufficient to support approximately 500,000 to 600,000 vehicles. Song Yining said self-developed cells can reduce costs by about 10% compared to external procurement, and self-production and external procurement will run in parallel long-term.

One way Leapmotor saves money is by having the same cell specification serve multiple vehicles. Changing a model requires production line adjustments, process re-validation, and yield ramp-up. The longer a specification is used, the more existing testing, control, and manufacturing experience can be carried forward. Song Yining also noted that Leapmotor has placed cell, module battery pack, enclosure, and liquid cooling plate production in the same industrial park in Wuyi, reducing packaging and transportation. What's saved here is repeated development, downtime, and handling, not just procurement prices.

Having several vehicles share one type of cell requires coordination in model development, and equipment investment must be amortized through sustained orders. Professional battery makers serving multiple customers often have to handle each one's customization requirements, whereas automakers can first coordinate among their own few models.

Geely is still allocating capacity across several paths. This restructuring of Chongqing Yaoning assets involves transferring the uncompleted battery factory project under Jiyao Tongxing to Shidai Geely, a joint venture between Geely and CATL, through an equity transaction arrangement. A person familiar with the matter revealed that after Geely's model sales grew, its own cell capacity could no longer meet demand, and the adjustment aims to accelerate new capacity coming online. Geely emphasized to Wall Street CN that its medium-to-long-term goal remains 50% overall battery self-sufficiency for the group. R&D continues to be invested in, while new factories are advanced by partners — the two arrangements can coexist.

A senior industry observer told Wall Street CN that some automakers already have the ability to evaluate batteries, and other battery makers are also improving. "Now that he has this capability, he thinks you're too expensive." Only when automakers better understand materials, specifications, and validation methods can they break down quotes, determine which performance to keep and which designs can be adjusted, and then compare different suppliers using the same requirement.

As negotiating room expands, automakers' investments also increase. R&D, prototyping, and validation that were previously included in procurement prices must now be partially completed by their own teams; choosing to build their own factories also means bearing the risks of equipment and capacity utilization. How far in-house development goes ultimately requires each company to calculate orders, personnel, and capital together.

The Leader's Confidence

Li Auto's battery switch first brings opportunities for Sunwoda and CALB to enter new projects, and will also reduce CATL's supply on some models. But counting an automaker's entire battery demand as orders lost by CATL would overestimate the speed and scale of the switch.

A DBS report dated September 17 estimated that Li Auto's full-year battery demand of approximately 30–40GWh equals about 3%–4% of CATL's shipment target for the year. The bank believes the switch will proceed gradually, other business growth can buffer the impact, and it retains its judgment on CATL's industry position. The report also lowered earnings forecasts and valuation multiples considering industry taxes, costs, and customer supply chain adjustments. According to Sina HK's summary, a JPMorgan report in September argued that investors may have overestimated the speed and scale of automakers' self-supply and underestimated leading battery makers' manufacturing capabilities.

In-house development like XPeng's brings another kind of change. Cells are still procured from battery makers, while battery packs and some development work are taken over by the automaker. Suppliers deliver fewer structural components and take on less work, so both revenue and costs change, and profit cannot be deducted in proportion to the revenue decline. CATL's half-year report does not separately list profits for cells versus complete battery packs, making it currently difficult to calculate a complete profit-and-loss picture for this kind of adjustment.

Domestic automaker procurement is only part of CATL's business. In the first half, the power battery business gross margin fell 1.78 percentage points year-on-year to 20.63%; energy storage batteries contributed approximately 53.3 billion yuan in revenue. By region, overseas business accounted for about 30% of total revenue with a gross margin of 29.97%, while domestic was 21.16%. These regional figures include power batteries, energy storage, and other businesses, and the gross margin for domestic power batteries is still not separately listed. When analyzing second-quarter results in July, DBS also linked the sequential gross margin decline to sales mix and the dilution effect of raw material price increases.

A March estimate by Soochow Securities concluded that on equivalent lithium iron phosphate cells, CATL's material cost is about 0.02 yuan/Wh lower than second-tier companies, and manufacturing cost about 0.01 yuan/Wh lower, the latter without considering depreciation differences. Based on this estimate for specific cells, even at similar selling prices, procurement, material usage, and manufacturing efficiency can create profit differences.

CATL's half-year report disclosed that in the first half, battery system capacity utilization was 94.86%, including both power and energy storage businesses. For asset-heavy factories, the fuller the production lines run, the lower the equipment depreciation and fixed expenses spread across each unit of product. After new factories secure automaker orders, they still need to complete debugging, ramp-up, and batch validation before they have a chance to approach the costs calculated at the design stage.

A senior industry observer told Wall Street CN that whether other battery makers can take on demand from multiple automakers depends on development, service, and supply capabilities. "It's not like the meal is placed in front of me and I can just eat it." After a battery maker wins a new vehicle project, engineers, equipment, and production lines all need to keep up; when several customers launch new cars simultaneously, these tasks must be completed at the same time.

The difficulty of mass production also lies in the fact that after one sample is made well, thousands or tens of thousands must be close to it. Capacity, internal resistance, and self-discharge vary, and once assembled into battery packs, these affect performance and long-term degradation. Zeng Yuqun emphasized consistency at the September World Power Battery Conference, which is exactly this issue. Automakers can lead formulation and structure, but manufacturing partners still must turn designs into stable batches.

After the battery is installed in a vehicle, development is not over. BMS can arrange charging strategies, but how much current a cell can withstand at different temperatures and aging levels requires material information and long-term validation to answer. A CATL representative explained that its models combine electrochemical mechanisms with real vehicle data, including battery records from old vehicles used for five or even eight-plus years. Automakers provide driving, charging, and vehicle operation data, while battery makers provide process, material, and failure information. Putting both ends together helps analyze anomalies and adjust strategies.

Fast charging and safety require even more joint development. According to CATL's explanation of system safety, structural deformation from a collision may affect electrical connections and heat dissipation, thereby changing thermal runaway risk. When increasing charging current or changing cells, hardware protection, heat dissipation, and control strategies must also be re-matched. Optimizing the BMS alone, or swapping in a better-performing cell, still requires testing under whole-vehicle conditions.

For next-generation models, increasing charging speed and reducing weight will impose new requirements on materials and processes. CATL invested 11.377 billion yuan in R&D in the first half. Even after automakers master the design of current cells, they can still jointly develop next-generation products with battery makers.

In-House Development Rewrites the Battery Division of Labor

Automakers once actively promoted Bosch brakes and Qualcomm cockpit chips, and now they are increasingly willing to give batteries a name of their own. A senior industry observer told Wall Street CN that these suppliers were once technical labels for OEMs, and Bosch still has technical strength today. By borrowing suppliers' names, automakers make it easier for users to understand and trust a technology.

After automakers brand jointly developed batteries with their own names, manufacturing and part of R&D may still remain with suppliers. The billing on the label changes, but cooperation does not necessarily end, and product credibility must be rebuilt through long-term delivery. CATL's existing brand recognition will also continue to enter automakers' procurement considerations.

When BMW announced its sixth-generation battery technology in 2022, it disclosed that it participates in developing cell formats, chemical systems, and integration solutions while awarding supply contracts to CATL and EVE Energy. It built pilot lines to understand the production process, while mass manufacturing still relies on suppliers' expertise. Automakers may understand cells better, but professional battery makers can still win manufacturing orders.

CATL itself is also moving toward complete vehicles. In May this year, it announced the Panshi chassis collaboration with Turkish automaker Togg, with Shidai Intelligent and Togg jointly developing three models on a new B-segment platform, with the first planned for mass production in 2027. Togg leads user experience, vehicle requirements, and digital architecture, while CATL's side provides the chassis solution, developing batteries, electric drive, thermal management, and more together.

BMW participates in cell development while letting suppliers handle mass manufacturing; the Togg cooperation covers more systems. CATL can take on projects of varying depth, and companies like Sunwoda and CALB can also use new projects to accumulate development and delivery experience.

The more cooperation content there is, the more things must be negotiated in the quote. If an automaker wants a unique cell, the factory may have to modify production lines; order scale and duration will affect how this investment is amortized. Whether jointly developed results can be used for other customers, to what layer manufacturing data is opened, and how anomalies are traced must also enter mutual agreements.

Li Auto's management emphasized in previous communications that whether contract manufacturing or external procurement, reasonable gross margins should be left for partners. Partners still need to improve processes, maintain production lines, and develop new products later. If automakers want this work to continue, they must leave room for investment in the price. Costs saved through collaboration may also be passed on to car buyers amid vehicle competition, and may not all become profit for one party.

Going forward, OEMs may have fewer internal specifications, but battery makers may face more designs. One automaker may unify cells across several models, while other customers may not adopt the same dimensions, materials, and charging requirements. While taking custom orders, suppliers must also make equipment and processes as shared as possible to avoid investing in a whole new production line and validation set for every additional customer.

If more automakers define their own cells and hand mass production to professional manufacturers, battery solutions will become more fragmented, while manufacturing may still see concentrated orders. For a manufacturer like CATL, serving one more customer means incorporating more designs into its existing manufacturing system, not merely adding another dedicated production line. Companies that can balance customized development and batch delivery will have the opportunity to continue expanding business within automakers' in-house development projects.