Episode summary
Ray Wang on How AI Is Causing DRAM Prices to Surge | Odd Lots
Odd Lots
· · Hosts: Joe Weisenthal, Tracy Alloway · Guests: Ray Wang
Summary
DRAM prices have surged because AI demand has collided with years of cautious memory investment. Ray Wang argues that the industry was set up for a shortage: capacity spending stayed constrained after the post-Covid downturn, while demand from AI training and inference accelerated much faster than expected. He says the market is now pulling in both commodity DRAM and HBM, but HBM is especially wafer-intensive and more profitable, so suppliers are shifting production toward it and leaving less supply for standard memory.
Wang describes DRAM as a classic commodity industry because products are standardized and competition mostly comes down to price and manufacturing cost per bit. That makes the business highly cyclical. The current cycle is different, he argues, because the new demand driver is AI, while the supply side is also constrained by clean-room capacity and slow technology migration to more advanced nodes.
He says the shortage is already affecting PCs, smartphones, cameras, and gaming hardware, with some companies hiking prices or trimming volumes. For now, he expects memory makers to prioritize server and AI demand, and he thinks the imbalance could last into 2027 before new capacity and migration help rebalance the market.
Key takeaways
- Ray Wang says AI is driving a memory shortage because training and inference both require much more DRAM and HBM than older workloads.
- Wang says commodity DRAM is already being crowded out because HBM is more profitable and consumes more wafer capacity per bit.
- The host says Apple, Nintendo, and other consumer-facing companies are already seeing pressure from rising memory costs.
- Wang says the current cycle may last into 2027 because clean-room limits and slow node migration will constrain new supply.
- Wang estimates Chinese memory makers are still roughly three to four years behind Korean leaders like Samsung and Hynix.
Notable quotes
“right now it was I think the use of GPT was 800 million. Right. And we are not included”
“So that's why you need a lot of memory, you know back to try to beat starts coming up”
“what's happening, you know, we see all of this like, spot price control price going out so much. Right? The margin of the commodity line right now, it's actually higher than NBM”
“this is a area that it's relatively easier to differentiate your products to get more market share, more meaningfully”
How the episode opens
Why does I need a lot of memory at all? So you just explain, you know, it's important for inference. It's important for training, whatever.
But like, why? Let's say you want to train a model, right? And then you need, you just need the pure tons of data.
What you need to do is right. Tons of data there to train your model. And then for the inferencing part, like you need one data after another to kind of compute one after another, sort of a chunk of sort.
Right. So to to do that process, how can you keep the previous data you process to keep in there and the process to another data that coming out after it? Right.
So that's why you need a lot of memory, you know back to try to beat starts coming up. People just like hey I wasn't whether I like a very simple questions people right now is sort nastier. People are doing like hey can you write of, plenty more.
How tensor can happen on people arrival example. They will give like 20 page report and they will wait like five minutes. Right.
And so if you think about a process like they do all the calculation, all the researchers, and also give you the output tokens that, that's not, and so you get from your prompt less significant way longer than the prompts that you are given. Right. And that's also the longer response that you are getting when you go back to the GPT three.
And remember, usage, we know back to GPT three was the workload, you know, the monthly daily users. But right now it was I think the use of GPT was 800 million. Right.
And we are not included. The users for the Gemini for called for AI.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai memory demand
- dram price surge
- hbm vs commodity dram
- supply constraints
- semiconductor capex
Companies and tickers mentioned
- SemiAnalysis
- Samsung
- SK Hynix
Watch the episode
Watch “Ray Wang on How AI Is Causing DRAM Prices to Surge | Odd Lots” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from Odd Lots
-
How a Chicago Restaurant Has Stayed Open For Nearly Three Decades | Odd Lots -
How Airlines Actually Hedge Higher Fuel Prices | Odd Lots -
Robert Friedland on the World's Monumental Shortage of Copper | Odd Lots -
Brad Setser on the US's Unusual Japanese Yen Intervention | Odd Lots -
The Tungsten Market Is Warning of an Upcoming War | Odd Lots -
These Are the Sharps Actually Making Money on Prediction Markets | Odd Lots
Odd Lots, this episode and its artwork belong to Bloomberg. Octagon is not affiliated with the show; this page is an independent summary.
