Loading summary
Transcript110 lines
- [00:00]
A
UKG Their HR pay and workforce management tools help business leaders empower their people. Because when work works, everything works. Learn more@ukg.com work@charmin we heard you shouldn't talk about going to the bathroom in public, so we decided to sing about it. Charmin Ultra strong you can use less better than the rest Charmin Ultra strong Booty past the clean text salmon weave texture it's the best Study up, teach a lesson on fresh your booty past the cle clean text Charmin Ultra strong Charmin Ultra strong with diamond weave texture cleans better than the leading one plaid brand so you can use less. Enjoy the go with Charmin.
- [00:45]
B
Bloomberg Audio Studios Podcasts Radio news. Hello and welcome to another episode of the Odd Lots podcast. I'm Joe Weisenthal.
- [01:05]
C
And I'm Tracy Alloway.
- [01:07]
B
Tracy, you know what I'm worried about? I don't know if I'm worried about it, but it kind of feels like the direction things are going, which is that like all energy, industrial commodities, anything that we use for any purpose is just going to like feed the AI beast. And us humans, we're just going to get left out in the cold. They're like nothing for you. We got to like feed, we got to feed it all to the AI. And maybe in 10 or 20, 50 years the AI is so powerful they just decide, why, why do humans get any of this? We just, we should keep it all to ourselves.
- [01:40]
C
A legitimate concern, I would say. I mean, to some extent we are already seeing this crowding out effect, right? So energy prices in certain areas have been going up and most recently memory chip prices. So this has been all over earnings recently. You had companies like Apple saying that because of a crunch in memory chip supply, they might have to either raise their prices or like cut down on amount of phones they use. Nintendo, if you pull up a chart of Nintendo shares right now, they are just getting hammered and supposedly it is all because of this memory chip shortage. By the way, do you remember, do you remember buying your first PC in like you bought one before I did, I think, but mine was in the mid-90s. Yeah.
- [02:26]
B
Yeah, I think that's about when I got one.
- [02:28]
C
Do you remember how much memory those things had?
- [02:30]
B
Oh, like nothing.
- [02:31]
C
Yeah, like I looked it up. Something like less than 10 megabytes for a bunch of them. You know how much PCs have nowadays?
- [02:38]
B
Actually, it's a good, I have no idea.
- [02:40]
C
I think it's, let's see, 16 gigabytes.
- [02:43]
B
Yeah.
- [02:43]
C
Isn't that crazy?
- [02:45]
B
It is, it is really Crazy. I remember like. Do you remember zip drives?
- [02:48]
A
Yeah.
- [02:49]
C
Oh my God.
- [02:49]
B
And, or like. And there were all these special, you know, obviously various memory peripherals that you could buy and they came in sort of dedicated cartridges and stuff like that. I think they were called Zip Drive. Yeah, I think.
- [03:01]
A
Right.
- [03:01]
B
There were zip drives. I don't remember.
- [03:03]
C
Yeah, it was. Yeah, that's right.
- [03:04]
B
Buying additional memory storage, etc. Used to be this big part of the consumer process. I'm actually glad, you know, my son really wants a Nintendo but if they can't produce them anymore because they can't get their memory, then I'll say, sorry, no video game.
- [03:17]
C
That's so mean. You're going to have to sit them down and be like, I'm so sorry son. There's this thing called the supply shortage. I'm going to use this as a learning opportunity to explain how global supply chains actually work.
- [03:28]
B
Totally. You know what the other, the one interesting thing is if you look at spot dram prices which we have on the terminal, this is I think maybe the interesting dynamic. There's a lot of AI related things on the terminal that have been surging for several years. Most notably some of the big chip companies, Nvidia, etc. If you look at a chart of spot dram prices, the big surge wasn't until late last year and then since then it's gone nuts. So for a long time, even as this AI story, the AI meme, the AI industry like propagated people were really paying attention to that area and then suddenly it's just gone like you know, the haywire. Haywire, yeah. And so obviously huge supply, demand imbalances. We need to understand what's going on because I think the, the stakes are, you know, it's getting so much. As you mentioned, a lot of companies are losing out pretty big, others are making a lot of money. Yeah, we sort of need to get a handle on what's going on with memory.
- [04:19]
C
Let's do it.
- [04:19]
B
I'm really excited to say we really do have the perfect guest. We are going to be speaking with Ray Wong. He's an analyst at Semianlysis and he's recently pub published a brand new report on exactly this topic talking about a memory super cycle and so forth. So truly the perfect guest. As I mentioned, he comes to us from Korea. Ray, thank you so much for coming on odd lots.
- [04:42]
A
Hey, nice to meet you both and thanks for having me.
- [04:45]
B
So many questions. Really appreciate you joining us. Let's just start with the basics, like what is it about AI or what is going on, what is the core idea such that demand for memory in various forms is absolutely seem to be booming right now in a way that must have caught the supply side of the equation very off guard.
- [05:04]
A
Yeah, I think to really look at kind of the sort of imbalanced supply demand dynamics, really we want to look at what happened a few years ago, right? Because your capacity to check your investment few years ago and what's happening in 2023 and 2024, it's sort of a down cycle, right? Because remember during the COVID all the companies are trying to flock in to buy tons of dram and there's a lot of purchases there and there's sort of like upcycle, right? But it's like a very short upcycle because when Covid gradually getting better, you actually don't need that much dram. And back then the company will be like hey, you actually don't have a sustainable demand, right? So they don't want to over invest to expand their capacity. And when you go into the down cycle, your capex typically being very constrained, marry company from their perspective, they try to be conservative. So that leads to 2024 and 2025, that your incremental wafer capacity that goes to DRAM is actually quite limited. But what's happening on the demand end is actually the demand was accelerating so fast, right? One way to look at it is to look at media's earnings, right? And when you were new, on one side your demand is accelerating so fast, on the other hand, on the supply side, you're supply just couldn't catch up. And there's more nuances behind the supply, right? Because before they are just supplying all the commodity dram like DDR, LPDDR, right? All the memory that goes to PCs, goes to laptops, goes to your mobiles, goes to Nintendo switches, right? But what's happening in this AI is there's a new thing called hbm. Think about it as like a specialized memory directly for AI accelerator. And that memory is extremely DRAM wafer intensive. To give you a sense, right? On the same wafer basis you can produce three more bits if you do commodity dram, but you can only produce one bit of HBM and that ratio will actually go in higher when you go to HBM4, HBM5 in the next few years. So in that way, on the same wafer basis you can only produce that much hbm. And right now what's happening is like hey, this HBM is super profitable and why not we dedicate more Wafer for hbm. But so here's the problem, right? You only have that much wafer and you need try to do everything. So that's quanting out a lot of the supply going to the commodity dram. Right? So we have sort of both things is happening and essentially comes down to like I think really things start really kind of people aware it's like probably Q2, 20, 25 that people realize hey, we actually couldn't find enough DRAM, both enough commodity DRAM and also HBN for their demand.
- [07:47]
C
So you mentioned the, the C word just then. Oh wait, not that C word, commodities. You mentioned commodities. And when I'm reading these articles about a memory chip shortage, people use commodity type language, right? So you hear super cycle a lot. You hear commoditized chips versus, I don't know, other types of chips. Are they customized? I don't think so. But anyway, how much does this particular industry actually resemble a commodities like industry? And then maybe explain a little bit more why historically it has been really, really cyclical.
- [08:21]
A
Yeah, so I think a couple reasons, right? You know, fundamentally the sort of the cost per bit for DRAM was actually coming down every single year significantly, right? Even in the most recent year was also coming down. But like the extent that they are coming down for every single year is actually slowing down over the past like 10, 20 years, right? So the cost was keep going down, right? So because it's hard to become the most competitive part, right. Another thing is there's sort of the committee that set up the standard for dram, right? So well, you are doing the similar products, right? You're just having, oh, this is kind of a little better power efficiency, the performance was a little better right here and there. So it's hard for a memory supplier to significantly differentiate their products compared to their competitor, right? So essentially what's the things that will compete in the end will be the market price prices, right. And then when you are selling the similar products and you are targeting similar audience, right? That sort of like are the commodity, you know, sort of semiconductor commodity products to me. And I think those are the two main characteristics I'm seeing for dram. And I think that's a little, little different for hbm.
- [09:33]
B
Tell me, explain this. So there's commodity DRAM and then there's hbm. What? HBM is a type of dram, is that what you're saying?
- [09:42]
A
Yes.
- [09:43]
B
Can you explain this further and walk us through what the commodity version and why there is this thing called hbm High bandwidth memory. What that's all about.
- [09:52]
A
Yeah. So essentially the emergence of HBN is really go down to a kind of model layer because this is when you try to scale and try scaling your AI model, right? Your FOB is increasing significantly every single year, right? Because your compute was increased a lot faster but your memory bandwidth was actually sort of. It's increasing but it's very limited. Especially if you are using the memory from kind of traditional commodity DDRAM that the bandwidth is actually very limited. So from a memory supply they are thinking hey, how can we expand the bandwidth to supporting the memory that going to scale the AI, right? So let's consult HBN. So the way they do HBN is so they are stacking multiply 8 or 12 or even future 16 deer and dice together, right? So that making you essentially having higher bandwidth and have enough capacity to support AI developments over the past few years and in the next few years. And that's very different, right? You are stacking all these dice together compared to hey, we just have one DRAM products as a chip, right? And that also comes down to a point that HBM manufacturing also the packaging, both front end and back end is a lot more complex compared to commodity dram. And that's why I say commodity DRAM is more like commoditized. So I think HPN is less so because it's a lot more complex. Right. And for supplier you can actually differentiate either from the front end or back end technology. And when your yield is better, right. It gives you better margin, better ability to compete with your competitor. Right. And stuff like thermal, things like that is the thing that will really stand out. Right. And that's very different.
- [11:32]
C
Commodity theorem getting more and more layers on hbm. It kind of reminds me of the razor wars where like you know, Gillette would unveil a three blade thing and then the next day someone has like 12 or whatever.
- [11:59]
A
UKG their HR pay and workforce management tools help business leaders empower their people. Because when work works, everything works. Learn more@ukg.com work premier protein.
- [12:13]
C
It's for getting after life, not just Fitness. With 30 grams of protein, 160 calories and no sugar added, helping people fuel their joyful lives. With Premier Protein, you can say yes to more. Whether it's crushing a big presentation at work, building an epic fort with the kids or hitting the hiking trail with friends, Premier Protein offers delicious flavors like cafe latte, chocolate, caramel, vanilla, strawberry and cake batter to name a few. Find your favorite flavor@premierprotein.com so if I'm a company Like Nintendo, how am I actually purchasing chips? Because I think this is going to be important when we talk about what's going on right now. But do I have like a forward contract with a major supplier or do I have, I don't know, a warehouse full of chips that I bought previously? How do they actually manage that?
- [13:02]
A
So there's a couple of ways they do it, right? They will do it in terms of like kind of, they will do a different way like in different environments and also different vendors. Right. So it's hard to say like hey, this is exactly, exactly the way they do it. But typically I think you can think about like hey, this is amount of the chips we're going to purchase in the next four months and this is amount we're going to commit, right? And we're going to sign the contract pricing. Whether this term will be four months or for the lta, you can go as long as like to a year or even two years, right? So that's typical way you think about it but typically you'll be like hey, we signed this contract for four months, right? And by the end, sorry, for three months, for a quarter. So by the end of the quarter you negotiate a new contract pricing for the next quarter. So that's typically the way it work. But you know, the problem for right now is for consumer electronics is like the pricing probably sort of, I will say secondary because if you couldn't secure value, you couldn't even make any money. So number one is securing volume, whether that make sure that fits to the demand outlook you are seeing. Two, it's whether you can negotiate the best pricing you can. But given the overall pricing environment was so volatile, was so escalating so much, it's hard to not take in some kinds of margin impact because essentially you try to get a chips and you cannot just get too much away from your vendors.
- [14:22]
B
Yeah, I want to get more into the demand destruction element because I think this is very important to think about where the equilibrium is going. Just to clarify, what is it about AI specifically? Is it in the training, is it in the inference? Where specifically in the process of building out AI services does this voracious demand for DRAM come from?
- [14:43]
A
So honestly, to be very honest, I think it's everywhere AI training you need tons of HBM that just everyone know you also need a lot of the CPU DRAM and that's LPDDR and the DDR that goes to the server. You also need hbn. Right? So that's on the training part in the inferencing you also need hbm, right? And you also need the CPU DRAM to go through loss workload. Right? But the main one will be the HBM and even the inferencing. Right now they are separating to pre fill and decode and I think 80% I think right now the most important thing will be to code, right? And decode is super memory bound and memory intensive and hbn, the memory importance of there will only increase to my understanding, especially given the long context window continue to expand and the inference usage adoptions continue to going up. Right. So I think those are typical way I would think about those. Right. And the last thing is I think a lot of people talk about these days is about agentic AI. Agentic AI, you want to power those. You need a lot of cpu, you know, a lot of CPU based server, right? So what's including the CPU server? There's a lot of theorem in there, right. And a lot of agent workload also mean you need a lot of inference as well, right. And that go back to my previous point. You need a lot of HBN and dran, right? So I will say like kind of dran and I got HBN and DRAN is kind of everywhere and we can kind of talk about storage, but storage is so everywhere. You know, in doing this process I.
- [16:08]
C
Have a basic question, but why does AI need a lot of memory at all? So you just explained it's important for inference, it's important for training, whatever. But why?
- [16:22]
A
Yeah, I think this needs to go down to the model layer. Because essentially when you are trying to train, let's say you want to train the model and then you just need a pure tons of data, you need to just write tons of data there to train your model. And then for the inferencing part you need to one data after another to kind of compute one after another. Sort of a chain of salt, right? Yeah. So to do that process, how can you keep the previous data you process to keep it in there and to process to another data that coming out after, right? So that's why you need a lot of memory. And go back to my previous point, right. The column context window basically means when the things that you are processing longer and longer, right? How can you have enough memory to process those? Right. And one for a good example, if you use ChatGPT, you guys definitely remember right back to ChatGPT which starts coming out, people just like hey, what's the weather? Write a poem right now is or last year people are doing like hey, can you write a 20 page report how tensor can happen on people. Right. For example, then we'll give like 20 page report and they will wait like five minutes. Right. And so if you think about the process like they do all the calculation, all the researches and also give you the output tokens that the answer you get from your prompt, that's significant way longer than the prompts that you are given. Right. And that's also a lot longer response that you are getting. When you go back to the GPT3 and remember the usage we know back to GPT3 was forgot the monthly daily users, but right now I think the user for GPT was 800 million. Right. And we are not included the users for the Gemini, for cloud, for Xai. Right. So if you multiply those effects, I think the memory the man is very significant and it's very clear to me.
- [18:08]
B
Yeah, it's really interesting. I saw some people were talking about this, but it really strikes me in my own usage of AI just basically how much more I use it as capabilities have grown like it was sort of Cute in late 2022, it's like, oh, get it to write a poem about this. And it's like, okay. And then I would go, you know, several weeks sometimes without having anything to query. And then you know, these days like I'm using it all the time like trying to do things with code or scrape data, et cetera. So to just overall token consumption has grown massively with increased capabilities. All right, let's talk about the demand destruction element. Any commodity is going to be. The price is going to be a function of supply and demand and eventually prices get high enough such that certain forms of demand just are no longer economical. And it's like, you know what, we're just not going to make as many video games or cameras or whatever else uses memory. Are we starting to see any of that yet? Such that certain uses and consumption of memory.
- [19:10]
C
So Joe doesn't have to buy a switch for his son.
- [19:13]
B
Yeah, that's right.
- [19:15]
A
Yeah. I actually bought a switch recently. So the price.
- [19:18]
B
You should resell it. You can resell it.
- [19:22]
A
I don't know. We'll see how the arbitrage there works, but I think we're already seeing a lot of impacts actually. But you kind of. It kind of vary in different kinds of products. Right. In PC we are surely seeing that Dell Nanovo answer Asus, they are all having a price hike for kind of different products. Right. And on the kind of demand side you also seen that I think for the Chinese smartphone market Right. A lot of the analysts, a lot of the research firm, a lot of the company was saying like they are cutting off their smartphone outlook for example. I think Mediatek recently left earnings. Right. They are saying they are cutting off the mobile to. I think it was 12 or 10 to 15% of the 2026 outlook. That's very significant. Right. Y saying this year is supposed to be 100 million but it will essentially goes to 90 million. Right? That's very significant. So I think we are already seeing that but kind of vary different places. Right. Apple will be a great example here because they are saying hey, like hey, we have all the price hike, right. But we actually manage pretty well. They are saying the real sort of meaningful margin impact will actually show up in the second half, 2026 instead of like the first quarter or second quarter. I think those sort of the things I'm seeing. So it's definitely already showing out in the market. Right. I think what you can expect this year will be two things. One is sort of demand destruction in PCs, right. Water, consumer electronics or the dispec. Whether it's the memory disp. That's very straightforward. But also the camera.
- [20:49]
B
Right.
- [20:50]
A
The camera is also very important. I'm hearing some of the display on the camera as well. And you know, also on the sort of the what we call like delay launch. Right. You don't want to launch your new products that have higher price than initially thought which will impact your new launch performance, which usually the new launch product has better asp better margin compared to your previous generation products. Right. I think those are the things that we're going to see in the coming months, in the coming quarters. I think we'll show impact and people will start feeling it.
- [21:19]
C
Joe, do you think we're going to get people like stripping old Nintendo's for memory chips? Like remember how during the commodities super cycle in like the early 2000 and tens. I guess there are all those stories about people like stealing electricity wires.
- [21:34]
B
Yeah, it happened again in 2021 too I think. Yeah, we might, we might get some of that.
- [21:38]
C
Well actually on that note, is there like a short term fix for this problem either in terms of design, maybe you can make the products more efficient in some way so that it needs less memory or recycling old memory chips. I don't know.
- [21:54]
A
Yeah, I think from the demand side it's a lot difficult from an OEM perspective. There's only that much you can do. Right. You can dispect. Right. But I don't think that addresses the fundamental issues. Right. You are just dispect your products and so you can ship. But there is double edged sword, right. When you downgrade your products that actually look bad compared to some of your peers who doesn't dispec, right. And make your product less competitive that impact your final sort of quarterly or annual shipment. I think what needs to be addressed or what can be done is probably on the supply side. Right? So supply side we are mainly talking about a couple of vendors, right? The Micron, Hynix, Samsung and for legacy ones you have wind bound Nanya and China like CSMT and jhitc. Right? Those people. So I think the most direct way right now this year because this year the really challenge is there's a clean room constraint. So clean room constraint is basically you have limited space that you can put all the equipments and start manufacturing chips in the fab. So you have clean range shortage in all three major memory makers, right? So your incremental wafer capacity coming online for 2026 will actually be quite limited. So in that sense how can you produce more bits, more DRAM bits while having only limited incremental wafer capacity? The only thing you can do it's non migration. So in different processes now you have the most advances. Right now it's 1C going down, have 1B, 1A and 1X, blah blah blah. Right. I think right now they are trying to do is migrate as much as they can to 1B and 1C which on the same wafer basis they will produce a lot more bits compared to a legacy knot. Right? So doing that way on the same way for basis you can produce more bits. So that should produce more supply even though your wafer capacity is constrained. But the challenge is two things. One, how fast you can rent about this non migration, right? Non migration is typically difficult. Well, I won't say difficult. It takes time to go into the most advanced narrow, you need EUV machine, you need all different kind of manufacturing processes and not every fab can already prepare to doing that. Right? So it will take time to ramp up all those production for the new events. Nut also you are having the challenge to like hey, okay, I want to do that migration, right? But how are we going to balance the capacity with hbn? Again that's another issue. You want to kind of think about it, right? Because some of the HBN processnap is actually using for Hynix they are using 1B so that's actually the events processor, right? So even you are including 1B but at the same time you want to do more hbm. So like the increase here is actually quite limited. So I think those are the way potentially fix the issue. But you know, know from our house view, right, even we factor all those in, I think this year we're still going to see quite a significant shortage.
- [24:47]
B
One of the themes that comes up when we talk about any commodities super cycle is that sure, higher prices will eventually elicit more production, more supply, but in the meantime, you know, there's a sweet spot for the existing producers. Especially when you have just a small number of producers that can produce at scale, they enjoy massive profits, right? And they're sort of going to be reluctant to build out new fabs because that's money going out the door. It also means lower prices. So there's always this sort of Goldilocks period for them before the supply comes online where they're just raking in profits. What do we see is the industry's impulse to invest and I'm curious like you know, whether it's Chinese producers, are they in a position to potentially undercut the profits of the Korean makers? Are the Korean makers seeing themselves as having a long window where they can enjoy large profits before they have to invest? Like what is the thinking on the supply side about those big capital outlays?
- [25:45]
A
Yeah, that's a very good strategic question. I think every management will probably give you a sort of different answer and kind of on a high level probably same. But I think in general, right, like you know, you know, look at like historical cycle, it's like if we are seeing quite sustainable demand in the next few years and then clearly the capacity couldn't catch all the demand, they are going to expand the capacity in some time, right? It probably not going to be like hey, we announce in 2026, the capacity really coming online meaningfully is probably 2028. So I think they will still announce, right? And we are seeing that from Micron, right? Micron is announcing a new NFAB in Singapore. They are expanding, they have two fab currently it's under construction in us, right. Also doing tons of non migration and recently acquired the new fab from PSMC which is a Taiwanese chip maker. And then all those moves you are already seeing some sign of that. And another sign you are seeing it's the Capex. Both Hynix, Samsung and Micron. Their drain Capex has actually increased quite significant this year and expect similar trend in 2027. Given I think they're going to probably try to expand more capacity and invest more in both HBN as well as Equipment that goes to advanced equipment. Right. That go back to the point I was talking about the non migration. If you want non migration you need to use more advanced equipment to produce more advanced chips. Yeah.
- [27:05]
C
It really does remind me of the oil industry right where like there's a bust in the underlying commodity price and everyone starts talking about being disciplined on capex spend and so it takes a while to ramp up. I wanted to ask about I guess the balance between HBM and other types of dram. Are people or how are the actual chip makers thinking about that? And is there, is there a chance that because hbm, my understanding is that it has higher profit margins. Is there a chance that everyone just pours into HBM and kind of leaves the more basic stuff in the dust?
- [27:44]
A
Yeah, so actually it's very interesting I think what probably things folk here where you wrote about this for our institutional client, it's that the margin typically we think HBN has higher margin which is definitely true but what's happening, we see all this spot price, contract price going up so much. Right. The margin of the commodity rand right now is actually higher than hbn. So here that created a real dilemma that Tracy, that you talk about it because when your margin of commodity is actually going higher why would you make more hbm? Like why? Right. But I think if you really think about a long term perspective from a memory supplier before the all AI boom your demand is really driving from couple a market PC, mobile, automotive, industrial, blah blah blah. HBN right now is surge as a new growth driver for the company which we believe will last for the next few years. And for the memory supplier they are thinking the same. And if you are not continuing to push your capacity, push your technology when you're lagging behind, it's hard to coming back. And they go back to my previous point that this is an area that is relatively easier to differentiate your products to get more market share more meaningfully. Right. So I think three memory makers are very, very value hbm even though right now the margin of commodity DRAM is higher. Right. I think they will still invest quite a lot on HBM and they will do their best to sort of balance the commodity DRAM market. But I think this year at least the number I'm seeing, it's still quite significant shortage. Yeah.
- [29:20]
B
How do the Chinese producers compare to the Korean producers? Is there a, is there a gap?
- [29:27]
A
Oh yeah, I think there's definitely the gap. I think like you know, if you look at the memory in general, I think it's probably three years Three years, it's probably four years. I would say that this is very like, you know, timeless of today. You know, who knows what's going to happen in the future. But I think, you know, there's definitely a gap. You know, when I think about the competition pressure from Chinese memory makers. We always want to kind of separate the Chinese market and also the ex China market. Let's also the memory supply was looking at it, right? So Chinese market is probably about 25% of the global drain demand. So really the competition is happening in China because for the leading Chinese memory supplier like CSMT for example, I believe like 90% or 95% of the revenue is coming from China and Hong Kong. So meaning most of the DRAM is really competing Chinese market competing with Micron, Hynix and Samsung. I will say that right now on the events on the high end, right, I think it was still dominated by stream memory makers but CSMD are gaining two momentum. One, they are probably getting sort of a low end and medium products and they are also benefited by Chinese government's policy in terms of the self sufficiency drive whether it's for the commodity dram. Also right now they are pushing for HBN development that support Chinese AI hardware.
- [30:42]
C
So if there's a shortage and it seems like the shortage is going to be with us for some time, how are the chip makers actually allocating what supply they have? Does it go to, you know, the company that I think is going to be really big and important in the future like I don't know, an OpenAI or something or does it go to an existing customer that's been buying in volume from me for years?
- [31:07]
A
Yeah, I think there's no doubt to me that some of the highest tier customer will get a volume for sure. Of course there's a more complex pricing negotiation behind but I will say the volume will still be the most important thing and I think that's probably what they're going to do. Right. And besides thinking about buy customer I will think about kind of by sector I will say the server DRAM and HBN two things together will be the top, top priority because for the whole a market in DRAM we are seeing actually HBN and server DRAM together it's more than half of the DRAM market. Right. And that's important and it's growing so fast. Right. And it's unlikely mobile. Right. Mobile is kind of flat. The mobile demand is really driven by the increased DRAM contents in the mobile which is quite limited. Right. If you buy iPhone over the past year, you know, they are actually quite limited on their drain content increase. So I think those are the top two priorities for the memory company to focus on in the next two, three years. And we'll see what happens after second half of 2027 or 2028 if there's some structural change. But I think ServerDrake and HBM will be the top prior.
- [32:34]
B
You know, the other day I was very lazy, which is not real for me. Just that one day, just one day and I used Claude code, you know, I have a bunch of screenshots on the desktop of my computer and I. It sort of makes my desktop look like a mess and I can't see folders so I like drag and drop them and put them in the recycling bin. And I was very lazy and I put in to Claude code, I said clean up my desktop. Clean up my desktop. And I was like, this is so insane. I'm using Anthropic's computers wherever they are on the other side of the country to clean up my own computer's desktop. Why do I even have my own computer?
- [33:10]
C
Did it do a good job?
- [33:11]
B
Yeah, of course. Yeah, it did great. And I was like, why do I even have my own computer? At this point I started wondering, I'm like, why do I even have a computer? Related to this sort of DRAM question, Could we just get into a situation in which people don't really have their own compute on their own device? Because more that like Anthropic's brain is sort of, or OpenAI's brain is sort of controlling the things we use. Why not just have a very low spec computer or low spec phone or whatever, or you know, low memory, whatever because I'm already using their computers, etc. And are we just going to see whether memory or elsewhere, this sort of migration where all of the interesting stuff and all of the memory and all of the compute happens elsewhere and my phone just becomes a sort of, you know, or my computer just becomes a. A screen with an Internet access.
- [34:07]
A
Yeah, I mean I don't, I don't know about the idea like you know, not having the personal devices.
- [34:12]
B
Yeah, I might like have a physical device, but it just sort of seems similar like over time. Does it make sense for me to have all of this actual compute inside my house when I'm just like, I might as well let them have it all.
- [34:26]
A
Yeah, I would think it depends on like what's your emperor's purpose? Right. You know, for example, laptop, for example. Right. Like if you are Doing like video editing. Mermaid heavy, right? You probably still a lot of that. You probably still need quite a lofty dram, right? If you're just doing like you know, document work, right? How code depends on how you're going to use it. If you are using like super intensively, I think you still need quite a good dram. Especially you are pulling all different API from different places, right? So I will say it just really depends on the end purpose. I don't think, at least I haven't seen there will be kind of structural sort of theoretical destruction in the devices for agenting AI.
- [35:01]
C
So going back to the beginning of this conversation and the cyclicality of the industry, people have been throwing around the term super cycle. Again, very commodities esque. Is this just a super cycle, you know, maybe bigger than previous turns that we've seen throughout history? Or do you think something structural has shifted here perhaps given the rise of AI and the fact that Joe is, you know, using Claude to clean up his desktop.
- [35:32]
A
This is actually very dangerous question to Rio. You are basically asking, right, whether this time is different, right?
- [35:38]
C
Yeah, right.
- [35:40]
A
I think like, you know, I think there's a couple of things that actually runs with the history, right? Non migration, fake capacity, those things. But I would say there's kind of difference I'm seeing, right? Because we rarely see in a sort of super cycle that there's a new demand driver coming online. It's not only constrained the demand, this thing is also constrained the supply, right? Because for example in 2010-2012 there's a super cycle driven by mobile, right? So mobile is basically, hey, we have a new product coming online and it requires a lot of theorem, let's. And our capacity couldn't catch up. But right now for hbn it's like hey, okay, we need to do hbn. But your HBN is actually constrained, your commodity derailing capacity. So I think this is the key difference that we are seeing for this, right? Apparently the demand accelerations is also last quite long, right? If you're really starting counting the sort of, I would say the AI acceleration in terms of demand, probably let's say second half of 2023, when Nvidia really start kind of emerging in people's attention to now, it's almost probably like let's say three years, right? Two year and a half, two weeks, three years. And we are looking at this cycle to last until let's say safely to say second half of 27, right? So we are looking at kind of four year cycle, that's quite rare. In the history. Usually the cycle if you look at historically for the memory is probably like 15 months, probably longest, like 18 months from start to its peak. But right now we are going through an area that was like hey, the demand was going up, not a price directly, but the demand at least was growing quite significantly over the past few years. And the pricing impact is happening right now starting from probably Q2, 2025.
- [37:18]
B
Well, when you say okay, the cycle could end by 2027 or the second half of 2027, what are the ingredients for it to slow down? Is it the more production capacity, is it a slowdown in total demand? Like what, what do you see potentially happening in 2027 that brings supply and demand more into balance?
- [37:39]
A
Yeah, so I won't say like it will end, I will say like it's a more tricky area to sort of estimate because you know, based on the number you know, I'm modeling, we are seeing here the demand is actually going higher and the shortage is actually going to worse. And the power of that is because a media AI server, their drain content was going to 3x compared to regular. Right. And then.
- [38:01]
B
Sorry, what kind of server? On Nvidia server.
- [38:05]
A
Lubin 200. Yeah, okay, sorry, Lubing. Sorry, Lubing outside Lubin 300. Okay. And if that's the case and the demand was continued to be that strong. Right. Your HBN shortage should even be worse compared to 26 and your DM should go to go worse. Right. Because the memory maker is going to try to make more HBN wafer.
- [38:25]
B
Right.
- [38:25]
A
And that kind of counting out more on commodity dram. I think the reason I say it's more tricky because two things I mentioned about a lot of non migration, it's actually happening this year. It's also going to happen in 2027. Right. So tons of non migration going to come in online and being completed. So then we'll add a lot of additional bits for the memory supplier. Additionally the wafer should have a lot more coming online by the end of 26 and for throughout 2027. Right. So I think those are two variables that you want to factor in. From what I'm seeing at least I think like 2027 theorem going to still be shortage.
- [39:02]
C
Yeah, we need a strategic memory chip preserve.
- [39:07]
A
Right.
- [39:07]
C
Put a floor on prices and the cycle. Actually on that note, do you see a lot of stockpiling in the market or panic buying right now where people are seeing these forecasts and they're freaking out and just buying whatever they can?
- [39:20]
B
How you get a bullwhip effect.
- [39:22]
C
That's right, yeah.
- [39:24]
A
It's hard to say no. Right. Like one of the good. I think the signal you have seen in Tinic, Samsung and Micron, the Dura inventory was dropping every single quarter since I think Q3, 20, 25. And that's just one other signal that you are not just buying the sort of products on the shelf. Right. And then the products are coming online in that quarter, you're also buying the, the products that in their inventory. You are trying to get as much of the product as you can. And given all the demand and how AI has developed so fast by monthly basis it's hard to say. Hey, there's no preemptive purchasing behavior happening. Especially when hyperscalers and the leading AI labs they are competing with each other very intensively. Securing capacity on the hardware is sort of the baseline. You want to get the best equipment that you can compete.
- [40:19]
B
So obviously if I'm buying a Nintendo Switch or if I'm buying a PC or maybe some sort of consumer device then the increased memory cost is a meaningful driver of that. And maybe you know, the company will have to raise prices, maybe I won't buy it or whatever and you might get this. What about like for the big hyperscalers when we see these huge capital expenditure numbers from the likes of a Microsoft or an Amazon, do these prices changes move the dial at these levels in terms of do they affect anything when we're talking about these buyers?
- [40:53]
A
Yeah, I don't think you know number one for sure, I think the CapEx is not really. Well the CapEx increase is not because of the DRAM price increase.
- [41:01]
B
Sure, yeah.
- [41:03]
A
Price increase does gonna have some kinds of impacts to their purchases of memory. Right. Assuming hyperscalers are the direct buyer of the memory, which is not necessarily the case in different locations. So if they, assuming they are the direct buyers of the memory for sure. If your initial goal was using 100 million to buy a certain amount of memory right now probably you need to do some discount there for sure. Especially this thing, the discount going to be more and more in the coming quarter. So one of the things they'll try and hopefully they are trying to do but it's still very difficult. It's like they try to have a long term agreement, right. To commit large value for a year basis and to hopefully to have better pricing. But it's really hard to achieve that because why not memory supply? I want to do that. Right. We can negotiate the pricing in a quarterly basis. We can actually get more money especially given the current Pricing and supply demand environment.
- [42:01]
B
Rei Wang, thank you so much for coming on ODD Labs. Really the perfect guest. Let's stay in touch and maybe we'll Revisit it in 2027 to see if things have eased.
- [42:11]
A
Yeah, hopefully.
- [42:12]
C
All right, thanks so much.
- [42:14]
B
Take care. Take care, Ray. That was great.
- [42:16]
A
Take care. Thank you.
- [42:30]
B
Tracy. That was fun. I do feel like there's a lot of moving pieces there, but just from AI in general, the crowding out effect is such a big part of the story. When you saw those big capital expenditure plans for 2027, like these are real macro drivers, they're going to show up in cpi, et cetera, like that. And because it's like, maybe we'll get a lot of productivity gains in the future, but right now the pace of spending is so furious, it is like a, a fiscal boom.
- [42:58]
C
Yeah, I mean, stimulus. This is what happens when some of the world's biggest companies and most cash rich companies decide that this is an existential threat. Right. There's no upward limit on how much they're willing to spend in order to survive. And so you could just throw money at HBM chips, I guess, and then you get the chip makers going. Well, actually like we want to produce a bunch of hbm, like forget about dram, all that stuff.
- [43:26]
B
I do think like, yeah, you know, when it comes to some of the data center and energy stuff, it's a little ambiguous how much it's affecting energy prices right now. But already the politics of that is very fraught. And then you're going to start upsetting the gamer community because they can't get act. You know, Nvidia has talked about curtailing.
- [43:46]
C
Big mistake.
- [43:47]
B
And then they're going to upset my son because he can't get a Nintendo Switch, et cetera, people are just going to start feeling it more and more. The sort of visceral reality that various resources that they thought they could get abundantly, it's like, nope, we've switched this line over to the data centers, over to AI. And it's become going to become more. This crowding out effect is going to become more and more real to people directly.
- [44:12]
C
Which is so ironic because when you think about AI, you know, it's this thing that exists in the ether. Right. But at the same time it has this huge physical impact on the real world. It's kind of funny and I guess not what a lot of people would have expected.
- [44:28]
B
No, I suppose not. The DRAM people, if you, anyone who has a terminal or just should just look up DRAM prices.
- [44:35]
C
Because this is, I thought you were going to say anyone who has a terminal should start stripping out their memory checks.
- [44:40]
B
Terminal should just unscrew the back and pull out the dram. No, but if you have, people should just look at the chart because really, here is this thing that's very sleepy. I mean, this was true commodity. You know, this was the low end in terms of what people were excited about in chips and so forth. And as Ray mentioned in the beginning, you know, costs generally were on this downward trend and so forth because growth was modest and technology continues to improve. And then it literally just looks like an L really in like the last four months just gone completely nuts. And so, yeah, kind of a fascinating spot to watch. And it's just interesting to the point that you were making how much it really is like a commodity super cycle. Yeah, a commodity cycle.
- [45:23]
C
Well, I guess in 2027, maybe we'll find out.
- [45:26]
B
We'll see if it balances out.
- [45:27]
C
Shall we leave it there?
- [45:28]
B
Let's leave it there.
- [45:28]
C
All right. This has been another episode of the Odd Thoughts podcast. I'm Tracy Alloway. You can follow me at Tracy Allaway.
- [45:34]
B
And I'm Joe Weisenthal. You can follow me at the Stalwart. Follow our guest Ray Wong. He's at Rwong07. Follow our producers Carmen Rodriguez at Carmenarmon, Dashiell Bennett at Dashbot and Kalebrooks Kalebrooks. And for more Odd Lots content, go to bloomberg.com oddlots for the daily newsletter and all of our episodes and you can chat about all of these topics 24. 7 in our Discord, Discord, GG, Oddlauds.
- [45:58]
C
And if you enjoy Oddlots, if you like it when we talk about DRAM and HBM and other acronyms of the memory chip industry, then please leave us a positive review on your favorite podcast platform. And remember, if you are a Bloomberg subscriber, you can listen to all of our episodes absolutely ad free. All you need to do is find the Bloomberg Channel on Apple podcast and follow the instructions there. Thanks for listening.
- [46:38]
A
Sam.