The Investor's Podcast — CATL: The $280 Billion Company You've Never Heard Of

1:08 The setup: 40% share, $280B, and nobody in the West can name it

• Manish opens on the anonymity: "this company controls 40% market share in a growing industry. Most people probably have never seen their products. But if you have been in a Tesla or an electric BMW or a Mercedes, there is more than a one-third chance" the pack is theirs.
• Scale: market cap ~$280B, "roughly the size of Shell." Stig admits he didn't know whether to pronounce it "cattle" or C-A-T-L.
• The thesis in one sentence: "more than a battery company… fast becoming the backbone of the energy infrastructure. It powers EVs. It stores renewable power. And now increasingly it's becoming the core power infrastructure player in the AI data centers."

2:39 Founding story: a flawed Bell Labs patent, Apple's iPod, and a spin-off engineered around nationality

• 1990s: Robin Zeng is technical director at SAE Magnetics, a TDK subsidiary making hard-disk components; he sees the consumer-electronics lithium opportunity and quits to found ATL (Amperex Technology Limited — "that's CATL without the C") with two colleagues.
• They had no chemistry of their own, so they licensed a Bell Labs patent with a fatal flaw — cells swelled after repeated charging. Zeng's team solved it and "turned this defective license into a product," winning Apple's iPod in 2003 and later the iPhone; ATL became the world's largest smartphone battery maker.
• TDK bought ATL and kept Zeng running it — "a full circle." But when Zeng saw China designating EVs a strategic industry with state support for domestic firms, foreign-owned ATL was structurally disqualified. So in 2011 he gathered Chinese investors, spun the EV-battery division out as CATL — 85% Zeng and co-investors, 15% retained by TDK.

7:09 The two customer relationships that made the company: BMW, then Tesla

• 2012: BMW wanted an EV and disliked every battery on the market. It approached BYD first — BYD declined to supply external customers — so BMW gambled on an unknown and gave CATL the China supply contract.
• Not a purchase order: "they basically embedded their engineers inside CATL's facilities to co-design the battery." Manish's read: "that education in German engineering and a stamp of approval from a global automaker is what put CATL on the map."
• "If BMW gave CATL the credibility, it was Tesla that gave them the global stage" — CATL entered the Tesla supply chain around 2020 when Tesla localized in China, "and that's when the growth of CATL went vertical": revenue ~$7B (2020) → >$60B (2025), an 8× move in five years.

9:41 Founder-controlled, and Ralph's first (favourable) management read

• Zeng owns ~22% and the founding team >35% — "founder controlled, deeply technical, R&D focused, with skin in the game and a very long time horizon." He is "sometimes called the invisible king of the global battery supply chain."
• Ralph, unprompted, adds the value-investor screen: "we as value investors always are looking for a company led by a founder who has more than a minor stake," estimating Zeng's personal position at >$53 billion.
• The culture datapoint he weights: Zeng grew up poor, and "the manufacturing facilities are in a city of about 400,000, located very close to his hometown… he's expressed his desire to give back and reward loyalty." (He returns to this at the end of the episode as a forensic-accounting positive.)

11:13 The competitive field: BYD at 16%, LG Energy Solution at 9%, and a widening Chinese lead

• Nine straight years at #1; "CATL is about two and a half times larger than the nearest rival." BYD ~16% (vertically integrated, mostly self-consumed), LG Energy Solution ~9%.
• The fast-charge exchange as evidence: BYD's March product goes 10→97% in ~9 minutes; CATL's answer a month later goes 10→98% in under 7 minutes. "In this R&D competition between CATL and BYD, the Chinese lead over everyone keeps widening."
• Berkshire footnote: BYD was "Berkshire's famous investment in China led by Charlie Munger back in 2008… roughly 20 times before Berkshire fully exited last year."

13:57 The flywheel: the moat is not on the checklist, it's the loop between the items

• The key reframe: run the checklist — cheapest producer, biggest R&D budget, best technology — "and you might assume that the moat is one of those. But the moat is that these advantages feed each other and create a flywheel."
• The loop, verbatim: "being the biggest makes them the cheapest. Being the cheapest makes them the most profitable. That profit funds a large R&D budget which helps them build the best technology and win more customers which makes them bigger still."
• Both ends are extended: upstream into mining stakes ("so they control the raw materials") and downstream by designing cells "straight into the car platforms." Conclusion: "for a competitor, it isn't about trying to catch up on one thing. They are chasing a flywheel that just keeps spinning faster every year."

14:28 Switching costs and product complexity — the two supporting moats

• Switching cost: a pack is engineered into a platform years before production and must clear "safety, crash, and durability testing before a single car gets shipped," so "a design win isn't a one-year order. It's locked in for the entire life of that platform — typically about 5 to 8 years," producing sticky, predictable revenue. Customer breadth: Tesla, BMW, Mercedes, Volkswagen "and nearly every Chinese EV maker except BYD."
• Complexity: "a grid-scale storage system looks like a giant block of battery, but a single gigawatt-hour installation can contain on the order of several million individual components" — Zeng's line is that it has "more components than a Boeing 747." The systems must last 20+ years, "so the reliability bar is very high and that makes them hard to manufacture at scale."

17:19 Under-priced engine #1: batteries as the shock absorber between GPUs and the grid

• The mechanism: "you have tens of thousands of GPUs all computing in parallel, then they pause for a split second to synchronize, then they fire another compute cycle. This massive parallel cycle creates spikes in power demand — sometimes swinging hundreds of megawatts in seconds. But legacy power grids are not designed to ramp up and down fast enough."
• The product: "that's the gap CATL's batteries fill. They sit in between this spiky compute load and the steady power grid. They absorb these surges." As total data-center demand climbs, "the appetite for this kind of storage buffering grows with it."
• Ralph's sizing: the industry expects data centers to need twice as much power by 2030 — 200 GW of continuous power. His scale ladder: 1 GW ≈ Seattle or San Francisco; 10 GW ≈ New York or Los Angeles; 200 GW ≈ the entire state of California. "All that additional power requires storage to furnish that clean steady power."

19:46 Storage is the better business, and CATL is buying the rest of the stack

• CATL is already the world's #1 energy-storage supplier, and "the storage segment carries higher margins than the core EV battery business because of the complexity and the durability requirements."
• The part Manish finds more interesting: "CATL isn't just selling batteries to these data centers. It is now systematically assembling a full stack energy ecosystem — storage through power distribution inside the data center. They are even buying into the power equipment makers and servicing the entire energy stack."
• The payoff if it works: transformation "from just an energy storage supplier to a strategic infrastructure partner for the entire AI compute buildout."

20:53 Under-priced engine #2: LRS — license, royalty and service

• The constraint: "CATL is effectively locked out of owning a factory in the US for geopolitical reasons." The workaround: "rather than fight it out… the American partner builds, owns and runs the manufacturing plant while CATL just provides the license of its technology and collects a royalty fee."
• The live example: Ford's Michigan battery plant — "Ford owns the plant, land, equipment and hires the workforce. CATL has no equity and only offers battery chemistry and technology." Economics undisclosed, "but if it is in the 3 to 4% range that these deals are typically done at, then that's a very high margin, capital light income," extendable beyond EV cells into AI-data-center storage.
• Ralph's supportive analogy before he turns on it: Kevin O'Leary on Shark Tank "frequently offers to invest, but not for an equity stake — he's looking for a royalty or a licensing fee… a huge return on the investment without the big investment in brick and mortar." His caveat lands immediately: "the scrutiny by the US lawmakers is a big but… a licensing deal of 3 to 4% of revenue would be wonderful, but it could easily get wiped out by the simple stroke of a pen."

25:46 Financials I: why flat revenue is not flat business

• Stig's question is the right one — revenue was flat 2023→2025. Manish: "the biggest cost for CATL is the raw materials… mostly lithium, and lithium prices became dramatically cheaper… and CATL passed those savings on as lower prices to its customers. So the company sold far more battery volume, made more profits, but the headline revenue barely moved."
• The second anomaly: "last year CATL earned about $11 billion in profit but generated close to $20 billion in operating cash flow — nearly double," funding capex, acquisitions, dividends and buybacks while still growing the cash balance.

26:50 Financials II: negative working capital, the Amazon playbook, and the rule that ends it

• The source of the gap: "CATL has a large negative working capital — it sells through its inventory and collects money from customers months before it has to pay its suppliers. So in effect its suppliers are its largest lenders, funding the growth almost interest free. This is similar to Amazon's playbook: get paid on e-commerce sales today, pay suppliers much later, and let that float fund the flywheel."
• The bull's own risk flag: "there is a limit to how much you can stretch your payables, and the Chinese authorities are asking large firms to start paying their SME suppliers sooner. So part of their interest-free float might unwind and a period of readjustment in how it funds its growth."
• His mitigation: "much of CATL's payables are to large suppliers, which are outside the SME rule… I don't think it will break the model. But it's still a real headwind to watch out for."

28:26 Golden shares, state support, and "even a pig can fly"

• What a golden share is: "a special management share typically given to the state or a government fund… the state gets some veto powers," used mostly in social media where content control matters; it "does not impact the day-to-day operations." Verdict on CATL: "a founder-controlled company with no publicly disclosed golden share."
• What state support actually was: (1) a restriction that "functionally locked out foreign competition from the China support program and gave CATL a captive domestic market to scale up" — scrapped after Elon Musk negotiated Tesla's entry in 2019; (2) consumer purchase subsidies that created EV demand — fully phased out at the end of 2022.
• The culture line worth keeping: at the 2017 subsidy peak Zeng sent staff a letter translating roughly as "if you stand where the wind blows, even a pig can fly" — a warning that a subsidy tailwind "can make even a mediocre company look great." The test came: "the wind stopped… LG, Panasonic, they all came back in," and CATL still leads globally on share with above-industry margins. "Robin Zeng didn't confuse policy support with a durable advantage."

31:33 The bull's own bear list: Pentagon list, price war, chemistry disruption

• Geopolitics ("the biggest"): "CATL is effectively walled out of the US market, which is the second largest market for EV batteries. The company is on the Pentagon's list of companies that work for the Chinese military" — contested hard, executives sent to Washington, "but so far without success." LRS is the counter, "and it's not a done deal yet."
• Domestic price war: "brutal… so much so that Robin Zeng himself has publicly asked the industry to stop competing purely on price." Counter: storage/AI-data-center demand is "a new, higher margin, faster growing segment where CATL is number one," though "storage is not immune to price pressure."
• Technology disruption: "as a rough rule, every two years the energy density in the battery packs climbs by about 20%," with solid-state or sodium-ion able to leapfrog CATL's LFP/NMC base. Counter: biggest R&D budget in the industry, no single-chemistry dependence, sodium-ion already scaling — "the disruptor and the incumbent may just turn out to be the same company here."

33:11 Ralph on LRS: defensive capitulation, a ghost in the machine, and "LRS leakage"

• The reframe: "LRS has often been spun as asset-light growth. In reality, it may be better spun as a defensive capitulation. By surrendering direct ownership and customer relations to Ford and GM, CATL is admitting that direct Chinese ownership is politically toxic under the US Inflation Reduction Act. Accepting a mere 3 to 4% royalty fee turns that invisible king into a low rent IP landlord."
• The structural fragility: "this model makes CATL what I would call a ghost in the machine — vulnerable to a stroke of the pen where US regulators can sever the IP at any time and leave CATL with zero physical assets to reclaim."
• The compounding cost: "by handing over the blueprint for high-density packs to Ford and GM… CATL is training the very competitors who seek to exclude it. Let's call it LRS leakage." Manish concedes it fully — "after all, CATL itself learned the manufacturing skills from other players such as BMW. So it's definitely possible for US manufacturers to learn from CATL."

37:57 The forensic bear case: the treadmill, the windfall margin, the float unwind

• Treadmill effect: "in a recent fiscal period, CATL reported 21.8% growth in shipped gigawatt-hour volume, yet topline revenue contracted by 9.7%. This price deflation acted as a massive 130% drag on growth." The cause is contractual: "CATL's long-term agreements utilize raw material indexation — the company is contractually obligated to pass manufacturing efficiency gains and commodity savings directly to their OEMs. CATL is running exponentially faster to stay in the same place financially… surrendering its pricing power to maintain utilization."
• Windfall margin: "the recent ~15% profit margin expansion achieved while revenue was shrinking is a mathematically temporary windfall. It's not a new baseline. It doesn't indicate technological superiority" — a "cost wedge" that regresses to "a tighter 11 to 12% historical band."
• Float unwind: the 2× cash-flow-to-profit delta "exists because CATL uses its dominant position to squeeze suppliers by [stretching] days payable outstanding… as this interest-free loan is called in, CATL's funding model will face a painful readjustment, stripping away the cash used for buybacks and dividends. The interest-rate float like that enjoyed by Buffett and other insurance companies will go away."
• ESS as high beta: "the energy storage system is a cyclical buffer to slowing EV sales… sensitive to utility capital budgets, global interest rates, and AI spend" — not a defensive segment but "a high beta on the global macro environment."

40:35 Ralph's four gravitational pulls: Europe, commodity beta, CBAM, key man

• Utilization risk: new plants in Hungary and Germany carry "massive fixed overhead. Any minor slowdown in European EV adoption due to subsidy cliffs will leave this capacity unabsorbed, turning growth assets into margin anchors."
• Commodity concentration: "despite the vertical integration, CATL remains a shadow commodity trading house. Its stock price may be viewed by bears as a proxy for lithium and nickel volatility, not a reflection of R&D prowess."
• Protectionism: "the European protectionist wall — the EU carbon border adjustment mechanism, CBAM — acts as a targeted tax on Chinese cells."
• Key man: "CATL is inextricably tied to Robin Zeng's personal standing, with 22% ownership, deep ties to Beijing. Any shift in his political standing represents an existential unhedgeable risk for shareholders. And I will remind folks about Alibaba." Manish's answer: CATL "operates in a sector that China wants to grow," and Zeng "maintains a low profile and rarely speaks in public," so the odds of offending regulators are low — "low probability, potentially high impact, and very difficult to hedge."

45:09 Valuation: ~18× EV/EBIT, 17% ROIC, and a 15%-a-year base case

• The arithmetic: market cap ~$280B, net cash brings enterprise value to ~$250B, against "$14 billion of operating profit over the last four quarters — about 18 times EV/EBIT," and "the P/E multiple is roughly 21 times, which for a global leader like CATL is neither demanding nor cheap."
• The quality: ROIC ~17%, ROE ~25%. The base case: "a good enough chance that the operating business roughly doubles in value over the next 5 years. And if the multiples remain where they are, you can expect about 15% annual return before you add a dividend yield of another point or two."
• Upside/downside named explicitly: upside if the AI-data-center and LRS engines surprise; downside "the domestic price war intensifies and the LRS model doesn't scale in the US due to geopolitical tension."

47:12 The dual listing: an inverted 30–35% Hong Kong premium, with TSMC as the reference

• Same company, same one-share-one-vote, same dividend per share: A-shares Shenzhen 300750, H-shares Hong Kong 3750 — "but after adjusting for the FX rate, the Hong Kong shares trade at a big premium of roughly 30 to 35%… that's actually opposite to the norm," where mainland lines trade rich because domestic investors face capital controls.
• Why it inverts here: "pure supply and demand. The Hong Kong float is relatively small and the demand from global investors is quite high," and the two lines are not fungible — you cannot buy in one market and sell in the other.
• Access and the compression thesis: institutions can use northbound Stock Connect; "for smaller retail investors, Hong Kong shares are the only option." CATL has been widening the float (2025 HK IPO, 2026 follow-on placement) "and yet the global demand was so high that the premium hasn't compressed much." Manish's long-run view: 10–20%, "using TSMC as the reference — TSMC has averaged around a 15% premium — but how and when that gap will compress is hard to say."

51:08 Stig's grid explainer: base load, intermittency, and why electricity is "a pipe and not a bucket"

• Why coal and nuclear persisted: low marginal cost, but also base load — demand is highly predictable (a morning peak before work, an evening peak when the TV and the cooker go on), "so you can more or less see that demand of electricity on a curve."
• Renewables broke the supply side: solar and wind are intermittent, and wind is not even linear — above a cut-out speed of roughly 90 km/h the turbine pitches its blades and stops.
• The physical constraint storage removes: "from the moment the wind hits the turbine's blade, electricity has to be used the moment that it's being made. You can think about it as it's a pipe and not a bucket" — plus transmission losses that grow with distance. "With today's battery technology much of that is changing."
• Grid frequency as the stability variable: 50 Hz in Europe, 60 Hz in the US, for purely historical reasons (AEG picked 50 to fit the metric system; Westinghouse picked 60 because "lights flickered less"). Japan still runs both — German equipment in Tokyo, American in Osaka — and cannot share power across the seam without expensive converters. The stakes: "every power plant and every appliance needs to be in sync. If supply and demand fall out of balance even for seconds, the frequency drifts, and if it drifts too far, the way that the grid protects itself is through blackouts."

55:52 The stadium analogy — why an AI data center is worse than a football crowd

• The picture: "a stadium with 70,000 people, and a person here is a GPU. Everyone is talking, eating, walking around at random… it all blends into a steady hum. That's a normal data center."
• The AI version: "it's 70,000 people who watch the same game. Everyone cheers at the same time whenever there's a goal… and the grid isn't built for that."
• The twist that makes it worse than a stadium: "you might get a few goals throughout those 90 minutes, but whenever you train a model it will take you weeks… and then you have this cheering and then steady hum, and it happens every second or two. So it's incredibly tiresome for the grid to work with." CATL's batteries "sit between the data center and the grid and absorb those shocks."
• Then Stig turns it into the sharpest challenge of the episode: "everything I just said proves that someone is going to sell a lot of batteries. It doesn't prove that it's going to be CATL. The storage market is growing faster than CATL's storage shipments — competitors are taking their share right now. Why does CATL win the market and not just participate in it?"

58:31 The answer: it wasn't demand, it was a 97% utilization ceiling

• Manish concedes the datapoint: CATL's storage installations grew ~30% last year against industry growth of ~80%.
• The explanation is capacity, not competitiveness: "CATL's utilization rate last year was around 97% — meaning they sold pretty much what they could produce," with new capacity easing the constraint this year.
• The evidence it is already unlocking: "CATL's energy storage sales for the first half of 2026 grew by about 88% year on year, and the storage volume nearly doubled year on year."

59:54 Battery leasing and swapping — a possible business-model change, not just a feature

• The consumer framing: "you buy the car, but not the battery, and you subscribe to the battery" — Stig's analogy is a phone plan, where "you don't pay the $2,000 upfront… and whenever the phone gets old you swap it for a new one." Since the battery is "roughly a third of the price of an electric car," removing it "lowers the sticker price significantly and allows more people to enter the market."
• The economics of swapping: close to 2,000 CATL stations and growing, packs co-developed with major Chinese automakers ("it locks in the customers"), network cost shared with partners, and an average swap time under 100 seconds — decisive for taxis and heavy commercial vehicles where "the vehicle earns money every hour it moves" and "time at a charger is lost income." Most Chinese drivers still charge at home; for them it's "a nice to have… certainly not a need to have."
• The model change Stig finds interesting: "they stop being a company that sells your battery once. Instead the battery stays in the pool that CATL and its partners own and operate, and they never lose the battery. It comes back, it gets checked, gets reused and eventually recycled — so CATL owns the whole lifetime of the asset," while the owner stops carrying degradation risk (his iPhone analogy: with an EV, an aging battery "knocks thousands of dollars off the car's value").
• Both flag the constraint: "the batteries are heavy, stations are expensive, the automakers have to agree on standards," and falling charge times erode the time advantage for small vehicles — which is why CATL is adding superchargers to the swap stations so one site serves both.

1:04:37 150+ strategic stakes, DeepSeek, and how to value a portfolio you can't see

• Scale of the practice: CATL holds stakes in more than 150 companies, spanning upstream mines, downstream automakers and power-infrastructure players.
• The standout: DeepSeek — "the lab that shook everyone in Silicon Valley in early 2025," which "in June took outside money for the first time — roughly $7 billion at a valuation above $50 billion USD," with CATL putting in $700 million (Stig flags these as press reports, not filed terms). Rationale: "AI centers are power hungry. CATL are investing in a future customer."
• The comparison Stig draws: "Nvidia committed more than $40 billion to equity stakes in 2026," and structurally "Chinese companies hold roughly 10% of their market cap [in strategic stakes], whereas Americans, it's only 4%" — Chinese stakes being older (15 years) and broader (supply chains, mines, customers, political alignment), though all such marks are "on paper prices" from the latest round.
• How Manish treats it in the valuation: "more strategic than financial… the simplest way is the long-term equity investment line item on the balance sheet — about $11 billion last quarter, slightly less than one times operating profit. I did not account for it in the 18×. If we include them at book value, it drops to about 17 times." On DeepSeek specifically: part of the data-center integration play, "although investing in a foundational AI model lab is going to the extreme end of vertical integration."

1:09:19 Why the format exists: "argue the other side better," and the anti-echo-chamber rule

• Manish's closing principle is Munger's: "you have to argue the other side better than the other player. So understanding the bear thesis is extremely important — and seeing whether you are missing anything."
• Stig's design rationale for Bull vs. Bear: "whenever you are bullish about a company you end up speaking with other people who are also bullish, and you end up in the same echo chamber… giving someone the permission to be the bear" removes the social cost — "if people are excited about a stock, you generally don't want to be the guy who's like, well, you're probably all wrong… that guy doesn't get invited back." They plan to swap roles next quarter.

1:16:18 Ralph's forensic screen: management first, and the trait he learned to look for

• Career context: CPA → forensic accountant investigating "partners cooking the books, massive Ponzi schemes… and multi-billion dollar bank frauds" → investor. His honest caveat: fraud work "was a helpful foundation certainly, but it wasn't until I found a community of lifelong learners… that I truly learned how to dig into the fundamental analysis of a business."
• The common denominator across frauds: perpetrators "start off as just purely a narcissist… but they're also sociopaths. When you study somebody like Madoff who injured so many people — it takes a really good sociopath to be that kind of person."
• How that became a screen: "in my study and looking at companies, one of the most important things I look at is management, because I've investigated these bad guys. So I look for that bad trait and what they look like and how they frame themselves." He half-dissents from Buffett — "if you've got a choice between management and industry, you pick the industry… I think that's true, but I put a big but on that… even if you've got a great industry like AI, you still have to understand management, their motives and what they're going for."
• Applied here: "like Robin [Zeng] at CATL — he's giving back to his community. He wants the people in his community to thrive and grow. Most sociopaths and narcissists don't feel that way. They care about number one."