Author | EV Editor | Dexin
Observations from Li Auto Power Day
On the new-generation L8 launched in June this year, Li Auto announced the removal of CATL battery options, adopting only its self-developed batteries. Subsequently, the Li Auto L6 and i8 also switched to self-developed batteries. It can be said that starting with this move, Li Auto single-handedly sparked a trend of "de-CATL" across the industry.
On the surface, it appears that Li Auto's battery supply chain is shifting from heavy reliance on CATL to using Sunwoda and CALB as cell OEM partners. The underlying essence is that automakers wish to break away from the "turnkey solution model" and reconstruct a new order with battery manufacturers.
Last week, Li Auto held its first Electric Technology Day, Power Day. The content focused on its self-developed "Lianshan" electric technology system, including the Lianshan Battery, self-developed silicon carbide (SiC) electric drive, self-built ultra-fast charging network, and cloud-based battery early warning platform.
Power Day itself was a response to public and potential owner concerns following the "de-CATL" shift: After replacing CATL batteries, will the reliability of Li Auto's self-developed Lianshan Battery decline? Is this genuine engineering iteration or merely a marketing concept?
What Does "De-CATL" Actually Mean?
First, two points must be clarified: Li Auto has not explicitly stated that it will cease purchasing CATL cells in the future; nor does Li Auto currently have plans to build its own cell manufacturing factory.
The binding between automakers and CATL began in the early stages of the NEV industry's development. CATL's capabilities were not limited to cell production; more importantly, it simultaneously possessed mature capabilities in BMS, thermal management, PACK integration, and long-term reliability verification.
Therefore, when automakers initially developed NEVs, most manufacturers tended to adopt battery companies' "turnkey solutions," which included cell manufacturing, BMS algorithms, thermal management strategies, and PACK integration, while the OEMs were responsible for vehicle installation adaptation.
Purchasing a complete turnkey solution essentially meant directly using a full battery system that had already been defined and verified by the battery manufacturer. The benefit was maturity and low trial-and-error costs, but the limitation was that the automaker lacked permission to modify the underlying parameters and control strategies of the battery system. The logic of system calibration prioritized the design goals of the battery manufacturer rather than addressing the high-priority needs of the automaker regarding overall vehicle energy consumption, thermal management, etc.
Through its self-developed Lianshan Battery system, Li Auto aims to establish a new division of labor: Li Auto is responsible for top-level definition of the battery system, BMS algorithms, PACK structure, whole-vehicle thermal management coordination, and cloud monitoring; battery manufacturers such as Sunwoda and CALB manufacture basic hardware like cells according to specifications provided by Li Auto.
However, one concept requires attention here: While basic cell specifications can be standardized, this does not mean that mass-produced cells are completely identical. So-called cell standardization refers to the fact that the industry's baseline safety frameworks, material systems, and basic performance indicators for cells have formed unified entry thresholds, with the gap in baseline safety capabilities among top-tier manufacturers continuously narrowing.
Yet, even if drawings and specification sheets are identical, subtle differences in manufacturing processes such as material purity, coating, winding, and formation across different factories and batches still lead to individual characteristic dispersion among cells, i.e., consistency issues. Standardization aligns the lower limit, while consistency involves controlling batch dispersion; the two are not contradictory.
Li Auto currently employs a multi-supplier parallel strategy, retaining CATL cell versions for some models.
The primary goal of this move is supply chain diversification, reducing risks associated with capacity, price, and delivery rhythm caused by a single supplier, while introducing competition among multiple cell manufacturers to optimize cost and delivery flexibility.
This supply structure is commonly adopted by current leading automakers. Li Auto's operation does not inherently equate to a technical upgrade or downgrade; the final outcome depends on the OEM's system management capabilities.
Analogous to automotive R&D, an OEM may not cast engine blanks themselves, but powertrain calibration, whole-vehicle matching, and fault strategies must be controlled by the OEM. Applied to batteries, cells are basic hardware, while the control, thermal management, safety strategies of the battery system, and whole-vehicle coordination are the systemic foundations that determine actual vehicle range, charging, and safety performance.
Two Routes of Lianshan Battery: Parameters vs. Reality
At Power Day, Li Auto publicly disclosed two technical routes for the Lianshan Battery, targeting pure electric/high-end range-extender vehicles and household range-extender vehicles respectively.
The performance data below comes from Li Auto's official laboratory test results and has not yet undergone large-scale long-term real-world user validation.
The first route is the 5C ternary battery. Official disclosures state that this battery can reach a peak charging power of 520kW. Paired with Li Auto's 5C ultra-fast charger, it can achieve10%-80% replenishment in 10 minutes under laboratory conditions. The new L8 equipped with this battery has a CLTC pure electric range of 430km.
Laboratory tests indicate that after a single cell thermal failure, this solution is unlikely to trigger pack-wide thermal runaway, reducing thermal failure risk by 90% compared to previous traditional solutions. Regarding cycle life, which users also care about, it improves by 30% over the previous generation 2C battery, with a target of maintaining above 75% State of Health (SOH) after 8 years/160,000 km.
The engineering goal of this solution is to address temperature rise and thermal runaway risks in high-rate fast-charging scenarios, serving high-speed long-distance replenishment needs.
The cost of 5C fast charging is higher cell temperature rise during the charging process, placing higher demands on thermal management and BMS real-time power limiting strategies. The degradation performance under long-term high-frequency fast charging requires validation through long-term real-world data.
The second route is the Lithium Iron Phosphate (LFP) battery. This battery set is primarily positioned to adapt to Li Auto's extensive range of household range-extender models, prioritizing a balance between cost, lifespan, and low-temperature usage experience.
LFP materials have low cost and good thermal stability, but suffer from two long-standing pain points: range attenuation due to reduced low-temperature activity, and SOC estimation drift. According to official introductions, this LFP solution relies on a self-developed ATR adaptive state-of-charge recognition algorithm and whole-vehicle waste heat recovery coordinated thermal control. Under laboratory conditions, winter equivalent range can improve by 30km, while simultaneously narrowing SOC estimation errors.
The two battery sets are not isolated designs but are embedded within Li Auto's complete self-developed Lianshan electric system, which includes SiC electric drive, ultra-fast charging network, and cloud quality early warning platform.
On the electric drive side, Li Auto uses self-developed SiC power modules, highly integrating the motor and electronic control to optimize system energy consumption and power response, matching the output characteristics of the Lianshan Battery for whole-vehicle calibration.
The replenishment network is a prerequisite for the 5C battery experience. The speed of ultra-fast charger deployment, site utilization rates, and maintenance quality directly impact the landing experience of the fast-charging solution. Li Auto currently has 4,200 ultra-fast charging stations and23,000 ultra-fast chargers (including 5C and 2C). According to plans, Li Auto will expand 5C ultra-fast charging guns to 8,300 before the Spring Festival of 2027.
The Lianshan Quality Early Warning Platform relies on battery operating condition data returned by vehicles to perform status monitoring and early fault warnings, shifting battery management from "post-failure repair" to active full-lifecycle monitoring.
Accompanying this is the Lianshan Battery Lifetime Guarantee Plan: If spontaneous combustion occurs due to cell quality issues, or if battery fire results from collision or chassis scraping, and the vehicle reaches scrapping standards, compensation in the form of a new vehicle with the same configuration will be provided.
Note the sequence: This guarantee is a user risk commitment given by the manufacturer. Currently, the complete claims assessment standards, disclaimer clauses, and judgment boundaries have not been fully disclosed. The commitment itself can change users' perception of risk but cannot be directly treated as technical evidence of battery system reliability. Its ultimate value depends on the subsequent execution of accident cases.
Objectively Viewing Three Types of Engineering Challenges for Lianshan Battery
The parameter data above can be considered laboratory results from filtered ideal operating conditions. When implemented across hundreds of thousands of vehicles, over several years, and for users in different regions, this multi-supplier self-developed system faces several realistic challenges:
First, consistency management of multi-supplier cells.
Even if cells are produced according to the same specification sheet, process differences between different OEM factories will lead to individual cell dispersion.
When two sets of cell sources are installed in parallel, Li Auto needs to build a unified incoming inspection, cell screening, and PACK end-of-line testing system to screen out cells exceeding dispersion limits, ensuring that battery pack performance approaches each other after installation from different suppliers.
Compared to CATL's integrated solution, whose advantage lies in completing cell production, PACK integration, and system verification within the same framework, accumulating decades of complete batch data and quality control experience, Li Auto's multi-supplier system requires continuous iteration in incoming quality control, cell sorting, and BMS adaptation to different cell characteristics.
This is not unsolvable, but it requires sufficient installation volume and time accumulation for validation.
Second, the gap in long-term aging data under 5C fast charging.
All current lifespan and thermal safety data come from accelerated laboratory tests, which cannot fully replicate complex real-world conditions, such as the long-term impact of different ambient temperatures, user charging habits, charging SOC ranges, bumps and vibrations on internal cell structures.
5C belongs to high-rate charging, imposing strict requirements on cell temperature rise control. Short-term fast charging performance meeting standards does not equal controllable capacity decay, internal pressure growth, and fault probability after 3-5 years of high-frequency fast charging. These conclusions can only be substantiated after accumulating large-scale user samples.
Third, increased complexity in after-sales service and OTA due to multi-version parallelism.
Installing multi-brand cells and multi-version battery packs simultaneously means that BMS calibration, OTA strategies, and fault diagnosis logic all need to adapt to different cell characteristics. Spare parts reserves, maintenance training, and after-sales diagnostic platforms need to expand synchronously.
The biggest difficulty lies in fault diagnosis: For the same fault phenomenon, root causes and handling solutions corresponding to different cell sources may differ. Whether this after-sales system can keep up with the O&M needs of multi-version batteries is a challenge at the long-term operational level.
Industry Perspective: Cell Standardization vs. System Capability Game
To date, the baseline for cell fundamental safety and basic performance has clearly converged in the NEV industry, with the gap in lower limits among products from top manufacturers narrowing. However, convergence of basic capabilities does not mean that different cells perform identically in extreme conditions and long-term aging, which is why multi-supplier consistency management remains an engineering difficulty.
In the past, when choosing NEVs, users were accustomed to treating the cell brand as a safety yardstick. This was a simplified judgment rule naturally formed because reliable alternative bases for judgment were lacking in the early industry stage.
Currently, an increasing number of leading automakers are attempting to seize the definition rights of the battery system. This does not mean cell manufacturers no longer possess system capabilities, but rather that OEMs hope the calibration targets of the battery system prioritize serving the whole vehicle's energy consumption, thermal management, and safety strategies, rather than the automaker accommodating the product framework of the battery manufacturer.
Automaker self-developed systems are not simply "replacing cell manufacturers." The benefits of this route include supply chain elasticity and customized capabilities at the whole-vehicle level; the cost is significantly increased R&D, verification, quality control, and after-sales costs and workload on the vehicle side, with risks borne by the automaker itself.
Returning to the initial question: After supply chain diversification, is the Lianshan Battery reliable?
At the current point in time, a clear conclusion of "reliable/unreliable" cannot be given.
The essence of this solution is an attempt at industrial chain division of labor: Cells are outsourced to major external manufacturers for OEM production, while the OEM controls system definition and algorithms. There are clear benefits, accompanied by corresponding engineering risks. The final yardstick for judging the Lianshan Battery is not the parameters at the launch event, but rather the long-term decay, low-temperature real-world tests, fault statistics of massive real vehicles over the coming years, and the actual execution of warranty terms in accidents.
For consumers, there is no need to simply apply labels: One should not judge a downgrade immediately upon seeing "non-CATL cells," nor default to assuming this new system is fully mature based solely on laboratory data from manufacturer launch events. The answer sheet for technology needs to be submitted to time and real user scenarios for review.