Episode 286 - CATL.1

[00:00:00] Welcome, welcome everybody. My name is Jeff Towson, and this is the Tech Strategy Podcast from Techmoat Consulting. And the topic for today, innovation lessons from my visit to CATL. Uh, I was out in Fujian Province a week or so ago visiting the headquarters of CATL. Spent a couple days with them. Uh, they had a big announcement, which was their big, uh, new testing center called ESBL, which, uh, I've written about.

Thanks to CATL for flying me out and putting me up for that. And I also want to talk about, uh, the company in general and kind of what I learned and especially in the area of innovation in batteries. For those of you who aren't familiar, CATL is the leader by a long margin in rechargeable batteries that go into electric vehicles.

In China, 50% of all the cars on the road, CATL batteries. Globally, 40%. [00:01:00] Uh, and now they're moving into energy storage, which are these massive stationary batteries that you put in things like data centers and power grids and things like that. So-- And they're number one in that too. They've been number one in power batteries, which are for cars and things like that.

They've been number one for nine years for energy storage, like grids. They've been number one for five years. So this is a hugely successful company, and I was kind of trying to learn whatever I could. Anyways, so that will be the topic for today. I'll probably start out with a rant about what's coming out of the US in terms of banning Claude, uh, Fable, which is kind of a sea change moment in US and global tech.

I think I'll tell you what I think it means, but then I'll get into what I thought from this trip out there. That'll be the topic for today. Let me do, uh, my standard disclaimer here. Nothing in this podcast or my writing or website is investment advice. The numbers and information from me and [00:02:00] any guests may be incorrect.

The views and opinions expressed may no longer be relevant or accurate. Overall, investing is risky. This is not investment, legal, tax advice. Do your own research. And that's super important here because CATL is a public company, so I am not giving anything in the range of investment or financial advice or anything like that, okay?

This is just strategy thinking and lessons, educational only. And with that, let's get into the topic. All right. Let me start with a bit of a political, eh, sort of strategy rant because there has been, like, a sea change moment in the last couple days, um, which is really, uh, pretty profound coming out of the US government, which basically for those of you who weren't sort of following, uh, yeah, the US government on a very quick announcement basically prevented Anthropic from letting its frontier level models, which are Fable [00:03:00] 5 and Mythos 5, basically said, "Look, non-US citizens can't use it," and it doesn't matter if they're overseas or if they're l- working in the US, including people who work at Anthropic.

And this just came out of the blue. It was very fast. I think they were given 45 minutes to comply. It was under national security grounds. Who knows? And then, you know, Anthropic basically just pulled the whole model because, one, I don't think they could implement that so quickly, if at all, and two, I think it was just probably a decent thing to do.

But yeah, that's a defining moment, and I'll, I'll tell you why I think-- And one, I think it, this actually gives us a f- pretty clear window into what will probably happen in terms of AI going forward within the US, within China, and in between. Okay, so go back to 2019. I was, um, visiting the Huawei headquarters in Shenzhen [00:04:00] a couple months after the entity list had sort of dropped out of the US, and that was basically a US, you know, decree that, you know, you couldn't work with Huawei.

And they applied it to US companies, and then they tried to apply it extra territorial, extra territorially. So they said, like, Taiwan companies cannot work with Huawei, even un- though they're not under US jurisdiction. So it was a very sweeping move, and at the time, it was going to have a lot of short-term implications, but the big long-term implication was it convinced literally every tech company in China in one moment, you cannot be dependent on the US technology supply chain, including GPUs, because it has been politically weaponized.

And there was no debate. It was just across the board from the highest levels of government to provincial government to every single tech [00:05:00] company. Everybody knew you had to act now And within the next year or so, I would-- You know, the companies I follow, I look at their annual reports. Pretty much every annual report I looked at in the next year, 2020, within the risk section, there was a new clause saying, "Our technology is heavily dependent on US supply chains, and these could be cut off at any time."

So when that sort of moment happens, one, it creates a risk that has to be disclosed in the annual reports, the filings. It forces every CEO to, to create a backup plan, an alternative supply, and it convinced the whole country in one move. And now I didn't know how, how quick that was going to happen, but it turns out that it's happened pretty quick, that China has created an, a pretty much an alternative AI tech stack, which, you know, I've been talking a lot about the cloud matrix, which is Huawei's da- [00:06:00] AI data center structure, which is kind of an alternative architecture where you can get similar performance to a US data center, AI data center, with less powerful chips, uh, using really good interconnects, some different architecture, a lot of power, and a lot of cooling.

Okay, and then last month, we had Huawei basically announce their new Tao scaling law, and, um, you know, that basically put forward a new architecture for lo- for circuitry that was going to use vertical scaling, 3D architecture, to overcome the limitations they have within lithography and creating ever smaller and more dense, uh, circuit boards And according to them, it will be able to match two nanometer, 1.5 nanometer chips within a couple years without having access to what NVIDIA can do.

It will use a different scaling law. Anyways, that [00:07:00] was their... So, okay, that happened, I think, much faster than people thought. It looks like the Huawei announcement is the real thing. This isn't theory. They've already deployed this in 286 different consumer products manufactured at scale. So it looks like it's real.

We'll see. Anyways, this is similar to that, but for every basically executive in the United States, and maybe internationally, especially internationally, they basically all had their, their entity list-like moment where it's very clear now you can be a business based in California, you can be using, you know, um, these sort of frontier level models for doing, let's say, you know, biochemistry and drug development, and it was-- I think it was Jason Calaca- Calacas, Kalinakis, the guy on All-In Podcast.

He basically said this in the last [00:08:00] day, that, you know, we do sort of biochemistry and biological research to design new drugs using Anthropic, and based on this event, we are going to have to build our own model or use an open source model, cause we can no longer count on this being available and not, and not being cut off.

So it's created this absolutely proven risk that you have to build against for American companies, for international companies, the same way it was done for Chinese companies. So it's a big sea change, and I think that's, yeah, it's a defining moment. Now, where does that lead? Okay, so let's assume that all the US and especially the non-US companies, they know they can never really be dependent on US frontier models cause they can get cut off any day of the week, and it just happened.

It's not theory. So where does that lead? [00:09:00] Well, Bill Gurley, the, uh, the big venture capitalist, you know, really smart guy, he put out a, a tweet which is pretty devastating, I think. Here's what he said. I'll read it to you He basically said, and I think this is a quote, but it came from someone quoting him and not his direct post, so it may be not 100% on.

He said, quote, "If a credible Western open frontier player does not emerge, the consequences cascade quickly. This is the inverse of the early internet wave. In the 2000s and the 2010s, Western companies, Google, Facebook, Amazon, Apple, Microsoft, dominated globally while China carved out its own walled garden."

This, the AI version flips that dynamic on its head. Without a [00:10:00] credible Western open frontier player, the only open models capable of running entirely, e-entire economies are made in China. If US policy further restricts Chinese open weight access on national security grounds, the US ends up with two or three closed cathedrals serving the US market, and the rest of the world picks the AI stack that is free, capable, self-hostable, and not embargoed.

Europe, Africa, Southeast Asia, Latin America, India, the Unit- the Middle East, roughly six billion people, Chinese open models become the global default by 2030, and the U-United States ends up technologically isolated from the majority of the world's AI users. We have done it to ourselves, un- end quote. The argument there, in my opinion, is, look, China engaged for a long time, and still [00:11:00] today, in what I would call info nationalism, where they didn't allow YouTube, they didn't allow Google, any of those businesses into their country because they didn't want foreign companies controlling the information flows in their country, which are very powerful in terms of culture, um, politics, news.

They, they sort of created a walled garden, and everyone within China used, you know, Weibo, WeChat, but everyone outside of China used the Silicon Valley models, Google, and so and so. He's arguing this is the reverse. The whole rest of the world is going to have to go for the open source, downloadable, hostable, and most importantly, non-embargoable models, which these days are coming out of China, because if you're dependent on the Western model, you can get cut off any day of the week.

So the US becomes the walled garden. The rest of the [00:12:00] world will go open source. Right now, the open source models on the table are overwhelmingly Chinese. Do I agree with that? Kind of. I agree with the premise that this is going to have a chilling effect on US AI frontier models, and we could very likely see it come down to two or three major players in the US that are politically acceptable, because the politicians are deciding now, and that is what everyone uses for these frontier models Will open source models emerge in the US?

They might. Will the Chinese models that are pretty popular right now ever, will they get blacklisted in the US? I think that would be the second shoe to drop. After this national security decree, if the US government blacklists the, the Chinese open source [00:13:00] models, then that moves us very closely to what Bill Gurley sort of described.

Now, I think there's a problem with what he just said. We have to look at the AI data centers. It's about the models, but it's also about the compute infrastructure. Now, the US has built 5,000 plus da- AI data centers. Uh, China has about 500. Those massive data centers like, you know, Colossus in Tennessee and the OpenAI ones, they are overwhelmingly using the proprietary models of these massive companies like Gemini and OpenAI.

They are not using open AI models, OpenAI models. So the advanced models, okay, you have proprietary and you have open source, but when it comes down to the advanced compute architecture, it's overwhelmingly these big proprietary models, and I don't think that's going to change, and [00:14:00] especially if the open source ones out of China get cut off in the US.

I think that combination of the big data centers in the US plus two to three proprietary models basically working hand in glove with the government, I think that gets you a walled garden. So I, I think you have to put both of those together. So let's say we know what the US looks like, and we kind of know what China's going to look like.

They have their own AI tech stack now. They already have good models. So the big question then for me is, it's really two things. Number one, is the US going to put a blanket blacklist on Chinese open source models? That would be the second shoe to drop. Then the question becomes, what is the rest of the world going to do?

Are they going to choose The models coming out of the US, whether they're proprietary or let's say there's an open source one, or do they now view those as politically too risky and they have to go with something else? And I think that will [00:15:00] probably be Chinese models for the simple-- I mean, my reason for that would be like, look, I have never seen China try to control the information flows outside of China.

I've never seen them do that. I've seen the US do that quite aggressively through the Silicon Valley companies. I've never seen China try to do that. They seem to be, and I'm talking generally speaking, China, Chinese companies, they seem to be quite happy to create open source, open weight, totally downloadable models that anyone can use for whatever they want.

They don't seem to be doing that. So if they keep that trend up, then yeah, I think the world will go Chinese open source. I think that's the default assumption for me at this point in time. It wasn't last week. It is now. And then what AI architecture are they going to use for that? Well, Huawei, Alibaba, and now Tencent are moving [00:16:00] fairly aggressively to build AI data centers outside of China.

They're definitely in Southeast Asia. They're showing up in Mexico and Brazil right now. So I think that gets you the models, and I think that gets you the AI architecture, and the AI architecture is also technology agnostic The new DeepSeek model, the powerful one, Pro Four, it runs as well on Nvidia as it does on the Huawei Ascend architecture.

I think the hardware is agnostic. So that combination, Chinese hardware, which is going international right now and getting quite good, and outside of Huawei, they can all use Nvidia chips if they want, and they can use those and Ascend, and they seem to be sort of approaching this with a heterogeneous approach in terms of compute.

That combination of cheap and very good hardware plus open source downloadable models that are frontier level, I [00:17:00] think as of two days ago, that might be the winning strategy now. That's my new default assumption for the future. And I actually think that's good. The truth is, I actually like China's approach to tech.

I like info nationalism. I like that each country determines what is censored, what is not, what is prioritized, what is not, what type of contact, con- uh, conduct is allowed, what values are promoted. I don't like a bunch of far left Silicon Valley people telling journalists in Indonesia what they can write and not write about their own country if they don't want to get sort of shadow banned.

I like the idea that all those controls are local, which is kind of what China did all along by saying YouTube has-- can't be here. And that's pretty much what the US did with their nationalization expropriation of TikTok. They basically said, [00:18:00] "We don't want the algorithm that shapes American information flows being hosted in Beijing.

We want it to be local. You have to sell US TikTok." I actually think that's the right approach. And when we move to generative AI, these frontier models, this is taking it to the next level times ten. If you control a search engine or a news feed, whether it's on video or whatever, you have ability to shape opinion.

If you control generative AI, you control reality for the most part. You can say what happened in history and what didn't. You can say what news stories happened and what didn't. You-- I mean, it is si- it is arguably the single most powerful tool that's ever existed. You can control reality for people f- in all, for all effective purposes.

So I like all of those things to be hosted locally. I, I don't like the idea of Anthropic or OpenAI [00:19:00] shaping reality for people in Indonesia. I think there should be open source models with, you know, sort of technology agnostic architecture and infrastructure that is high quality and low cost. I like that solution that you can deploy country by country, and then people decide for themselves at the cultural and societal and political level how they want to behave in this sort of really stunning level of power that's being deployed.

Anyways, that's where I am on it politically. Anyways, that's a bit of a rant. Um, but yeah, I think it's, I think it's a major event. I think this is a sea change in what's going to happen. And, uh, yeah, we'll see. But yeah, I'm-- that's my baseline prediction now. We'll see. I also think it has to do a lot with the fact that the US has, does not have the credibility it used to have, let's say, 10 or 15 years ago in a lot of countries.

Some it does, but there's a lot of countries where that is really not the case anymore. [00:20:00] That's a political opinion. Okay, that's it for this subject. I'm going to switch over to CATL. But yeah, I've been thinking about this all day. Kind of blew my mind All right, let me start talking about CATL, and I'll do this in two podcasts cause I've already been going on for, like, 20 minutes.

Yeah, I flew out to, um, to Fujian or Fuzhou. Uh, the province is Fujian, right across the, you know, strait from Taiwan, and the capital's Fuzhou, but really everyone goes to Xiamen. Right? Xiamen is, uh, it's not the capital of the province, but it's the larger city. It's, it's beautiful. If you've never been to Xiamen, I really like Xiamen.

I find it really pleasant place to go. It's all mountainous. It's on the water. A lot of good food. It's, they're good with history. Anyways, I used to fly there when I was living in Shanghai. I would go down for the weekend, uh, Fujian. It was great fun. Anyway, so I went to a l- pretty little city actually, which is, um, north of Fuzhou, which is called Ningde.

And [00:21:00] Ningde is, uh, not a big city, but it happens to be where CATL was founded, and that's always sort of remained its headquarters even though that it's probab- I mean, if I had to guess, I think they're moving to Xiamen because, you know, they got 100,000-plus employees now. You, you're trying to get the top-tier engineers of the world.

It's kind of hard to get them to live in Ningde. But in contrast, people are quite happy to move to Xiamen. Like, you can actually get top-tier talent to move to Xiamen. It's right up there with Shanghai, Hangzhou in terms of really nice place to live. So I think it's actually well-positioned to be a tech city or at least a engineering city in this case So yeah, maybe I, I think they're going to probably end up moving their, their headquarters there at some point, but I don't know.

So anyways, I, I sort of, uh, flew into Fuzhou and I took the train up to Ningde and, um, you know, I met with them and sort of went through the, the basics of [00:22:00] CATL. I looked a lot at, you know, what they've been doing and I'll give you the basics of the company. But really the key question-- I've written four articles about this that are on the webpage now about CATL.

I sent out two of them, uh, by email, but there's another two on there which I haven't sent out, and I'm going to do one more. So there's going to end up being five articles about me sort of taking apart CATL on my, on my webpage. You can go look there. But the question I'm getting to, which is, I won't get to this podcast, but in the next one, how did they win?

They have won so big and, you know, it looks like it'll continue, but who knows? Um, you got to kind of wonder what happened here, cause it, this is a stunningly successful company. So whenever I see a stunningly successful company, I try to take it apart. That's kind of my strategy business, right? So I'm going to sort of give you that in article number five, which I [00:23:00] haven't finished yet, but I want to sort of tee up the basics in this one.

So to me, there's sort of three things to think about. Number one, okay, what are the basics of CATL? I'll go through that. Two, what have they been doing on the innovation side? Cause that's a big engine of what's going on, and innovation in batteries, which is chemistry, is very different than engineering in, let's say, uh, innovation in, like, rockets, which is engineering.

And that's very different than innovation in software, which basically you can innovate in software very quickly. Engineering is harder and slower. Science is even more difficult. So engineering's a big part of the story, but you got to break it down. And then the third question is, look, how did they win?

And I'll, I'll get to that in the next podcast. So those are kind of my-- what I want to go through in this, um, in about 10 minutes, and then I'll finish up. Okay, so here's-- Let me start with the sort of basic background. Uh, this is all from their 2025 annual report. They're publicly traded. [00:24:00] Uh, nothing in this is non-public information.

This is all coming from annual reports and public stuff, and this is not investment advice in any way, shape, or form. Okay. The world's largest manufacturer of basically rechargeable batteries for electric vehicles and for energy storage systems. Now, those are two different markets. Uh- Electric, you know, power batteries go into electric vehicles.

They have to go in the floor of the vehicle. They have to sort of deliver power very quickly, and they have to be compact. So energy density is huge. Energy storage systems, these are big 20-foot containers you park next to the power grid or the solar plant, and they have a different set of dynamics. You basically want them to be low cost.

But energy density, how compact you can make it, it's not that big a deal, and you don't need to discharge really quickly, you know, to hit the gas [00:25:00] pedal. So different types of batteries, but they're number one in both in China and globally. Now, the core business has been power batteries. That's what you put in EVs.

But really, you can put that in scooters, you can put that in boats, anything small that moves, and there's a lot of stuff that's starting to get batteries in it. But no, I mean, 70 to 75% of their revenue has been from the power battery business, um Energy storage, that's kind of their big growth opportunity.

They've been doing power batteries since about 2011. They've been doing energy storage for about five years now. In power batteries, they're number one in the world. Globally, 40%. In China, 50%. So every EV you see in China, every two EVs, one of them has a CATL battery in there. They supply them for Tesla, they supply them [00:26:00] for NIO, they supply them for BMW.

They're the major supplier for most all the auto companies. When-- And they've been number one for nine years. Huge. Energy storage is newer. They've been number one for about five years, and the percentage is much lower, 30-ish percent, something like that. Now, when you think about energy storage, which I'm not going to talk about too much here, you know, the big customer there was power grids.

You know, these big electrical grids that are increasingly doing things like putting solar panels. Well, then you got to store that somewhere. So they put batteries that attach to the grid. Maybe they do peak shaving, you know, things like that, but okay. AI data centers are an interesting sector within that, that business because everyone's building the AI data centers, they all need batteries.

So energy storage is interesting. But I'm not going to talk much about that in this one. Power batteries is kind of the, [00:27:00] their core engine. All right. Let me give you some more numbers. Global EV batteries CATL had 39% of the market share in 2025. Uh, their batteries business, their EV battery business within CATL grew 42% in 2025.

Now, that got my attention. Wait, you're already 40% of the global market, which makes you huge, and you're growing at 42% as well? I mean, when do you see massive market share in a big business and that type of growth? Usually, growth slows when you get that big. 42%, really? When you look at the Chinese EV battery market, they're up 46, 48% 2025.

You know, they're number one. The... Let's say number two would be BYD. They make batteries that go in cars. They're at 25%. [00:28:00] They're a good 15 percentage points down in terms of market share. Now, when you look at the storage business, their market share, 30% 2025, their sales within that also grew at 29 to 30%. So you see the same pic- This is really what got my attention.

I'm like, "That's really unusual." Okay, but maybe, maybe the margins aren't so good. No, it turns out the margins are pretty good. Um, their revenue for 2025, $61 billion That was up 17% from 2024. So yeah, again, big and growing fast. Okay, what's your operating profit? $20 billion. So of 61 billion in revenue, you got $20 billion of operating profit.

So I mean, that doesn't look-- that doesn't sound like a hardware company. Hardware companies [00:29:00] don't make those types of operating profits. They don't show that type of global dominance usually. Sometimes, but not usually, not 40%. And if they do show 40%, they don't show growth on those levels. So the, the numbers really kind of jumped out at me that, you know, this is a company that's dominating China and dominating globally.

They're throwing off cash and they're high growth. I mean, that's... I mean, that was my first take, and I kind of wanted to know how they were doing this. Uh, some other stats, you know, as I mentioned, they do batteries for most of the major EV brands, Tesla, Volkswagen, BMW, NIO, Geely, SAIC. Um, globally, about one in three vehicles, EV vehicles, have their batteries.

In China, it's about half. They have about 130,000 employees globally. Most of that is in China still. They have 13 sites where they make their batteries. They have six R&D sites. Again, [00:30:00] most of those are in China, but they're in kind of the middle of an, a global expansion right now, so they're opening in Germany, Hungary, they've opened.

But their footprint is still kind of mostly China. All right. Let me give you some sort of interesting factoids from their history. Founded-- Actually, everyone says they're founded in twenty eleven, CATL. Not actually true. Well, kind of true. They were founded in nineteen ninety-nine, but they were called ATL.

Uh, CATL stands for Contemporary Amperex, uh, Technology Limited. So whatever. The company was founded as ATL, Amperex Technology Limited. They added the C, CATL, when they spun off or at least created a separate subsidiary in twenty eleven for their, basically their EV battery business. Prior to that, they were doing stuff in batteries for [00:31:00] consumer electronics and other places.

So actually kind of a lot like BYD started out, consumer stuff, um, but they kind of moved up the, the value chain into larger and more complicated batteries, and then they jumped on the EV tidal wave and rode that pretty well. Uh, as mentioned, started in Ningde. Interesting. That's where they built their first factories.

Okay, they start making EVs for passenger vehicles twenty eleven, twenty twelve. That's really the start. The first big event happens in twenty twelve. They get a strategic partnership with BMW, where BMW basically asks them to supply the battery systems for the Zinoro electric vehicle brand and its Chinese joint venture.

So yeah, they basically did a w- deal with BMW for China, at least at that point. Now, that was obviously a big, you know, big acceleration in their [00:32:00] fledgling new business. There's probably a lot of benefit there, where suddenly they had to make batteries that conform to BMW standards and German engineering standards, safety requirements, qual-quality standards, quite high.

So that kind of moved them up to the highest level of quality, uh, very early on, either by choice or because they had to. That's pretty interesting. Uh, by twenty fifteen, four years in, their manufacturing output of batteries, eh, two gigawatts, two point two gigawatts. That puts them in the top in terms of overall production of batteries, uh, in the top tier.

Not the top one, but in the top tier. They had about six thousand employees at that point, so pretty rapid growth as soon as the BMW deal kicked in. And you know, when I look at these numbers, what I see is a company that was in the right place at the right time [00:33:00] and moved really fast. Now I wrote, I rewrote my books, Moats and Marathons.

I just put part number one up for free on my website if you sign up for the email. And one of the things I rewrote as I kind of redid these books, I really re- I really underestimated how important it was just being in the right place at the right time and being really, really fast. I like the strategy stuff.

I like structure. I like to think about the business models. Those are all important. But I think I really underestimated... Look, be early to a big wave and be really fast. That is a huge part of success. And I think it was, um, I don't know, I think it was the All In podcast. One of them said that, like, investing in tech is basically big wave riding.

You sit out there on your surfboard, and you wait for the really, really big wave, and you have to be in the right place at the right time to catch [00:34:00] it That's being early. But at the same time, when the wave hits, you also have to paddle like crazy to get yourself on top of that wave. To me, that's 50% of what happened with this company between 2011, 2015.

They caught the wave, they paddled like crazy. By 2015, they were riding the crest of this wave, which is, you know, pretty awesome. All right, some of the other announcements within their core business over the next 10 years. Uh, 2017, they surpassed Panasonic to become the world's largest supplier of power batteries for cars by sales volume.

That was about 12 gigawatts total. Uh, they opened all these facilities all over China. Their headcount grew to 14, but they were basically number one by 2017. Uh, they IPO'd in Shenzhen 2018. 2018, they also started going global, [00:35:00] which in practice means you start signing supply contracts with automotive manufacturers outside of China.

For them, that was Volkswagen, Daimler, Jaguar, Land Rover, and, you know, suddenly you're making the batteries and supplying them with their factories in Germany, Hungary, places like that. Uh, what else? Uh, 2020, they got another big boost. They signed a high-volume supply agreement with Tesla to provide lithium ion phosphate batteries for their Giga Shanghai plant.

That, I mean, that really took them up to the next level, got them a lot more scale. Pretty impressive. 2022, they began overseas manufacturing in Germany. Uh, what else? Eh, that's pretty much it. 2024, they had 13 major production bases all over China. Um, they start building mega facilities in [00:36:00] Europe. And by 2026, they're a global manufacturing company, 120,000 employees.

Their engineering workforce is about 20,000 of them. Uh, that's researchers, software people, electrochemical specialists. And that sort of scale in their core business is what enabled them to jump into energy storage as a secondary business. Let me mention a couple other things before I move on. That's their core business, but they've also started to expand as energy storage is their big expansion.

Uh, they've done some other things that are interesting. They, they got into battery recycling Uh, they bought the world's largest recycler of batteries, a company's called Brunpht, which I think is German. Um, basically that gave them a 99% recovery rate for, uh, nickel, cobalt, manganese, and then 96% for lithium, which are the core components for [00:37:00] making most of their batteries.

Uh, as mentioned, 2018 they started moving into energy storage as a new business. Uh, 2019 they started doing a joint venture with State Grid of China, which is a huge massive SOE that does power grids. That's to build energy storage, um, for the main grids of China. Uh, some innovations that were interesting, 2021 they launched their sodium ion battery.

I'll talk about that, I think I'll probably do that in the next podcast. They've had four or five major innovations. I want to get into why innovation matters and why it doesn't, and they've had four or five major innovations that are worth paying attention to. Uh, some of them are in chemistry of batteries.

That's this case, that we're going to make batteries out of sodium ions instead of, uh, lithium. Uh, but there's some others that they've done, uh, super fast charging, uh, energy swapping, where you swap [00:38:00] out the batteries in cars, things like that. Okay. That's kind of a... I think that's a decent summary of CATL.

Let me finish up this introduction by talking about how they talk about their innovation. Now, if you read their annual reports, you'll see that they sort of break their innovation down into four buckets. Uh, the first is basically materials and electrochemistry. Okay. I mean, this is how do you make batteries.

I've s- I've actually spent a lot of time in the last couple weeks reading about batteries. If you, if you want to get into this, read the four articles I've put up. I mean, I get into kind of a lot of the details of how various lithium versus sodium batteries work, why thermal runaway causes batteries to catch on fire, how that happens in practice, why you need these types of anodes and cathodes, and why you need blah, blah, blah.

I mean, if you're interested in that stuff, I've detailed it out pretty good. I spent a lot of time sort of [00:39:00] reading about it, trying to get smarter. But yeah, the first bucket for their R&D and their innovation work is it's all in this area, materials, chemistry, electrochemical systems. And what are they trying to do?

They're trying, number one- Increase the energy density. How can we pack more energy into the same physical space? Because, yeah, these things are heavy. We can't put massive batteries in cars. We got to get them lighter, but we got to get more power in there. Energy density, at least for power batteries, is huge.

How do we get the batteries to have a longer lifespan? There's usually a limit to how many times you can charge it and recharge it. Three thousand is a common number for power batteries. Um, you can get up to eight or nine thousand for things like energy storage, but that's a big number. That's kind of the lifespan.

How fast can you charge the battery? Now, [00:40:00] they've got some pretty cool breakthroughs in fast charging. Uh, their latest one, if it's in a car, you can charge that thing in about ten minutes. That's pretty amazing. Um, safety is a big, big deal. Um, as you know if you've flown on an airplane, batteries are a big problem in terms of catching fire.

If they get damaged, you can get thermal runaway. One cell, because basically what happens, which I won't go into, if a cell get-- You know, these batteries are made up of lots and lots of cells packed tightly together. If one cell gets damaged, it can trigger a series of accelerating exothermic reactions, where one reaction happens, like the, the barrier between the anode and the rest of the gel breaks down either because of damage or heat.

That gets you an exothermic reaction, which means [00:41:00] basically a lot of heat and gas gets expelled. That can trigger the electrolyte, which connects the anode to the cathode. That can then break down. That creates a secondary exothermic reaction. When you get that, which means then a lot more heat, a lot more gas.

That can often cause a short circuit where the anode connects directly to the cathode, which it's not supposed to do. You get a short circuit, more energy. You get, basically get a series of cascading chemical reactions that cause these batteries to go up in flames, and you can't put them out. I mean, it, it's a chemical reaction, and if it happens in one cell, it triggers the next cell and the next cell and the next cell.

That's why you're not allowed to, you know, charge your batteries on phone, you know, sorry, charge your phones with your battery on airplanes anymore, cause they're worried about these batteries going up. Um, anyways, I detailed all the [00:42:00] chemistry in my own simplistic way in the notes and in the article if you're curious.

But yeah, reliability, lifetime, and safety is a huge issue. Okay, so that's bucket number one for their R&D. Bucket number two for their R&D, they call it system structure innovation. I would call it engineering-based innovation. The first bucket is about science. Can you create a new type of battery in this world based on the physical world, chemistry?

Well, that's super hard. The nature is super difficult to innovate in. But if you move to engineering, which is just building machines, which is what Elon Musk does, you can move much faster. So when they talk about system structure innovation, they're talking about how we put the battery cells together in a structure, how we put them into modules, how the pack architecture puts [00:43:00] together.

You can do a lot of cool engineering there. That moves much quicker. So the goal is there, how do we optimize performance? How do we do thermal management? How do we increase mechanical strength? How can we create greater efficiency in the whole battery pack? That bucket is actually pretty cool, and I'll talk about some innovations there.

Bucket number three, they call this extreme manufacturing or intelligent manufacturing. Basically, this is production engineering. How can we produce these batters with lower-- batteries with lower cost? That's actually a huge deal. Um, how can we do it with higher precision, so we get fewer defects? How can we get reduced energy consumption when we make the batteries?

How can we increase the product safety? How can we increase the product reliability? All of that would go under [00:44:00] manufacturing Processes and engineering and efficiency. That's a huge bucket. Uh, number four, which I'm not going to talk about, business model innovation. They've been sort of experimenting with moving into service-based businesses, like they have several thousand energy swapping stations around China.

I went and visited one of them, where you basically pull up in your car, you go through a little tunnel, like a car wash. You pull your car into the tunnel, the bottom under your car opens up, and they basically replace your battery while you're sitting in the car, which I did. And then it, you know, takes four minutes, something like that, and then you leave the...

It l- really looks like a car wash. You leave, you have a new battery in your car that's fully charged. So they're getting into sort of battering swapping stations. Um, [00:45:00] for passenger-owned cars, this is actually kind of difficult cause there's lots of cars and people live all over the place. It makes more sense for commercial vehicles like trucks because they all take the same highways and you can put these stations along the highway.

You can pull up, it'll automatically swap out your battery and swap in a charged one. The truck moves right out. So they're, they're sort of experimenting with these new, um, battery as a service business models. Um, it's actually kind of interesting cause it basically means... Here's why it's actually kind of cool.

It basically means you buy the car, but you don't buy the battery. The battery is a service. You buy the car. Now, if you're not paying for the battery, the cost of the car goes down quite a lot cause it turns out the battery's pretty expensive. So, if you're a consumer looking to buy a passenger vehicle [00:46:00] and they say, "Look, you can buy the car and own the battery and then just charge it, or you can buy the car and just get battery as a service," and the car drops by like 30% in price I don't know.

I made that number 30% up, but it's actually kind of an interesting model when you think about it. Okay, so that's kind of their four buckets, how they talk about it. If you read their annual report, you'll see them talk about these four buckets. What I'm going to do in the next part, and I'm going to finish up here, is I'm going to sort of break down how I think their innovation plays out and why it matters as a business model and then tie that to why they've won.

I think innovation is a part of it, but not how they're describing it really. I'll give you my take on that, and that'll be in part two. Okay, so we'll call that an introduction. I think there are some good lessons here, and I, I mean, I'll tell you why I'm doing this. [00:47:00] You know, I've spent a long time looking at businesses that are basically business plus software.

That's where I've lived and then I kind of expanded into their into AI, because AI is... It's a type of software, but it's different. I've been moving more and more into hardware cause I think that's a big part of the future. I think, you know, the software business models we are so comfortable with, network effects, YouTube, platform business models, I think most interest-- No, let's say I think a lot of the interesting businesses are going to be digital physical hybrids.

It's going to be a combination of hardware and software or software and services. And I'm-- I've kind of recognized that I'm weak on the hardware side, so I've been digging a lot into manufacturing hardware. And I started with data centers, which I think I'm pretty good at now, and now I'm moving into batteries.

And the [00:48:00] next I'm going to do... I'm not going to do EVs. I don't really want to dig into cars, but I'm going to look into manufactured products like robots. Uh, I'm going to look into drones more. I'm going to look into smart refrigerators. I think that's the future. I think it's digital physical hybrids, and especially because it doesn't look like generative AI has any major competitive advantages.

It doesn't look like it has network effects. There are some advantages in proprietary data, but it looks like the business models that are going to have real competitive power are not going to be pure AI companies. They're going to be AI companies that also have more traditional competitive advantages that probably come from hardware and physical assets.

Now, I've said this for years. I've said for years, I like e-commerce. I like JD because it is a combination of software, [00:49:00] e-commerce, but physical assets as well, logistics, trucks. I like digital physical hybrids as highly defended, uh, business models. I think a lot of this AI stuff, to be defended and to make longer term profits, you're going to be looking at digital physical hybrids.

So that's why I'm digging more and more into hardware and manufacturing. Uh, I'm going to do a lot more of this actually. I've been reading a lot about factories and the cost structures. But yeah, I started with AI data centers, now batteries, cause batteries are pretty straightforward, and next will be big factories and drones and robots, which are coming.

Anyways, that's what I'm doing. Okay, that is it for me for today. It's about 11:00 PM here in, um, on, uh, Sunday night. I'm flying out to Brazil in a couple days. I'm going to be there for a couple weeks, São Paulo, Rio. Um, [00:50:00] some pretty awesome meetings. I'm pretty excited about it actually. I'm going to try and sneak out to Rio for a couple days cause that's, like, my favorite thing.

Uh, but f- actually some kind of big... I don't want to say big. Uh, yeah, for me they're pretty important. Like, this is kind of a big deal, so I'm, I'm really excited about that. And I'll talk about it when I get there. But yeah, you'll know the companies I'm meeting with. You, you know their names already. Anyways, that's it.

I'll, I'll give you more details about that shortly. Anyways, that is it for me. I hope everyone is doing well, and I'll talk to you next week from São Paulo. Bye-bye.