The Memory Selloff: Record Profits, Collapsing Stocks

Today, we need to discuss the sell-off in AI and memory stocks, including sharp declines in Korean equities and SK Hynix despite insane earnings. 

We also cover the fragility of leverage-driven markets, China’s memory industry, and the outlook for continued AI-related demand for chips and infrastructure.

------
🔒 Check Out Our Sponsor: LEDGER AGENT STACK 🔒
https://developers.ledger.com/docs/ai-tools/overview/?utm_source=Audio&utm_medium=Podcasts&utm_campaign=Limitless

------
🌌 LIMITLESS HQ ⬇️

EMAIL US:            [e-mail]
NEWSLETTER:    https://limitlessft.substack.com/
FOLLOW ON X:   https://x.com/LimitlessFT
SPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQ
APPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890
RSS FEED:           https://limitlessft.substack.com/
------
TIMESTAMPS

0:00 Memory Stocks Crash
3:52 Why SK Hynix Is Dumping
7:02 Leverage Fuels the Selloff
8:30 China’s Memory Surge
12:00 ASML Rival Panic
15:47 AI Money Shifts Upstream
17:23 Open Source Misconceptions
18:56 GPUs Need More Memory
21:04 Agentic AI Drives Demand
23:08 What Investors Should Do

------
RESOURCES

Josh: https://x.com/JoshKale

Ejaaz: https://x.com/cryptopunk7213

------
Not financial or tax advice. See our investment disclosures here:
https://www.bankless.com/disclosures⁠

Josh works with Anthropic as a contractor. All views expressed are his own and do not represent Anthropic, its leadership, or its affiliates. Nothing in this episode is investment advice.

Episode Transcript

Josh: The U.S. stock market just lost more money than we did in the 2008 financial (00:00):
undefined

Josh: crisis over a matter of seemingly a couple of days. It's been the worst memory (00:04):
undefined

Josh: crisis, the worst stock market crash in quite some time because the numbers have gotten so high. (00:09):
undefined

Josh: Memory has run up to astronomical levels. But what's interesting here is this (00:14):
undefined

Josh: is backed up against some counterfactual evidence. (00:18):
undefined

Josh: SK Hynix, that Korean company that could seemingly do absolutely no wrong, (00:20):
undefined

Josh: they recorded more profit than they did revenue this last quarter. (00:24):
undefined

Josh: So the numbers are great, but the stock market is saying, wait a second, (00:28):
undefined

Josh: something's not right here. (00:32):
undefined

Josh: So we have some insight. In this episode, we're going to unpack this, (00:33):
undefined

Josh: why the market is reacting the way it is, and evaluating whether or not it's (00:35):
undefined

Josh: right or if it's a pretty serious overreaction. (00:39):
undefined

Josh: I'm currently sitting in the middle of a hotel, in the middle of nowhere, (00:42):
undefined

Josh: on some spotty Wi-Fi, trying to get this episode out to you guys. (00:45):
undefined

Josh: Because, I mean, this is an incredibly important topic. This feels very timely (00:48):
undefined

Josh: and important to navigate. So, Ejaz, let's unbundle what's happening here in (00:51):
undefined

Josh: the market. I mean, worse than 2008 is crazy. (00:55):
undefined

Ejaaz: Yes. So I'm seeing a lot of commentary in the markets right now, (00:58):
undefined

Ejaaz: which is the AI bubble is finally popping. It's unwinding. It's been happening for four weeks. (01:00):
undefined

Ejaaz: The Korean stock index itself is down over 30%, which is just the largest drawdown they've ever had. (01:06):
undefined

Ejaaz: And so I want to explain what the damage is. And then let's get into whether (01:12):
undefined

Ejaaz: this is right or wrong, because I honestly have some pretty strong opinions (01:16):
undefined

Ejaaz: as to why this is completely wrong and why it's like one of the bigger opportunities right now. (01:19):
undefined

Ejaaz: So what is the damage? What you're seeing on the screen right now is the Korean (01:22):
undefined

Ejaaz: stock market. And the reason why I'm showing you the Korean stock market versus (01:27):
undefined

Ejaaz: the American stock market is because they're influencing each other quite a bit. (01:29):
undefined

Ejaaz: Now, Korea has three of the biggest, or rather two of the biggest memory manufacturers, (01:33):
undefined

Ejaaz: SK Hynix, which is the number one, and Samsung. (01:38):
undefined

Ejaaz: And they provide pretty much all the memory that is required for the GPU makers, (01:41):
undefined

Ejaaz: the CPU makers, or whatever type of AI infrastructure you can dream of. (01:45):
undefined

Ejaaz: So it's a very essential component. (01:48):
undefined

Ejaaz: Now, the issue is, it's down 32%, and that is because there's a lot of memory (01:50):
undefined

Ejaaz: providers that have been dumping completely. So if I pull up SK Hynix right (01:55):
undefined

Ejaaz: over here, over the last month, it's down almost 25%. (01:59):
undefined

Ejaaz: If I pull up SanDisk, right? SanDisk Corporation creates a different type of memory for AI. (02:04):
undefined

Ejaaz: They're down 50%. This is the darling stock, by the way, which was up like- (02:08):
undefined

Ejaaz: This is trending like a meme coin, man. Yeah, I know. (02:13):
undefined

Ejaaz: But it also pumped like a meme coin. (02:15):
undefined

Ejaaz: I think it was up like something ridiculous, like 4,000%. (02:18):
undefined

Josh: And even though- Well, can we look at the year chart? (02:21):
undefined

Ejaaz: Yes, let's look at the year chart. (02:23):
undefined

Josh: Just to see what that- Yeah. (02:24):
undefined

Ejaaz: The year chart is still up 2,200%. So even though it's down 50%, it's still up 2,200%. (02:26):
undefined

Josh: Oh, this is so unbelievable. (02:33):
undefined

Ejaaz: And so the question on everyone's mind is, why are these stocks, (02:34):
undefined

Ejaaz: specifically AI stocks, specifically memory stocks, dumping (02:40):
undefined

Ejaaz: on the back of some amazing news, which is these companies are still pulling (02:44):
undefined

Ejaaz: in more revenue, more profit, as you mentioned earlier, than they've ever done in a single quarter. (02:48):
undefined

Ejaaz: SK Hynix released their quarterly earnings literally yesterday. (02:53):
undefined

Ejaaz: I was reading it last night before in preparation for this episode. (02:57):
undefined

Ejaaz: And Josh, to your point, Their revenue increased by 354%, but their profit margins increased by 555%. (03:00):
undefined

Ejaaz: They made more money in this quarter than they did in the entirety of 2025, (03:07):
undefined

Ejaaz: last year. So the question on everyone's mind is, (03:12):
undefined

Ejaaz: Do these stock price movements make sense at all? (03:15):
undefined

Josh: The thing that like, well, first of all, the answer is no, because I read through (03:19):
undefined

Josh: this earnings report too, and I was immediately confused because how on earth (03:22):
undefined

Josh: do you have more net profit than revenue? (03:26):
undefined

Josh: It's 118% net margins. It's like unbelievable profit margins. (03:28):
undefined

Josh: It's an incredible business. It seems like it's doing remarkably well. (03:32):
undefined

Josh: And yet the market seems to just be kind of done with this. (03:36):
undefined

Josh: It's like the Toy Story meme where they just kind of throw the toy out like, (03:39):
undefined

Josh: I'm done with this toy. But the numbers don't make any sense at all. (03:42):
undefined

Ejaaz: Can I tell you why? (03:46):
undefined

Josh: Yeah, please. Like, I'm trying to understand. I'm reading through these notes (03:47):
undefined

Josh: here. I'm like, okay, what's wrong here? (03:51):
undefined

Ejaaz: Okay, so there's two reasons why I think SK Hynix or memory stocks, (03:53):
undefined

Ejaaz: AI stocks in general, are dumping. (03:57):
undefined

Ejaaz: Number one, analysts at these different firms on Wall Street or wherever set (03:58):
undefined

Ejaaz: targets, right? So expected revenue growth for a lot of these companies. (04:04):
undefined

Ejaaz: Now, SK Hynix technically missed their revenue target by around $1.7 billion. (04:08):
undefined

Ejaaz: You're seeing this on the screen right now. (04:14):
undefined

Ejaaz: What I would like to draw your attention to, Josh, is this. You see that? (04:16):
undefined

Ejaaz: That is 250% of revenue growth from the previous year. (04:20):
undefined

Ejaaz: And that is almost 600% in operating profit from the previous year. (04:25):
undefined

Ejaaz: So whilst they may have missed some random analysts' projections, (04:30):
undefined

Ejaaz: They've still excelled and compounded at a much more rapid rate than any other (04:35):
undefined

Ejaaz: company in the world. It is extremely impressive. These things are printing money. (04:39):
undefined

Ejaaz: But the question should then be, why were the analysts even predicting this (04:43):
undefined

Ejaaz: target in the first place, right? (04:47):
undefined

Ejaaz: And I'll give you an answer for this. SK Hynix specializes in this thing called (04:49):
undefined

Ejaaz: HBM, high bandwidth memory. We've spoken about this a lot on the show, right, Josh? (04:54):
undefined

Ejaaz: Now, they are the number one provider of HBM. They only dedicate their chip (04:58):
undefined

Ejaaz: fab capacity to create HBM. (05:03):
undefined

Ejaaz: Now, if you look at the other competitors, Samsung, Micron, they do HBM, (05:06):
undefined

Ejaaz: obviously, but they also do some other cheaper memory for like your mobile phone (05:10):
undefined

Ejaaz: or for your computer. It's called DRAM, and you're basically able to use that (05:14):
undefined

Ejaaz: for other different types of gadgets. (05:18):
undefined

Ejaaz: Now, because SK Hynix is so focused on HBM, they have run out of supply, (05:20):
undefined

Ejaaz: so they can't possibly sell anymore. (05:26):
undefined

Ejaaz: So they're selling off on the best news ever, which is they've sold out all (05:28):
undefined

Ejaaz: their entire supply for this quarter. (05:32):
undefined

Ejaaz: And so they have to move on to the next quarter's supply in order to get higher profit margins. (05:34):
undefined

Ejaaz: Now, if you look at Samsung, if you look at Micron, they have dumped, (05:39):
undefined

Ejaaz: but they have dumped less because they're selling more DRAM. (05:42):
undefined

Ejaaz: Have they made as much money as SK Hanex? No, because HBM is priced higher, uses more wafers. (05:45):
undefined

Ejaaz: So the whole thing basically is ridiculous. (05:51):
undefined

Ejaaz: SK Hynix has created a really good product. They have sold all of the product (05:54):
undefined

Ejaaz: that they could potentially make in that single quarter. (05:57):
undefined

Ejaaz: And because they've done that, they have now not been able to reach a specific (06:00):
undefined

Ejaaz: target that this random Wall Street analyst has set for them because they've (06:04):
undefined

Ejaaz: dedicated all their chip capacity to the specific thing. (06:08):
undefined

Ejaaz: So let me know if that makes sense. But basically, it's crazy. (06:10):
undefined

Josh: If I had to summarize it probably in three points, it's like, (06:13):
undefined

Josh: okay, the first one is the stock market rating, like a lot of the analyst ratings, (06:15):
undefined

Josh: where basically you could, (06:20):
undefined

Josh: you could think of it like if you if you give your kid a hundred dollars for (06:22):
undefined

Josh: straight a's and they bring home like (06:25):
undefined

Josh: four a's and then an a minus and everyone like freaks out because they're like (06:27):
undefined

Josh: oh my god no this isn't what you promised yes that's the first thing the second thing is the (06:31):
undefined

Josh: record earning surprises like basically like you mentioned they've fully sold (06:35):
undefined

Josh: out of their inventory for 2026 yes there is no more capability for them to sell more by the way, (06:40):
undefined

Josh: And 27, there's no ability for them to sell more or sell it at a higher margin (06:45):
undefined

Josh: because it's already pre-sold. So therefore, you eliminate a lot of the upside surprises. (06:49):
undefined

Josh: And there really is only downside surprises possible in the case one of these (06:54):
undefined

Josh: deals don't work out how they expect or things fall through. (06:57):
undefined

Josh: So the upside is kind of capped in terms of surprises. Downside is not. (07:00):
undefined

Josh: And the third is there is this two times leverage ETF that started trading just (07:02):
undefined

Josh: a couple weeks ago, July 13th. So I mean, of course, being a Korean market, (07:07):
undefined

Josh: a lot of people, it's funny, like there's this thing. (07:11):
undefined

Josh: It's kind of known with the Korean stock market, where they are the most aggressive (07:15):
undefined

Josh: gamblers per se. They like to take on the most risk. (07:19):
undefined

Josh: And this two-time leverage fund, I'm sure, fed right into that. (07:21):
undefined

Josh: So there was a lot of leverage baked into the price of these stocks. (07:24):
undefined

Josh: And any sort of sell-off event creates this cascading liquidation event. (07:27):
undefined

Josh: And I'm sure we saw a lot of that as well with a 2x leverage stock. (07:32):
undefined

Josh: So the convergence of those three things... (07:36):
undefined

Ejaaz: Actually, on that, Josh, there's some news from Chucan I saw this morning that (07:37):
undefined

Ejaaz: JP Morgan... So you mentioned the leveraged ETFs and you're right. (07:41):
undefined

Ejaaz: Like this has led to a lot of the dump because there's people like couldn't (07:44):
undefined

Ejaaz: afford the stocks that they were buying. They were too leveraged up. (07:47):
undefined

Ejaaz: And JP Morgan this morning reported that the leveraged ETF drawdown, (07:49):
undefined

Ejaaz: the liquidation specifically is about 90% complete. (07:53):
undefined

Ejaaz: So if you wanted to kind of like extrapolate, you'd probably see this bottoming sometime soon. (07:57):
undefined

Ejaaz: So like this drawdown can't go on forever and we're probably nearish a point (08:02):
undefined

Ejaaz: where it's gonna reach its bottom before like everything starts to like settle (08:07):
undefined

Ejaaz: and maybe kind of like recoup. (08:11):
undefined

Josh: Yeah, well, it seems like, I mean, we had this earlier in the year where there (08:13):
undefined

Josh: was that big sell-off. I remember Bill Ackman famously saying like, (08:15):
undefined

Josh: hey, this is oversold. The market is wrong. (08:18):
undefined

Josh: You're overreacting. We are probably getting close to something like that now. (08:21):
undefined

Josh: Again, not financial advice, who the hell knows? (08:25):
undefined

Josh: But there are some signs that things are shifting. (08:27):
undefined

Josh: And I want to shift our attention to China now to talk about what's shifting (08:30):
undefined

Josh: over there because there is some, you could say that China played a fairly large (08:35):
undefined

Josh: role in this and will continue to play a fairly large role going forward, (08:38):
undefined

Josh: what you'll notice is that we're not really talking much about (08:42):
undefined

Josh: the united states stocks like this is very much an international this is a global (08:45):
undefined

Josh: marketplace now because everyone is so interdependent on these supply chains (08:49):
undefined

Josh: and china has a very big one with memory and there's a company i'm going to (08:53):
undefined

Josh: try to pronounce this right shangjin (08:57):
undefined

Josh: memory technologies yes cxmt is a ticker basically. (08:59):
undefined

Josh: And they had themselves a public ipo a public debut in which they traded up, (09:02):
undefined

Josh: 466 percent in one day which is instantly the most valuable china listed company (09:07):
undefined

Josh: ever, which is more than Alibaba or Tencent. (09:14):
undefined

Josh: And they raised about $9 billion. (09:16):
undefined

Josh: So you're thinking, who on earth is this company? I've never heard of CXMT. (09:18):
undefined

Josh: Well, they're the world's number four DRAM maker. (09:23):
undefined

Josh: Now you'll notice we normally talk about the top three DRAM makers. (09:27):
undefined

Josh: A new entrant has entered the category. (09:30):
undefined

Josh: And I have to ask, Ejaz, is this like a little concerning because there's more (09:32):
undefined

Josh: distribution of people who are able to make this memory? (09:36):
undefined

Josh: I mean, over the last four years, I believe. They've gone from a 1% market share (09:39):
undefined

Josh: to a nearly 10% market share. And it seems like that number is going up only. (09:44):
undefined

Josh: They have the backing of China behind them. You know the Chinese CCP is going (09:48):
undefined

Josh: to be really pushing for them to win. (09:51):
undefined

Josh: Is this playing a role into the memory problem as well? Yes. (09:53):
undefined

Ejaaz: And it's not as much of a problem as people make it out to be. (09:57):
undefined

Ejaaz: So let me actually ask you this question. (10:01):
undefined

Ejaaz: Of the non-Chinese memory makers, so SK Hynix, Samsung, Micron. (10:03):
undefined

Ejaaz: Who would you think is the biggest region that they're selling all their memory (10:09):
undefined

Ejaaz: to? Is it the West or is it China? (10:13):
undefined

Josh: I would assume it's the West because we have all of these GPUs. (10:15):
undefined

Ejaaz: You'd be right. It's overwhelmingly the West. And the issue there is there's (10:19):
undefined

Ejaaz: not enough supply to meet the West's demand, right? (10:23):
undefined

Ejaaz: So guess who is starved of memory? (10:25):
undefined

Ejaaz: It's China. China star. (10:29):
undefined

Josh: Oh, those open source guys, huh? (10:30):
undefined

Ejaaz: Yeah, those open source guys. So listen, they're not getting access to any of (10:32):
undefined

Ejaaz: the American chips. NVIDIA has a trade restriction. They can't sell them frontier chips. (10:35):
undefined

Ejaaz: And they don't get access to any of SK Hynix and Samsung's memory chips because (10:40):
undefined

Ejaaz: they're selling it to the West. Micron is obviously selling it to the West as well. (10:45):
undefined

Ejaaz: So they have to kind of do their own thing. That company is the number four (10:48):
undefined

Ejaaz: memory provider now, CXMT, because of course, Chinese AI labs like Moonshark (10:52):
undefined

Ejaaz: creating Kimi K3, GPU creating GLM. (10:56):
undefined

Ejaaz: They also need memory for their GPUs to train their own AI models. (10:59):
undefined

Ejaaz: So CXMT stepped up and basically IPO'd and went up 500% in a single day, (11:02):
undefined

Ejaaz: making them the most valuable company in China. (11:08):
undefined

Ejaaz: Their valuation, I think right now, is roughly around the price of Micron or (11:12):
undefined

Ejaaz: the market cap of Micron. (11:18):
undefined

Ejaaz: And they did that in like a single day, just like the craziest IPO ever. (11:19):
undefined

Ejaaz: Now, the reason, again, for why this is the case is we are starved of memory (11:23):
undefined

Ejaaz: in AI. It's just a very simple thesis. (11:28):
undefined

Ejaaz: You need memory to remember everything that you type and talk to Claude and ChatGPT about. (11:31):
undefined

Ejaaz: You need memory to keep your agents running 24-7. And that memory demand isn't (11:35):
undefined

Ejaaz: just a linear line. I'm trying to figure out what this looks like in the camera, (11:40):
undefined

Ejaaz: but it's not a linear line. It is a completely exponential line. (11:43):
undefined

Ejaaz: And if you look at the demand growth for any of these memory supplies. (11:47):
undefined

Ejaaz: It literally looks like this. And you know what else looks like this? (11:50):
undefined

Ejaaz: The profit margins and the revenue that we're seeing. So whether it misses targets (11:53):
undefined

Ejaaz: by like a billion dollars or not, it does not matter. So that's one thing. (11:57):
undefined

Ejaaz: But there's two other news items why China is causing stocks to crash, Josh. The other one, (12:01):
undefined

Ejaaz: Have you heard of this company called asml based in the netherlands (12:07):
undefined

Josh: Yeah might have heard that like singular company that the entire world is. (12:10):
undefined

Ejaaz: Built up yeah yeah do you remember they create these like 300 million dollar (12:13):
undefined

Ejaaz: machines which are used by tsmc yeah exactly euv extreme ultraviolet lithography and they use (12:17):
undefined

Josh: Where they shoot they shoot little pieces of of light at tin and then the tin (12:24):
undefined

Josh: turns into light that doesn't exist anywhere else on the planet that was a banger this crazy scientific. (12:27):
undefined

Ejaaz: Company yeah exactly that was a good episode by the way for the ogs who know we're referencing. (12:32):
undefined

Josh: Go listen to that one because ASML is a crazy company. (12:37):
undefined

Ejaaz: It's such an awesome company. Anyway, so this company, it's one of a kind. (12:39):
undefined

Ejaaz: It's based in the Netherlands. They create these $300 million machines. (12:43):
undefined

Ejaaz: And I think they pump out a couple hundred a year. They're so hard to make. (12:47):
undefined

Ejaaz: They have teams and teams of people trying to create these things. (12:51):
undefined

Ejaaz: It is incredibly difficult to do. And it's very secretive. They have not released (12:54):
undefined

Ejaaz: any blueprints such that it has been super hard to replicate this. (12:58):
undefined

Ejaaz: They've tried many times in the West. Elon Musk has tried. (13:01):
undefined

Ejaaz: You just haven't been able to do it. And it is pinnacle to have these machines (13:04):
undefined

Ejaaz: to create next generation AI chips. So NVIDIA, very close to ASML. (13:07):
undefined

Ejaaz: China, a company in China, announced, very surprisingly, that they've been able (13:12):
undefined

Ejaaz: to replicate a version of these $300 million machines. (13:17):
undefined

Ejaaz: It's called DUV. It's called Deep Ultraviolet. So, it's not quite extreme, (13:21):
undefined

Ejaaz: but it's Deep Ultraviolet. And I have to stress, it is a prototype machine. (13:24):
undefined

Ejaaz: This hasn't been scaled. And let me ask you this, Josh. How many of these machines (13:29):
undefined

Ejaaz: do you think they're creating for the rest of this year or in a year that they're (13:33):
undefined

Ejaaz: targeting for a year? dude (13:37):
undefined

Josh: Not many it's like low hundreds dude. (13:37):
undefined

Ejaaz: No it's five (13:39):
undefined

Josh: Five oh shit okay that's way less than. (13:41):
undefined

Ejaaz: I'm aiming for 10 so like oh god it's a nothing burger but the market saw this (13:43):
undefined

Ejaaz: news and were like oh crap china's about to flood the market with euv machines (13:49):
undefined

Ejaaz: the the cost of all these gpus is gonna go down we're gonna have so many more (13:54):
undefined

Ejaaz: gpus we should just dump nvidia we should dump amd we should dump all these (13:57):
undefined

Ejaaz: memory stocks it makes no sense it has a massive overreaction (14:01):
undefined

Josh: Yeah it seems like we have this baked in trauma. I mean, there's like the, (14:04):
undefined

Josh: the bear thesis is kind of like the solar panel idea where China famously, (14:07):
undefined

Josh: they subsidized and then flooded the market with solar panels, (14:12):
undefined

Josh: completely collapsing prices everywhere. (14:15):
undefined

Josh: And because China is able to manufacture things at scale, they're able to kind (14:17):
undefined

Josh: of compete at a margin that other companies cannot. And the Chinese government (14:21):
undefined

Josh: is willing to back these companies and subsidize those companies in order to (14:25):
undefined

Josh: destroy the demand in other marketplaces. (14:29):
undefined

Josh: It's how China has always won. They've used their manufacturing capability and (14:32):
undefined

Josh: that connection with the government to subsidize these things to reach low prices (14:35):
undefined

Josh: that other companies cannot compete with. (14:39):
undefined

Josh: This is not the case for solar. This is just not really true. And it's, it's like, (14:41):
undefined

Josh: solar is this static technology it is this commodified thing whereas memory (14:47):
undefined

Josh: is very dynamic there's many different types of dynamic memory there's many (14:51):
undefined

Josh: different ways of making it there's very (14:55):
undefined

Josh: many different like custom architectures for it and that's just not really how (14:56):
undefined

Josh: it works you can't build a memory company to subsidize the prices of (15:00):
undefined

Josh: and and lower the cost relative to all the others because one the demand is (15:05):
undefined

Josh: so high and two there's so many different types i mean dram is kind of like (15:08):
undefined

Josh: tap water that's kind of what they're going for hbm is that premium bottled (15:12):
undefined

Josh: stuff. That's like that blue glass bottle that you see all the time. (15:15):
undefined

Josh: They're totally different things. (15:18):
undefined

Josh: And yet the market is reacting to this news as if they are the same. (15:19):
undefined

Josh: And I think that disconnect is probably where we feel a little optimistic and (15:23):
undefined

Josh: feel like perhaps this is a little bit oversold. (15:27):
undefined

Josh: Now, maybe from here, we get into the kind of unbundling of this thing and talking (15:30):
undefined

Josh: about where the money is going, because it's not just leaving the system. (15:35):
undefined

Josh: It is kind of shifting places. (15:39):
undefined

Josh: There is this unbundling of the AI trade happening. (15:41):
undefined

Josh: Maybe we could shed some light onto So kind of where that's headed to now. (15:44):
undefined

Ejaaz: The relative way to think about where the money is going right now in AI is, (15:47):
undefined

Ejaaz: it's going from the hyperscalers. It's going from the top AI labs, (15:51):
undefined

Ejaaz: such as Anthropic, OpenAI, Google, Meta. (15:55):
undefined

Ejaaz: They're spending copious amounts of money. I think the figure for this year (15:59):
undefined

Ejaaz: is something crazy like, what was it, like $250 billion or something like this on AI CapEx alone. (16:03):
undefined

Josh: Unbelievable amount. (16:10):
undefined

Ejaaz: Well, Google just recently reported their quarterly earnings. (16:11):
undefined

Josh: I think it's more than that. (16:13):
undefined

Ejaaz: Yeah, I think it might be more than that. It feels low when I said it. But Google's (16:14):
undefined

Ejaaz: Quarterly earnings reported that they've, for the first time since they IPO'd, (16:18):
undefined

Ejaaz: so 21 years, I believe, they've gone negative cash flow, which means they're (16:21):
undefined

Ejaaz: spending more money than they are taking in. The balances have been depleted, right? (16:26):
undefined

Ejaaz: And the craziest part about this is they're doubling down even more because (16:30):
undefined

Ejaaz: they see the opportunity. Now, think about it. Google's doing this. (16:34):
undefined

Ejaaz: Amazon's doing this. Meta's doing this. (16:37):
undefined

Ejaaz: These guys aren't stupid people. Like, they will only be doing this if they (16:39):
undefined

Ejaaz: see that there's real revenue coming through. (16:43):
undefined

Ejaaz: Amazon CEO, Andy Jassy, famously said this in his previous quarter. (16:45):
undefined

Ejaaz: He said, we are investing all this money because we are literally getting revenue (16:49):
undefined

Ejaaz: back from it almost immediately or six-month delay. So it makes sense for us (16:53):
undefined

Ejaaz: to just keep compounding this, right? (16:57):
undefined

Ejaaz: So the money is going from these companies into these semiconductor companies. (16:58):
undefined

Ejaaz: And that's what we're seeing. That's why SK Hynix had a record quarter where (17:03):
undefined

Ejaaz: they made the most money that they ever had, more than they did in 2025. (17:06):
undefined

Ejaaz: So the unboggling is this free cash flow going from the hyperscalers to the (17:09):
undefined

Ejaaz: semiconductor companies. So if you wanted to look at a layer to potentially (17:13):
undefined

Ejaaz: consider investing in or being focused on, it still is, as boring as the answer (17:16):
undefined

Ejaaz: is, semiconductors in general. (17:21):
undefined

Ejaaz: The other thing that I think is playing into this, Josh, not to bring up China (17:23):
undefined

Ejaaz: again, but like we have to because they've been so relevant this week, is open source. (17:26):
undefined

Ejaaz: A big critique from people right now is, huh, if I have an open source model (17:30):
undefined

Ejaaz: that I can run at home and is cheaper to run, why on earth would I need to be (17:34):
undefined

Ejaaz: spending millions and millions of dollars a year on AI? Why do I need all these GPUs? (17:38):
undefined

Ejaaz: Well, that is fundamentally wrong. And Gavin Baker actually did a really good (17:43):
undefined

Ejaaz: job explaining this, where he basically said, (17:47):
undefined

Ejaaz: Number one, these models, these open source models aren't cheap to run at home. (17:50):
undefined

Ejaaz: If you look at Kimi K3, it costs like between 1.1 to 1.2 million dollars to (17:55):
undefined

Ejaaz: run effectively at scale. (17:59):
undefined

Ejaaz: So like it's not available to the average consumer. (18:01):
undefined

Josh: Not getting that on your home PC. Not at all. (18:03):
undefined

Ejaaz: Number two, and this is a really good point. He said the hyperscalers or the (18:05):
undefined

Ejaaz: cloud service providers like Google, like Microsoft, like Meta, (18:08):
undefined

Ejaaz: like Elon Musk's SpaceX now, which is lending compute to Anthropic at 1.2 billion dollars a month. (18:12):
undefined

Ejaaz: They locked in really cheap contracts. those contracts expire at the end of the year. (18:17):
undefined

Ejaaz: What do you think they're going to do after those contracts expire? (18:23):
undefined

Ejaaz: They're going to re-rate it like 2x. And he makes the point here that like the (18:26):
undefined

Ejaaz: spot prices for GPU rentals have not slowed down. (18:30):
undefined

Ejaaz: They're 2x higher than the contracted rates that they were at the start of their contracts. (18:33):
undefined

Ejaaz: So the point is, whether you have an old GPU, whether you have new GPUs, (18:36):
undefined

Ejaaz: whether NVIDIA releases Vero Rubin in abundance, these GPUs are in such over-demand (18:39):
undefined

Ejaaz: that the prices for these things still go up. (18:45):
undefined

Ejaaz: So every fundamental building block for this unbundling, Josh, (18:47):
undefined

Ejaaz: is just money going to semiconductor stocks and semiconductor companies. (18:50):
undefined

Ejaaz: And I don't see any other way out of it right now. That's what it looks like. (18:54):
undefined

Josh: Yeah. And as we talk about the GPUs, I mean, we just had that fun visual on (18:57):
undefined

Josh: screen here, which shows you the anatomy of these chips and how nearly 50% of (19:00):
undefined

Josh: the costs are associated with this high bandwidth memory and everything else takes up for 50%. (19:05):
undefined

Josh: So the most important object in the world right now is the GPU. (19:10):
undefined

Josh: The most critical component of the GPU, which accounts for about half of the cost, is this memory. (19:14):
undefined

Josh: So it's like, okay, well, we have a seemingly infinite demand for GPUs, (19:19):
undefined

Josh: therefore infinite demand for memory, therefore... (19:23):
undefined

Josh: Infinite demand for all of the supply of all of these companies like when they (19:26):
undefined

Josh: get re-rated it should go up, (19:29):
undefined

Josh: right like that math just seems like it checks out so that's it feels like a (19:32):
undefined

Josh: little confusing as to why this is happening again there's a lot of external (19:35):
undefined

Josh: factors there's a lot of kind of overreactions baked in but it seems like (19:38):
undefined

Josh: as we're just looking at this like kind of pragmatically all of these numbers (19:42):
undefined

Josh: are checking out also it's insane as i'm looking at this that (19:45):
undefined

Josh: nvidia sells these chips for forty thousand dollars at an 84 percent gross margin (19:48):
undefined

Josh: like oh my god good for you man good for you (19:53):
undefined

Josh: but there is this interesting like inverted capex thing happening where traditionally (19:55):
undefined

Josh: in technology for the last two three decades all of the funds have gone from the bottom up (20:00):
undefined

Josh: so it's been from the consumers from the enterprise paying into these huge margin (20:06):
undefined

Josh: companies like nvidia to raise their profit margins and for the first time we're (20:09):
undefined

Josh: having the reverse effect where all of the. (20:13):
undefined

Josh: Companies that have collected all this money over time like google are now spending (20:16):
undefined

Josh: it for the first time ever in its history faster than it has made it and the (20:19):
undefined

Josh: downstream effects of that seem to be pretty huge and i mean we have lisa sue she's here (20:23):
undefined

Josh: reading um yahoo finance actually who is just sharing the idea that ai adoption (20:29):
undefined

Josh: is faster than any of us thought and there is no end in sight at least from (20:34):
undefined

Josh: what we can see of where the demand for tokens is going to stop where the demand (20:38):
undefined

Josh: for all this compute is going to stop. (20:42):
undefined

Josh: And as of right now, there's still this like tremendous shortage. (20:44):
undefined

Josh: If anyone was producing more memory, I think it would just get eaten up right (20:47):
undefined

Josh: away. And I think that's kind of the conclusion of this episode, (20:50):
undefined

Josh: generally speaking, is that like, hey, there's no end to the demand curve in sight. (20:52):
undefined

Josh: And the more GPUs, the more memory, the more power we could apply to all this, (20:56):
undefined

Josh: the better off everyone's going to be and the more hungry everyone's going to (21:01):
undefined

Josh: be to generate more tokens. Yeah. (21:03):
undefined

Ejaaz: And just to be clear, a point that Lisa Sue makes in this clip that we're showing (21:04):
undefined

Ejaaz: on screen right now is it's not just general AI demand that is causing GPU prices (21:09):
undefined

Ejaaz: and memory stock prices or memory demand to accelerate. (21:15):
undefined

Ejaaz: It's also this thing called agentic AI. AI agents in general have exploded over (21:18):
undefined

Ejaaz: the last couple of months. (21:24):
undefined

Ejaaz: And guess what these AI agents need to be able to access tools, (21:25):
undefined

Ejaaz: to do these tools, to orchestrate all the tasks, to run 24-7. (21:28):
undefined

Ejaaz: You know, you see all these fun viral examples on Twitter where, (21:31):
undefined

Ejaaz: you know, you set and forget a prompt and you come back and like you have a (21:34):
undefined

Ejaaz: full triple A game like we saw this week with Cloud Opus 5, it all requires CPUs. (21:37):
undefined

Ejaaz: CPUs require a lot of memory. So the point is, as AI agents scale, (21:43):
undefined

Ejaaz: you're going to need more of these fundamental things. (21:48):
undefined

Josh: That's funny you should say that because speaking of agents, (21:51):
undefined

Josh: we have a sponsor for the show Ledger, which we have to shout out here because (21:53):
undefined

Josh: if you are building with AI agents, you're worried about security. (21:57):
undefined

Josh: An agent with unchecked access is a brilliant but probabilistic teammate with (22:00):
undefined

Josh: access to important secrets. Ledger's agent stack fixes that with four open source tools. (22:04):
undefined

Josh: DMK skills and CLIs for Ledger wallet, Ledger Enterprise, and Ledger Enterprise (22:09):
undefined

Josh: multi-sig that let your agents work with crypto while you stay in full control. (22:13):
undefined

Josh: You could tell your agent, rebalance my wallet, ETH to 40%. (22:17):
undefined

Josh: Do it in the fewest swaps. It does the math, drafts the swaps, (22:21):
undefined

Josh: and stops. The transaction lands on your Ledger signer and nothing moves until you approve it. (22:24):
undefined

Josh: Agents propose, humans approve ledger signers enforce and i mean this is the (22:28):
undefined

Josh: best part obviously it works with cloud code codex cursor all the places that (22:32):
undefined

Josh: you are actively working um and it's open source ready today available on developers.ledger.com (22:35):
undefined

Josh: also linked in the description and thank you so much to ledger for supporting (22:41):
undefined

Josh: this episode now let's close this thing out you jazz we gotta let people know, (22:44):
undefined

Josh: what do you do with all this information now like okay we've given you kind (22:49):
undefined

Josh: of this like idea here's what's happening here's (22:52):
undefined

Josh: the market demands here is the sell-off that's happening we are in a currently (22:54):
undefined

Josh: in a spot that's worse than 2008 although it doesn't feel like it probably because (23:00):
undefined

Josh: we speed ran it and because everyone's already up so much, (23:02):
undefined

Josh: but like what do we do here? What do you do with this information? (23:05):
undefined

Ejaaz: Okay so here's my grounded take and you know that's Rich coming from the show (23:09):
undefined

Ejaaz: well we're quite optimistic but I'm going to try and be grounded. (23:13):
undefined

Ejaaz: The AI hype (23:16):
undefined

Ejaaz: Definitely drove markets to pretty insane valuations. (23:19):
undefined

Ejaaz: And I'm not denying that. But what I will say is the future of AI and the economic (23:23):
undefined

Ejaaz: value that it'll generate is just like the Wild West right now. (23:31):
undefined

Ejaaz: It is incredibly hard to predict and even conceive. There are many theories (23:35):
undefined

Ejaaz: out there, many skeptics out there. (23:39):
undefined

Ejaaz: So the bet you need to make if you're listening to this is, do you believe that (23:41):
undefined

Ejaaz: AI LLMs, Claude, ChatGPT, that AI agents, that GPUs are going to be in absurd demand going forward? (23:47):
undefined

Ejaaz: Do you think the demand for AI products, are you using AI more over the last (23:55):
undefined

Ejaaz: couple of months? Like, you know, answer that question. (23:58):
undefined

Ejaaz: If you believe that scales out, remember, it's only like something upset, (24:00):
undefined

Ejaaz: like 5% of the population that's even using AI LLMs beyond just a Google search. (24:04):
undefined

Ejaaz: If you believe that scales, then you believe and bet that these semiconductor (24:08):
undefined

Ejaaz: companies that are creating the fundamental materials. (24:13):
undefined

Ejaaz: You know, we showed this on our screen earlier to build these chips that are (24:16):
undefined

Ejaaz: required, whether you like it or not to run whatever types of AMLs, (24:20):
undefined

Ejaaz: whether it's open source or closed source, then you're betting that these companies (24:23):
undefined

Ejaaz: are going to be more in demand. (24:25):
undefined

Ejaaz: And you're betting that these profit margins and revenue is going to increasingly growth. (24:26):
undefined

Ejaaz: And guess what? Those are the fundamental drivers of whether a company's stock (24:31):
undefined

Ejaaz: prices is going to go up. And I think that the market is completely unjustified right now. (24:34):
undefined

Ejaaz: And I think we're going to look back on this in a year. I'm making my stake. (24:38):
undefined

Ejaaz: I'm making my claim. I'm making my prediction. (24:40):
undefined

Ejaaz: And we're going to think that these stock prices will add absurd valuations. (24:42):
undefined

Josh: Yeah, well, here's kind of how I think about it too, is like on a personal note, (24:46):
undefined

Josh: I'm more of an investor than a speculator. (24:49):
undefined

Josh: And that is why it's very easy to feel constantly optimistic. (24:52):
undefined

Josh: It's like, I very firmly believe in the idea that we are going to need a lot (24:56):
undefined

Josh: more tokens, a lot more compute, a lot more energy over a long period of time. (25:00):
undefined

Josh: How long it takes to get there is unknown, but that doesn't really matter. (25:03):
undefined

Josh: If you have a low time preference where it doesn't matter if this takes six (25:07):
undefined

Josh: months or six years or 60 years you just kind of directionally know where it's (25:11):
undefined

Josh: going to go then making these bets and dealing with the volatility makes things much easier, (25:15):
undefined

Josh: This directionally feels like a trading opportunity. This is people who are (25:18):
undefined

Josh: selling off their profit. This is people who are positioning themselves to make a short buck. (25:22):
undefined

Josh: That doesn't need to actually be the case if you believe in this long-term. (25:26):
undefined

Josh: And I think that's probably where we can wrap up this episode today. (25:29):
undefined

Josh: So with that, yeah, thank you all for watching. That's the state of memory. (25:31):
undefined

Josh: It was crazy sad to find out that we lost more money recently than 2008. (25:36):
undefined

Josh: And we don't feel like it because clearly we've been printing a lot more dollars (25:40):
undefined

Josh: and those numbers need to go up a lot higher to feel the same thing. (25:42):
undefined

Josh: But that is kind of where we stand. there is this discrepancy between memory (25:45):
undefined

Josh: stock prices and the actual demand for these items and yeah i think that's pretty (25:49):
undefined

Josh: much it so if you enjoyed the show please don't forget to share it with a friend (25:54):
undefined

Josh: who might also enjoy rate us (25:56):
undefined

Josh: on your favorite podcast player leave a comment if we are too optimistic or (25:58):
undefined

Josh: if we need to dial things back a little bit or (26:02):
undefined

Josh: or if you disagree in why and what stock you are investing in and choosing to (26:04):
undefined

Josh: gamble on um each has any closing thoughts while we wrap this up. (26:08):
undefined

Ejaaz: That is it thank you so much for listening and we will see you on the next one guys (26:11):
undefined