Battery suppliers are shifting from being answers to becoming options

When hospitals buy AI imaging solutions, the first thing they check is whether the model is accurate. Later, doctors ask more practical questions: Can it integrate with PACS? Can reports be written back into the system? Who is liable for missed diagnoses? No matter how high the metrics are, if no one in the workflow dares to sign off, the product fails to land.

CATL's position among automakers over the past few years was much like that "answer." Li Auto started using them from the ONE; the MEGA debuted with the 4C Qilin battery, and the new-generation MEGA uses the 5C ternary lithium. Automakers are willing to highlight charging speeds at launch events because they value CATL packaging yield rates, capacity, safety, and brand together. It saves automakers from having to explain a lot.

Now the customer list is starting to shrink. Li Auto announced that its self-developed batteries are already installed in the L8, L6, and i8, with plans to roll out across all future models; users who lock in orders after September 7 will switch to Li Auto's self-developed 5C ternary batteries. Xiaomi officially announced the Dragon Armor battery, partnering with CALB and Sunwoda; AITO has also introduced CALB and Gotion High-Tech, with an 81 kWh battery pack visible on the M6. The market immediately interpreted this as "de-CATL-ing." CATL's stock price dropped, closing at 335.49 yuan on September 8, down 3.65%, wiping out 54 billion yuan in market cap that day.

Beyond lost orders, control over definitions is even more critical

Automakers switching suppliers appears on the surface to be about cost, supply, and differentiation. Digging deeper, it's about batteries shifting from being a "black-box capability provided by others" to a "system they want to define themselves." Previously, CATL provided a good cell, and automakers used it to tell stories about range, fast charging, and safety. Now Li Auto, Xiaomi, and AITO all want to keep variables like cells, PACKs, vehicle strategy, after-sales degradation, residual value, and OTA updates in their own hands.

This is very similar to medical AI. Early AI imaging companies gained fame through detection rates, but later hospitals started demanding quality control, audits, and doctor feedback workflows. The model shifted from being a selling point to just part of the process. Battery suppliers face the same reality. Once fast charging, long range, and safety become standard features in high-end cars, the halo effect of a single supplier fades. A shrinking customer list likely means the industry is starting to treat them as a "replaceable module."

Self-developed batteries: The hard part is long-term delivery certainty

The easiest thing for automakers to underestimate when developing batteries in-house is the validation cycle. Cell formulas, manufacturing yields, batch consistency, thermal runaway, low-temperature degradation, crash repairability, and residual value assessments—none of these can be covered by a PowerPoint presentation. Li Auto investing in Sunwoda Power and establishing a joint battery company with them looks like concrete action. But when it comes down to the end-user, the questions are simple: How much range do you lose in winter? Is charging slow? How much does it degrade after three years? If there's an accident, can you afford the repairs? This is just like actual user feedback in medicine: Has it been clinically validated? The key is whether you dare to trust it long-term.

It's the same with batteries. What automakers want is long-term certainty. CATL's strength lay in providing certainty to major clients: capacity, quality, delivery, and safety endorsements. Now that multiple suppliers are entering the game, automakers are swapping multi-source supply, in-house development, joint ventures, and equity ties for certainty. I wrote previously about copper foil expansion, noting that scaling up depends on who can smooth out standards, collaboration, and delivery. This is even more obvious with batteries.

For CATL, the stock drop is just market sentiment. What's worth watching is how the business model changes. If they continue selling cells, margins will get squeezed. If they move towards materials, equipment, processes, testing, and recycling to form standards, customers will find it harder to leave. Just like in medical AI, we ultimately have to deliver equipment, imaging workflows, quality control, reports, and audits together. The clearer the boundaries of responsibility, the more willing customers are to use the service long-term.

After automakers turn away, batteries retreat from being brand stories at launch events back to being systems engineering inside factories. Only those who can turn uncertainty into verifiable, replicable, and accountable delivery deserve to keep earning high margins.