In short: The whole episode. Manish (bull) frames it as a category error to call it a battery company: "this company is more than a battery company… it is fast becoming the backbone of the energy infrastructure" — 40% global EV-battery share for nine straight years, ~$280B market cap ("roughly the size of Shell"), revenue from ~$7B in 2020 to >$60B in 2025, and on track for ~1,000 GWh of battery output in 2026 ("enough electricity storage to power all of Singapore, or the city of New York for about a week"). The moat is explicitly plural: "the moat is that these advantages feed each other and create a flywheel… 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" — plus 5–8-year platform switching costs and storage-system complexity ("more components than a Boeing 747"). Two under-priced engines: AI-data-center storage (batteries buffering GPU load spikes "swinging hundreds of megawatts in seconds"; H1-2026 storage sales +88% y/y, volume nearly doubled; higher margin than EV cells; a full-stack ecosystem including power-equipment stakes) and LRS licensing (Ford's Michigan plant — Ford owns everything, CATL takes a 3–4% royalty, "very high margin capital light income"). Valuation: EV ~$250B against ~$14B of trailing operating profit = ~18× EV/EBIT (~17× including the $11B long-term equity-investment book), ~21× earnings, "neither demanding nor cheap," ROIC ~17%, ROE ~25%, "a good enough chance that the operating business roughly doubles in value over the next 5 years… about 15% annual return before you add a dividend yield of another point or two." Governance: founder-controlled (Zeng ~22%, founding team >35%), no publicly disclosed golden share, and Zeng's "even a pig can fly" letter as the anti-subsidy-dependence culture. Ralph (bear) contests the durability of every number: the treadmill (+21.8% volume, −9.7% revenue — raw-material indexation contractually hands efficiency gains to the OEMs, "surrendering its pricing power to maintain utilization"); the windfall margin (~15% net "is a mathematically temporary windfall. It's not a new baseline" — regresses to 11–12%); the payables float ("an interest free loan from its supply chain… 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"); LRS as "defensive capitulation" that makes CATL "a low rent IP landlord," "a ghost in the machine… US regulators can sever the IP at any time," and trains its own competitors — "LRS leakage"; plus Hungary/Germany utilization risk, "a shadow commodity trading house" proxying lithium and nickel, the EU CBAM tax, the Pentagon list, and key-man risk ("I will remind folks about Alibaba"). Stig (host) stays out of the call but says what excited him most was "batteries that were not for vehicles, but really for stabilizing power grids and AI data centers." Dual-listed: A-shares Shenzhen 300750, H-shares Hong Kong 3750, same economics, H at a 30–35% premium (reverse of the norm; small float vs global demand), Manish expecting long-run compression to 10–20% on the TSMC precedent.
CATL makes batteries — about four out of every ten electric-car batteries sold anywhere in the world, and more grid-storage batteries than anyone else. Almost nobody in the West can name it, partly because it sells to carmakers rather than to you: if you have ridden in a Tesla, an electric BMW or a Mercedes, there is better than a one-in-three chance the pack under the floor was theirs. It is worth roughly $280 billion, about the size of Shell.
The bull case, made by Manish Karira, is that the usual list of advantages misses the point. Yes it is the cheapest producer, has the biggest research budget and the best technology — but those are not three separate moats, they are one loop. Being the biggest makes it the cheapest; being the cheapest makes it the most profitable; the profit pays for the research; the research wins more customers; that makes it bigger again. A rival cannot beat that by catching up on one item, because the wheel is already turning faster than it can run. CATL has also bought into the mines that supply its raw materials and designs its cells directly into car platforms years before those cars exist — and because a battery has to pass crash and durability testing before a single vehicle ships, once it is designed in it stays for the five-to-eight-year life of that model.
Two newer businesses are the reason he thinks the stock is mispriced. The first is powering AI. A data center full of chips does not draw electricity smoothly — thousands of processors compute together, stop for a fraction of a second to compare notes, then all fire again, which makes the power draw jump by hundreds of megawatts in seconds. Power grids cannot move that fast. CATL's batteries sit between the building and the grid and soak up the surges, like a shock absorber. That business earns better margins than car batteries and its sales nearly doubled in the first half of 2026. The second is a workaround for politics: CATL is barred from owning factories in America, so instead it lets Ford build and own the Michigan plant while CATL supplies the recipe and collects a small royalty on everything made there — almost pure profit, with no factory to pay for.
The bear case, from Ralph, a career forensic accountant, is that the profits are borrowed rather than earned. Shipments rose 21.8% in a recent period while revenue fell 9.7%, because CATL's contracts oblige it to hand cost savings straight back to the carmakers — running faster to stand still. The unusually fat margin it earned while shrinking is, he argues, a temporary accident of falling input costs that will drift back to the historical 11–12%. And the flood of cash on its books is not really profit: CATL collects from customers months before it pays its own suppliers, an interest-free loan from the supply chain that Beijing is now forcing companies to repay faster. He also thinks the Ford deal is a defeat dressed as a strategy — it hands the blueprint to the competitors trying to shut CATL out, and Washington can cancel the licence with a signature, leaving CATL owning nothing. Add a European carbon tax aimed at Chinese cells, factories in Hungary and Germany that need Europe to keep buying EVs, and a company welded to one founder's standing with Beijing, and you have real risk.
On price, the shares change hands at about 18 times operating profit and 21 times earnings, with returns on capital in the high teens — Manish's word is "neither demanding nor cheap." His base case is that the business roughly doubles over five years, which at an unchanged multiple is about 15% a year plus a small dividend. One practical wrinkle: the same company is listed twice, in Shenzhen and in Hong Kong, with identical rights — but the Hong Kong line costs 30–35% more, because the Hong Kong float is small and foreigners can only buy that one. He expects that gap to narrow toward 10–20% eventually, using Taiwan Semiconductor's US listing as the precedent, but will not guess when.
45:37The company has a high return on invested capital of around 17%. And a return on equity of about 25%. So there is 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.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.