Memory has grown to nearly two-thirds of AI chip component costs | Hacker News Reader

Memory has grown to nearly two-thirds of AI chip component costs

epoch.ai

445 pointsintelkishan499 comments

An interesting implication of this is that AI inference and training has a path to a ~3x hardware cost reduction (and maybe ~2x total cost reduction) without any technical innovation whatsoever, we just need to wait for dram supply to meet demand (either by manufacturing scaling or just waiting for the current rate of manufacturing to fill the demand spike).

The memory makers will not expand demand drastically. It is in the nature of their business to keep the market under-supplied, otherwise the following oversupply will kill them. Instead, supply is just rerouted from less profitable segments such as mobile and personal computing.

China is about to flood the market and prove this notion wrong. If there is demand they want to meet it with supply.

But to your point, that is exactly how American companies like to play now. No one is stopping them from screwing over the consumer.

I have a Micron near me and they are building another chip facility but we are years away still so I suspect China will beat them to the punch.

Not just DRAM market, but the GPU market soon. China is the great equalizer of the world.

Yeah, more global competition in DRAM would be great.

SK Hynix and Samsung are South Korean.

China wants a sovereign DRAM capacity. They're playing an entirely different game then the commercial suppliers in the West.

Help us Xi Jinping, you're our only hope.

China can afford and has the political will and power to centrally plan parts of the economy it feels like planning. Cars are obvious examples and if dram is next, western manufacturers should brace for impact.

man i keep thinking. why cant india get into stuff like this. Do their own manhattan project to build factories and tech for this and immigrate experts with high salaries.

I suspect Chinese factories will get built first, but quality may take a few years to really nail down.

Basically:

China floods the market with cheaper but less QA'd parts, makes a gazillion dollars, is able to spend said money to fix yields / QA issues and streamline operations, by the time that happens Micron and maybe a few other existing players will have new memory production, and then we'll have a flood of cheap, reliable memory. 4yr, maybe?

>China is about to flood the market and prove this notion wrong.

China is very far away from flooding the DRAM market.

China does not have sufficiently leading edge fab capabilities to fill DRAM demand.

This is wrong. It is NOT in their nature to keep the market under-supplied -- eg, Samsung, the industry's largest company, was notorious for expanding their capacity during the industry downturn to gain market share while everyone else was cutting back to minimize loss.

I'm guessing you are also probably unfamiliar with the terms like "chicken game" which refers to the cutthroat, high-stakes price wars where dominant semiconductor manufacturers intentionally overproduce and slash prices. This is literally how the industry went from dozens to just three majors today since the 80's.

You're making the point for him. Undersupply in a boom, store cash to ramp up capacity in a downturn. Presevres capital and avoids overcapacity during the turning

Sure, but the key word here is "was"

The industry is so naturally prone to oversupply that the only stable equilibrium is undersupply. Aggressive expansion kicks off a price war, which immediately undercuts the logic of the expansion.

This only changes with new entrants, which will come, especially from China. But it takes time to build fab capacity, so the medium-term modal outcome is consistent undersupply.

That works when there are dozen suppliers. Does not when there are three.

If the existing memory makers retains control of the market and don't defect from the optimal-long-term equilibrium for themselves, that's true. It just takes one player to defect for short term gains as we've seen with some past boom-and-bust cycles. Alternatively, it takes a sufficiently-resourced player with enough incentive to enter the market themselves (NVidia, Google, Amazon, the PRC government through one of many companies...)

CXMT is scaling up incredibly fast, they are on a clock (south koreans) their monopoly will end relatively soon, although I'm guessing that the AI companies will crash before that anyways.

> their monopoly will end relatively soon

Corsair DDR5 DIMM modules with CXMT RAM started appearing on Friday.

Relevant article posted on HN about this a few days ago: https://davidoks.blog/p/ai-is-killing-the-cheap-smartphone

I struggle to think of a line of business as cyclical as DRAM, maybe like certain kinds of mining would be my only thought.

The DRAM fabs have been on a roundabout for 40 years going from getting accused of price fixing and cartel behavior, to struggling to keep the lights on.

And imo it's not really their fault, it's all the lead time of advanced semiconductors, combined with the commodity dynamics of oil. And the goal is to match that supply to the demand of everything from consumer electronics to more datacenters than you can shake a stick at.

It's maddening to try and solve that, so at this point I really don't fault them for prioritizing survival.

Reminds me of how Samsung is giving out $340,000 per person bonuses. Shows you how much of a stronghold they have in market.

What you described only works if the manufacturers agree to price fix. Otherwise, in a free market, they'll race to increase their earnings by meeting the demand.

Supply and demand always balance out. There is no way manufacturers aren’t going to compete away these inflated margins, as long as they feel like this demand is sustainable.

> It is in the nature of their business to keep the market under-supplied

What?! If they did an anti competitive agreement sure. Otherwise no as each supplier is incentivized to produce more than its competitor and less than the demand, while divesting just enough to survive the oversupply risk.

Apple could always decide to build their own fab or some such thing.

If you factor in Nvidia’s profit margin due to the scarcity of the current bleeding-edge chips there is a path to a much larger cost reduction still.

There’s a lot to criticize Sam Altman for saying or popularizing culturally but I’ve come to think his “this is the worst it will ever be” is, in the long run, actually a very intriguing and underrated point.

In a decade training LLMs to the current level of sophistication, which is in my opinion rather advanced and probably has lots of additional upside just from constructing better RL training regime independently of hardware advancement, will become just as table stakes as running a database is now. I highly recommend everyone look into the Allen Institute’s projects in GitHub and HF because they have open source training materials (including an LLM from scratch off common crawl, and some quite interesting tunes of qwen) to get a taste for what will be in the near future afternoon projects or educational material. The future is going to be wild

These crazy hardware price increases will probably delay everything by at least 2-5 years. Then add at least 5-10 years for all these refinements and optimizations to permeate universally.

Until everything matures, most likely the current iteration of OpenAI and Anthropic will be long gone, along with their current business models.

I wonder if we will see an adoption of alternative floating point formats. IEEE floats are notoriously terrible at lower widths (<= 16 bits). Floating point formats such as posits do much better at 16 or 8 bits. If you could train at 16 bits per value instead of 32, and suffer a much smaller inaccuracy penalty than you would from IEEE32 to IEEE16...

This has been around for quite some time, to the point I had to read this a couple times to understand what you meant. Mighta predated LLMs even.

That's already the case with say bf16

Notoriously terrible?

Posits do a little better if your numbers are biased enough toward 1, but not much better. A 16 bit posit in a near-ideal situation matches an 18 bit IEEE float, and in a pretty wide range of situations loses to either fp16 or bf16.

Training anything at 8 bits is going to be tough, and it's hard to say if the flexible exponent is worth the precision tradeoffs.

> A 16 bit posit in a near-ideal situation matches an 18 bit IEEE float

Unsure what you mean by this... A posit16 has up to 11 bits of precision. There's no such thing as an 18 bit IEEE float.

> and in a pretty wide range of situations loses to either fp16 or bf16

Many papers have compared neural networks at 16 bits or 8 bits, and posits beat the hell out of floats and it's not even close. Which is very much expected. As they're particularly suited to this task. But also in other domains, like numerical weather simulations, where tests have shown 16-bit posits can replace 32-bit floats.

> A posit16 has up to 11 bits of precision.

Is this excluding the implied bit?

In that case a short float has 10, but if you're messing with formats you can staple on an extra bit of precision and an extra bit of exponent.

> There's no such thing as an 18 bit IEEE float.

There's a lot of custom sizes out there. But if you keep following IEEE rules then there's no special circuitry needed, just a small scaling factor.

nVidia also laid out a 19 bit format that's a superset of both fp16 and bf16.

> Many papers have compared neural networks at 16 bits or 8 bits, and posits beat the hell out of floats and it's not even close.

Can you link a paper that shows posits beating floats at different sizes?

I found a 2021 paper that compares various posits to 32 bit floats, and finds that the model quality is close for some of them. It does not compare any smaller floats.

> Which is very much expected. As they're particularly suited to this task.

Posits show their value when you need a huge exponent range and your numbers focus very closely around 1. How strongly do neural nets fit that pattern?

And how often is their advantage better than 1 or 2 bits?

If you can keep your weights within a range of 9 orders of magnitude, I expect fp16 to do just fine since it loses a bit on some numbers and gains a bit or two on other numbers.

> But also in other domains, like numerical weather simulations, where tests have shown 16-bit posits can replace 32-bit floats.

Can you link this too? I found a 2019 paper that shows them beating fp16 and falling short of fp64, but no fp32 comparison. They also noted that 16,0 posits and bf16 did badly.

They did conclude that 16 bit posits were probably good enough to beat out measurement error and be suitable for the bulk of simulation, but that same chart showed that fp16 was almost good enough. So again I wonder how many bits you'd actually need, since if you're considering rebuilding your FPUs it would be silly to exclude "float sizes that aren't powers of two".

For some reason I still haven't heard any predictions on when new fabs will come online to meet the current demand. This shouldn't be too hard to find out, since the building time of fabs is very predictable process.

The difficult question is more whether foreseeable memory demand will remain at the current level, grow even further, or shrink again.

It's very easy to find out when the new fabs come online. Try asking Claude or ChatGPT.

They say around 2028. See https://manufacturing.economictimes.indiatimes.com/news/hi-t...

Original source (paid): https://asia.nikkei.com/business/tech/semiconductors/memory-...

No new DRAM fabs are being built. That's why you don't see any predictions.

Seems unlikely. Increased demand usually causes increased investment in increasing supply.

Not necessarily in a notoriously cyclical industry where everyone has already been burned by doing exactly that multiple times

It sure looks like Sam Altman's masterful gambit to corner the memory market has had unforeseen consequences.

Is any of this actually unforeseen? Buying the vast majority of the world’s supply of something does have mostly predictable consequences.

Yes, but that's not what they are insinuating.

“Unforeseen consequences” in the same way death of the target is when someone aims a loaded gun at their head and pulls the trigger.

This line of thinking makes sense if we're talking about opex like power usage. This is capex though and we'll be financing this overpaying for a long time after the hardware has "aged out". Not really sure there is an upside to it.

Also, inference cost predictions were made before this price jump, so we really haven't started paying for it yet. Inference will not be getting cheaper.

Supply will not meet demand. What incentive do the handful of dram manufacturers have to end the party? This is what happens when legal monopolies finally win control. Dont't worry. The patents will expire in a few decades. Our grandkids will see DDR5 get cheap again. The system functions as intended.

Patents is not the issue here. Not even close.

The up-front investment of a memory fab is measured in billions, and takes years to construct and get running. The margin on the chips themselves is terrible, so without scale its not worth even trying. DDR5 is a industry standard that takes some effort to conform to, but the licence fees is a drop in the bucket to the cost of creating a fab.

The fabricators were cautious about increasing production, and slow to start planning. It takes further time to build up capacity, and if the demand drops down, they may end up producing dram at a loss when the market flips over to oversupply. The demand whiplash could kill any company that dared betting on increasing production. See the "bullwhip effect" https://en.wikipedia.org/wiki/Bullwhip_effect which has killed semiconductor fabricators before.

There is a discussion to be had about how to maintain national semiconductor production in Europe and US as a strategic industry, but historic attempts have all failed.

Billions is nothing in this market - if the market is supply constrained in the medium term then the hyperscalers will purchase their own route to manufacture (e.g. through coinvestment).

Also that's not what the bullwhip effect is - although I know what you are saying. The bullwhip scenario is about the effect of communication and batching through various layers in the supply chain, this is more similar to the cobweb effect/theory.

I have fairly simplistic view of the economics involved here. Could you explain why the ability to sell more chips wouldn't be sufficient enough incentive to increase supply?

Not the person you’re replying to, but RAM has historically been a boom-or-bust business, and companies that invest to meet demand during a boom cycle usually have that new capacity come online just in time for the bust.

If it was just variable costs and new capacity was available today they’d do it. But there are substantial fixed costs and delays to increasing capacity, and that uncertainty makes it risky.

That's such a nonsensical argument, it holds for every other business too and in this case it's just a lame excuse for monopolization. If you are that chicken and can't stomach competition you should not be in business anyway.

The current RAM manufactures were convicted of conspiracy to manipulate prices back in the 2000s or thereabout, doing so is their modus operandi, but this time the government is participating in the racket.

Bringing on new fabs takes many years and billions of dollars. You're exposing yourself to a lot of risk if you build now and find that the gold rush is over by the time your new capacity is online.

Let's imagine you're drilling oil instead. You have to spend billions of dollars over years finding and developing a new oilfield to make any profit back. And once you have it, you have to continuously spend enormous amounts of money to keep producing it, which means your effective price floor is higher than the current stable price.

Now it's 2021 and someone gets a tanker stuck in the Suez, sending the price of oil sky-high. How long does the ship have to be stuck before you spend those billions of dollars on a bet that it'll recoup before someone gets the ship out?

Although on the flipside, let's pretend it's 2017's and you are Nvidia selling GPU's for Bitcoin - maybe demand will dry up at some point? Do you stop scaling production as this might be the max of the market, or do you follow the market and increase production?

It's always easier to see the right move in hindsight!

Nvidia doesn't own fabs though, TSMC does. By 2017, ASICs for Bitcoin were well underway. Ethereum hadn't switched to PoS, and wouldn't until 2022. For that specific question, the answer is yes, because the GTX 1080 Ti is/was a monster card, and the crypto miners have a somewhat predictable demand for them, so there's some modeling you can do based on demand for the 2016 generation of cards. The question is ofc, if you're Nvidia, what are you optimizing for? Let's say, without foresight that Ethereum would move to PoS in 2022 and that AI would replace that demand, how many 1090 Ti cards do you make, how many 1070s, how many mobile 1080s, how many Titans? In order to answer that, someone at Nvidia would have to have, for better or worse, really had to have gotten into cryptocurrency in order to understand that market. Because you, as Nvidia, know how much better the 1080 will be for mining Ethereum, certain predictions can be made on demand.

Question is, without hindsight, 2022 rolls around, Ethereum moves to PoS, do you sell NVDA?

Its a lot easier to commit to spending billions of dollars in a hypothetical then reality.

Gold is an even better example.

The price has doubled but it takes 10-15 years on average just for the permitting of a new mine in the US.

Offshore drilling for oil is a 7-8 year project.

People in software I think can't but help to have a Laissez-faire bias because software is so Laissez-faire. Most other businesses are highly regulated.

What’s the lifespan/refurbishability of the capex elements like the “GPU” modules or even the DRAM soldered into them?

For lifespan, AWS is still running a ton of T4 GPUs from 2018, that power a lot of computer vision models. A ton of these will have a long life, not all ML is about frontier LLMs.

How can it be economically viable to still run them?

You can get 100x the output with the same energy use.

What demand? Can't shake the notion that it's fictive considering the amount od data centers being built and GPUs sitting in containers, where they will spend quite some time before being even integrated, even more until used...

Really wondering what this might mean for local LLMs when RAM costs plummet...

Well, no: manufacturers charge more than input price generally, here specifically, Nvidia wouldn’t lower prices because RAM went down.

2-3x is completely dwarfed by the remaining improvements in training which is still in its infancy relatively

Unless there's a new paradigm, scaling up is all they can do to improve performance. They've shrunk down all the way to 1-bit models and all the low-hanging fruit is gone. There's no way for them to get much smaller, so they have to get bigger and faster to meet expectations.

This hasn’t been true for the past 2 years

Is this based on an assumption that Opus 4.7 & co are equivalent or smaller to Opus 4.5 & co? I highly doubt the advanced models (Opus, Pro, etc) aren't biggen than the standard ones (Sonnet, Flash, etc) and fairly sure newer models are bigger than older ones.

this is just not true at all, there are massive leaps from algorithms, data, etc. every year. scale is one axis of many and you need to get them all correct.

What novel data hasn't already been used in training? What new algorithms are there? Can you post some links so we can read about them?

Probably, but at some point we're very likely to run out of significant training improvements and it's not clear that we'll see that point coming from a long way out.

Likewise it's probably dwarfed by improvements in how we make dram - continuing the roughly exponential (maybe a bit less recently) scaling of chips - but not necessarily.

The 2x from returning to previous costs is interesting because it's practically guaranteed, and it's on top of everything else. We're just currently "overpaying" (relative to the stable market price) for the manufacture of dram because of a sudden increase in demand.

my reply from the other thread fits here too:

> this is just not true at all, there are massive leaps from algorithms, data, etc. every year. scale is one axis of many and you need to get them all correct.

> either by manufacturing scaling or just waiting for the current rate of manufacturing to fill the demand spike

Or the more likely scenario that the AI bubble bursts and the hyperscalars realize they have built too many data centers.

> a path to a ~3x hardware cost reduction

Really?

How long do we have to wait until that ... cost reduction hits us?

For supply to meet demand. Depends very much on how aggressively producers scale and on how demand grows or shrinks.

Safe to say at least a year or two. It'd be shocking if it took a decade.

All the projections I've seen have said that the earliest we might see the curve flatten is 2030.

It just takes that long to get a fab up and running.

I bought 96GB of RAM a couple of years ago for ~$250. That same RAM now costs $1200!

I paid $279 for crucial 96gb DDR5 5600 MHz SO-DIMM ram October 22 of last year. Amazon has the same kit going for $1,048.90 right now.

CORSAIR Vengeance 96GB (2 x 48GB) SO-DIMM DDR5 5600 CMSX48GX5M1A5600C48

Bought an extra one by accident, paid $218.99 March 2025

Goes for $1400 now. I haven't gotten around to selling it.

Nice, you were lucky. =3

I bought 192GB of DDR3 a year ago for literally $60 ($5 a stick). It's about $22 a stick now, so more like $350 today. What on earth is _anybody_ doing with DDR3?

Demand for DDR3 is up because people who want DDR5 or DDR4 but can't afford either any more are choosing DDR3 and old DDR3-compatible systems to put it in, instead of what they really want.

All memory products use many shared resources in the supply chain, so if there is high demand in one product line, others have to raise prices to compete for the resources or stop making those lines altogether.

That is to say at least you were able to buy them at $350 today, with the current trajectory there will be no supply at all in few months.

You could set up swap space on Intel Optane media, it'll be about the same performance as DDR3 and sells for ~$1/GB on the secondary market. Though it will be a lot more power hungry than Flash, let alone DRAM - so not suitable for all uses.

there's an economic term for this: substitute good. https://en.wikipedia.org/wiki/Substitute_good

Being desperate?

I’m so mad I didn’t max out my main server when I had the chance. Used enterprise sticks were dirt cheap on eBay.

Used enterprise HDD’s also jacked up now. It’s absurd lol

People spend that much a month on restaurants

I bought a couple of used computers with 256 GB of DDR 4 (total) a year ago. The ram is worth more than I paid for the whole machines now.

Someone was selling an Epyc machine with 512GB RAM @ 500 EUR last year. I regret not buying it now ...

My main computer has 64GB. I bought that one in late 2022 or so.

Looking at the current prices, even of the same RAM, is just insane. Those companies really need to pay us compensation damage here. The whole "free market" notion does not work when you have de-facto monopolies and mega-corporations abuse average Joe and average Jane.

I just found two 4tb Samsung EVO drives - unused - while organizing my garage.

I forgot to add, I paid ~500 each, Samsung for the same drive is quoting $2k on their site (maybe a new sku). These were bought 2ish years ago. Strange things are a foot at the Circle-K.

2x16gb for $105 total April of 2025. $600 for that now. Makes no sense.

Makes prior assumptions that getting tens of gigs of ram is cheap thrown out the window. Would likely lead to super fast SSDs such as optain being way more valuable

The price of SSDs is similarly depressing.

It is one of the thing with consumer when they remember they brought it at the absolutely lowest price point when DRAM maker were bleeding money.

Those are not normal pricing. Before the pricing collapse in early 2020, 96GB DDR5 would have cost about $450 to $500. And I will need to restate again the cost of DRAM hasn't really changed much in the past 20 years. Its price just goes up and down in cycles.

So in reality it is more like going from $500 to $1300. But consumer felt it was more like going from $200 to $1300.

Crucial are already selling DRAM made by CXMT. And China are already throwing money at it. I doubt the memory bubble will burst in next 12-24 months. As in going back to money losing DRAM pricing. As they will all pivot to HBM or other money making products. But the bulk of lower end consumer DDR5 or LPDDR5 will goes to Chinese Foundry. Assuming they have figure out how to do them well. Which they have improved but are still so far away from industry leaders.

Normally memory maker will push the next DDR standard to market just to push out Chinese competitors, I am not sure it will work the same this time around. DDR5 have plenty of other usage / demands.

> Its price just goes up and down in cycles.

Historically the price has always trended downward. When I first got into computing $200 could buy you 128 MB (yes M) of ram. Really nice systems had 512 MB.

That's obviously changed over the decades as process shrinks have lead to higher memory density. We should generally expect that ram will cheaper up and until the point where process shrinks stop happening. They've definitely slowed, but they haven't stopped.

> Crucial are already selling DRAM made by CXMT.

Crucial was disestablished this year.

paid a bit more than that just for a half-decent 16 gig stick recently :)

i compensate by never paying for AI

Ramflation

yea, but people now have more money.

Everything I read seems to suggest that RAM capacity is going to grow at 20-25% a year, which just doesn't seem good enough. Even in consumer use cases, phones and laptops would benefit greatly by double the amount of RAM. And then obviously, the AI need is gigantic.

I don't see it going away. I mean, it may not grow as fast as now, but I don't see it growing away either. I get why the memory makers do not want to bankrupt themselves, but it feels like there's got to be some way to push that risk off onto model providers and other people in the ecosystem to allow us to grow ram capacity more like 50% per year.

The openai deal would be absorbed by two years of that. And it would be inefficient for the RAM makers in a competitive market to leave buyers unsold-to.

I don't actually know what the rate of growth before October was, I'm sure someone round here will though.

In theory the new futures markets for chip components would help here, since it would allow DRAM suppliers to insulate themselves from that risk.

I mean the biggest risk is Chinese CXML benefits and capturing markets that others are leaving hanging and then being able to compete and push out the others when costs start to normalize.

As for 20-25% growth not being enough, I think it's not that far off, if we assume data center build out plans hit a wall and slow down significantly, and the AI heat starts to cool off.

I don't think 20-25% may be enough in the short term but if the AI build out stops within this year, we have a massive oversupply instead of a under supply.

Looking at the history of the memory industry the biggest risk is that a firm would over produce and go bankrupt. Maybe this time is different but so far no memory chip maker has gone under because their competition increased capacity.

I might be wrong but your second point can't be true if the first one is true.

Let me explain, imagine CXML grows massive and builds a lot of fabs, so much so that it becomes the leader in multiple segments, then the market demand cools off.

Then CXML the company that invested massively has oversupply so it undercuts every other memory company.

Aka, Samsung, SK Hynix are dead, and to protect Micron now US has 10000% tariff on the supply of memory.

Imagine. Because that has happened, if you don't play the boom and bust game someone will because the market is very large during a boom, and generally the player scaling more isn't the one with margins to protect and generally has the ability to undercut others.

Asian memory chip giants were made by under cutting European and American companies, American companies adapted by moving manufacturing to Asia, and European ones got bought for pennies or dissolved.

Is there any indication research is being focused on reducing menory footprint of inference for frontier class models? Is the low hanging fruit already gone there?

Low hanging? how low hanging are we talking, the basic stuff is gone. Largely big challenges around quantization were solved 2 years ago, and we have just been improving from there.

But can massive gains still be made? Definitely.

The entire AI hype is based on the paper Attention is all you need, and Attention is basically loading a huge matrix of all the tokens in memory, how well you can optimize this attention layer is basically how most architectures are trying to solve for performance and memory usage.

Only one with significant gains in it is DeepSeek (or so I would like to believe because others don't make their work open for folks like me not in Big AI Labs to read). Their MLA architecture reduced KV-cache memory requirements by upto 90%, ofc that's purely architectural change.

With some quantization like Turboquant from google you could push it down to ~1/3 of that. So 96% memory savings when talking about kv-cache.

But the models are close to being saturated for quantization based memory optimizations. We will have to see some architectural changes for a significant shift now.

If they manage to make memory more efficient, they’ll just increase the context size and/or model size.

We just haven’t reached the diminishing return of gen AI capabilities yet.

Models will get more useful if you have higher context size or higher param size. Then people will just use the models even more, leading to even more memory demand.

What is the risk? Competition is good for consumers.

The risk is to the business not the consumers

According to the recent article HBM memory is 3x less efficient wafer area wise than LPDDR; but the bandwidth is more than triple.

What if its in everyone's interest to buy computers at say 1/3rd the rate and switch everything over to HBM?

the discrepancy between compute and memory has been growing for ages, perhaps a painful switch to HBM is exactly what we need?

Would you rather have 3 intermediate computers with low memory bandwidth, or wait a little longer statistically so that we can all enjoy a new computer at 1/3rd the rate but much higher bandwidth than the area ratio?

These are fundamentally different points in design space though, hbm doesn’t have a 10mw idle draw like lpddr does.

I hear people are doing AI workloads on apple hardware, which is LPDDR but with a wider memory bus (1024bit). This requires the SoC to support this; from what I understand not many of any beyond Apple offer this. A wider memory bus may be all we need.

Can’t put HBM in smartphones and laptops. The power drain is too great.

Not many workloads are RAM bandwidth limited. Power and latency are much more common bottlenecks, and HBM loses on both of those.

Isn’t memory bandwidth super relevant for AI?

It is like the most important performance figure. When I use an LLM that mostly fits on my GPU, the GPU will run at about 30% of its maximum power consumption - probably because the memory can't feed the ALUs fast enough. Similarly for the part that runs on the CPU, the CPU cores will show 100% utilization but not consume as much power as they usually do under full load. The GUI will also be choppier than usual under full load (noticeable, but not too annoying) presumably because pixel pushing also needs some nontrivial memory bandwidth which is hard to get.

Yes and so we use HBM for AI (among other things), but that's an exception. For things like games or displaying webpages, its not very important and we generally don't put HBM into things for that.

Multicore workloads do tend to hit RAM bandwidth limits before they hit power constraints. If you do the math, running at max frequency and core utilization would usually imply you could only access a byte or so per core clock cycle. Perhaps a mere handful of bytes for the highest-performance systems with in-package RAM.

What percent of the time do you think the average consumer computing device spends fully clocked up, let alone fully saturated on every core?

I'm not moving past my DDR4 build (and the 32 GB of DDR4 2133 MHz backup chips I still have around from way back, before I got the current 3200 MHz ones) until the prices go back to being at least partially sane. This also means that CPU manufacturers are not getting my money (since the 5800X is fine for now) and I have no reason to get a new GPU either (though admittedly the B580 isn't perfect).

Memory manufactures don't want your money anymore, Micron just left consumer market 6 months ago and says we want to be B2B from now on, and who can blame him? https://investors.micron.com/news-releases/news-release-deta...

What if this is the lowest that prices will ever be?

As Yogi Berra famously said, "It's tough to make predictions, especially about the future." But based on historical tech industry trends, a price increase in one component that's this rapid and extreme, is likely to eventually regress somewhat toward the long-term trend line - even if that trend line experiences a longer-term shift upward.

As always, some interpret certain recent events as reason to conclude "but this time it's different." Occasionally they are correct. But that doesn't change the fact that it's reasonable to assume some of the recent extreme, rapid price inflation is due to shorter term market distortion. It's also pretty clear that some of the recent increase in demand represents a stable increase in the long-term trendline. The question is how much is long-term stable and how much is short-term distortion.

Then I will make my build last as long as it can, in protest of that. I do expect at least a performative price drop in the coming years, though.

Then I better divert all of my investment into memory maker stocks.

Awful time for gamers and PC hobbyists not fully into AI.

This is 100% going to kill the home built pc market. When I started building gaming pcs, the top top card was 750$ (NZD). Now they’re 10,000 just for the gpu and another 1-2000 for ram.

People used to get into gaming pcs as an affordable hobby, now it’s making general aviation look like plan B.

This has already happened. Home PC market is practically dead already due to memory, ssd and graphics card price inflation. Makers of components like PC cases and power supplies etc. are seeing demand down 30-40% year over year and this is going to put many suppliers out of business. NVDIA has stopped even listing gaming revenue on their earnings reports. Both NVDIA and AMD are not seriously interested in supplying the consumer GPU market anymore either.

The only hope left is really Apple, but even apple has conspicuously delayed the launch of M5-gen mac minis and mac studio. Mostly because even Apple can't source enough DRAM to fully supply all their product lines.

Indeed, Gamers Nexus is doing interviews with PC component manufacturers, and some are hurting bad right now. The PC market is no longer in competition, but rather survival mode. =3

https://www.youtube.com/@GamersNexus/videos

there's much more than triple A video-games running at 240 Hz on Ultra settings... a 200 USD laptop/computer has enough power to run hundreds of interesting indie games and AAA from the past

My 2019 gaming PC is considered unusable ewaste by most pc gamers. The RX5700 XT GPU is super cheap second hand right now and I've been able to play every game I want including new releases like Kingdom Come Deliverance II on great settings with no noticeable issues.

You don't even have to drop down to old indie games. You just have to turn off the FPS counter and stop pixel peeping screenshots.

Yeah sure, but some folks were in it for the hot rods too.

Were in for the hot rods.

You can still play fantastic games with amazing gameplay, great storytelling, and even requiring quite a GPU. But you won't upgrade your GPU or RAM. If it gets broken, people have already gotten their money back instead of replacement (whether that is legal or not, depends on your jurisdiction, and regardless: it is happening). So the demand and adoption of say 240 Hz 4k OLED gaming is going to slow. I currently sport two 1440p IPS capable of 144 Hz, with an AMD 6700 XT, 64 GB DDR4, and a 5700X3D. I'll wait upgrading that to a 4k rig.

What I will do is buy a Nintendo Switch 2 before the price increase hits. Why? Great gameplay for kids.

It might kill the console gaming market, too. Typically consoles get cheaper over time post-release. Instead, all the latest gen consoles are getting price hikes and at least one company is potentially pushing back the next gen release (PS6). A PlayStation 5 for $900? I'll just wait and be happy with my perfectly usable Switch 1 (since the 2 is also more expensive than it should be).

I don't understand the threat to the PC market.

Prices haven't risen THAT much and are quite affordable. And if you look at the improved quality of upscalers (DLLS 4.5 for example), gaming is now more affordable than ever, despite the increased cost of components.

Of course, the 5090 prices are insane, as are for SOME memory models, but that's nothing new and represents a fairly small market share.

> When I started building gaming pcs, the top top card was 750$ (NZD)

When I started building gaming PC, the top $700 cards didn't even provide comfortable performance or graphics. Back then, you were supposed to have several of this connected SLI or somethin. And even then, it wasn't always reliable, and it resulted in stuttering, lags, and graphical artifacts (in cases when it worked). Today, even $700 graphics cards are a much better product from a user perspective than the high-end cards of that time (and that's not even taking into account that $700 cards back then were much more expensive).

> When I started building gaming PC, the top $700 cards didn't even provide comfortable performance or graphics.

When would this have been? I can not remember a time this was accurate for the games of the time, outside of a handful of meme titles like the original crysis that made bad hardware bets. Most of them fulfilled the needs of the software and hardware of the time. I'd say the biggest issue was that for a time, software and hardware were advancing so rapidly that you wouldnt get very long out of your hardware, but that's just the reality of rapid development and not the fault or failure of any specific hardware release.

> Back then, you were supposed to have several of this connected SLI or somethin.

SLI was aimed squarely at enthusiasts, not at joe-average PC gamer and it was certainly never a requirement. It existed as a halo feature for people chasing maximum performance, benchmark scores, and bragging rights.

Improved quality used to be the justification for buying new hardware at a similar price to the old hardware when it came out new. Now the 5060/70s are 4 figure cards.

As for how much the prices have actually risen, it’s not hard to see if this is true or not. If doubling of prices doesn’t raise your eyebrows, I’m not sure what will.

Yes, this will definitely renew interest in Stadia type products.

Why? Those servers still have to pay the same price for components plus a markup for the service. In theory you can serve more gamers per GPU, but these GPUs have to be physically located in your city to have a usable latency, and that means you'll have issues with peak utilization being most users gaming at the same time of day.

I just don't see the cost savings of sharing a GPU overcoming the extra expense + profit such a service would need.

The GPUs do not have to be "psychically located in your city" to have usable latency.

Of course, less latency is always better although running a traceroute between my IP and major city (Sydney) from 1,500 km equates to about 11ms latency with optimal routing. (Real life test, traceroute via an ISP Looking Glass).

> Those servers still have to pay the same price for components...

Not if Nvidia is running the service.

Seems quite possible to me that Nvidia sells to the public just enough graphics cards to keep any frisky antitrust investigators off its back and reserves the rest for GeForce NOW, its "pay monthly for limited access to a remote gaming PC" service. The cards for NOW are billed to the BU running NOW at or below cost, the few cards available to consumers and System Integrators naturally have a huge markup due to extremely constrained supply, and Nvidia uses the fact that they are the thing behind the LLM Boom to ensure that they have -what a System Integrator in 2022 would recognize as- a reasonable price for just enough RAM for the computers that NOW rents access to.

Downvoters: notice the speculative nature of the previous paragraph. I'm not claiming that this is happening right now. I'm claiming that it's quite possibly more profitable for Nvidia to bill monthly for limited remote access to computers with Nvidia graphics cards in them than it is to sell those cards at retail and to SIs.

Don’t you worry - Microsoft and Amazon will have you covered with cloud streaming.

Can’t afford a computer because they bought up all the supply? They’ll conveniently sell it back to you with a subscription!

You’ll own nothing and be happy.

It's more likely to kill the AI market. They're overbuilding capacity and most of it is going unused. The upcoming haircut is going to kill a lot of the major players.

They've intentionally crafted an unsustainable business model in an effort to get users in the front door and raise their MAUs. We've seen this story before. We should know precisely where it's headed.

When you consider how much an employee costs, AI makes a ton of sense. Lots of businesses are stacked with staff doing basic data entry / shuffling. Even if it’s 1000usd a month, AI is still a bargain.

> most of it is going unused

Sorry that “it is going unused”? From what I've read, most AI providers are capacity constrained.

I think it's the opposite. Sure in short term hobbyists are getting squeezed, but the amount of capital that they can put into pushing the edge is small compared to Fortune 500. Sooner or later hobbyists will benefit, especially if the market crashes.

I fully agree, the billion dollar question is when it will come.

If it crashes after it kills the PC we’ll be left with… nothing? Path matters as much as destination

It's impossible to kill gaming like this. Even if hardware was completely unaffordable, people would just use old stuff for longer and then upgrade after prices restore.

Why would it kill PC? There will always be hobbyists, e.g. I can't imagine pro e-sports players running on a Mac. Personally, half of the reason I moved away from Windows is Microsoft stalling/degrading Windows experience.

Price of PCs causing a collapse in demand, then mass bankruptcies of companies making PC components so supply chains get demolished and when prices come back down there’s no one left selling anything so you can’t build a pc at any price

Apple will survive, but it's like having a car with the hood welded shut and controlled from Apple headquarters. Not much fun for hobbyists.

Apple is very far from anything gaming related - tons of games just dont work. And top tier gaming is pc gaming.

also for ones fully into AI

I wonder why the hyperscalers aren't vertically integrating more and building their own fabs. Sure, a fab costs a billion dollars, but they're currently spending hundreds of billions of dollars purchasing chips from NVidia and others.

I'm not sure if they should vertically integrate, it would probably be a better idea to directly fund the expansion of capacity, much like Apple does when they scale up a new technology for iPhones.

However, that the hyperscalers and AI companies aren't doing this says a lot about their true beliefs about how much future demand AI will have.

AI companies claim they will need a ton of massive expansion, but are unwilling to take on the risk of the capital needed for that expansion.

I'm hearing a lot of sad whining from AI folks about how these chip makers are holding them back, but who actually has the money to finance the expansion easily? Chip makers have been through this game far longer, when Sam Altman went around claiming it was time for $7T of fabs the AI companies made it clear that they were willing to make ridiculous claims, eliminating credibility.

What's needed now is for them to funnel a tiny amount of their massive piles of cash into financing fabs directly.

Oracle is getting sold because of how much capex they're spending on new data centers in the middle of a high rates environment. It's not like they're stockpiling cash due to doubting AI.

> [...] better idea to directly fund the expansion of capacity [...] > > However, that the hyperscalers and AI companies aren't doing this says a lot about their true beliefs about how much future demand AI will have.

With what money? They have to spend the money they get on hardware ASAP else they are left behind.

Because fabs are about the most complex cutting edge technology out there: the "rocket science" of our day (or one of them). And merely having the money is not sufficient. It would be very easy to blow several billion dollars and end up with nothing to show for it.

Just look at how Intel has struggled to compete in recent years, and they have been in the business for decades.

Intel struggled because they bet the company that Moore's law was over back in ~2014, and instead of upgrading their fabs to EUV they sent the money back to shareholders.

They forgot Moore's main lesson: only the paranoid survive. They thought they could coast, and it nearly killed them.

A fab takes years to build even when you have the necessary know-how. If you don't it'll take some additional experimenting before you can compete with the established manufacturers. By the time you can produce a usable chip the shortage might be over.

A fab costs $15-20bn and it takes at least five years to build. Plus it requires expertise that none of these companies have.

Another guy answered it ITT. Intel did that, it’s not great because fabs are expensive and risky and it’s less risky to amortize the cost across multiple customers instead of just yourself

Fab margins are on average super thin compared to the margins of big tech companies, and come with a lot of risk because of that. It's not something they are likely to be keen to integrate.

A fab costs a billion dollars (really a lot more) and 5 years. It doesn't do anything for anyone today.

Bought a second hand Dell server a week ago. The entire rig with a 12-core CPU and 32GB DDR4 ecc RAM cost as much as I'd pay to buy 64 GB of DDR RAM alone. I hope there's an end to this absurdity soon enough otherwise the pain will affect other markets too. I read the other day that PC case sales have collapsed by more than 40%.

Poor people are already being priced out of cheap phones due to rise in RAM-related unit costs. https://www.cnet.com/tech/mobile/smartphone-sales-to-plummet...

It makes me sad for the Neo 2.0. More ram is the only thing stopping me switching to it from a Pro.

The Macbook Neo 2 (or likely just Macbook Neo A19) is likely to have 12GB of RAM, given the 12GB iPhone Pro Max.

I feel like by the time the AI bubble bursts the PC market will be irreparably damaged. Manufactures who have been making "enterprise" parts aren't going to go back to making consumer parts because there will be no market for it. And with a glut of datacenters not making any money on slop, they are going to be repurposed for saas, stuff like OnShape but for every application.

Most users don't seem to care about storing everything they generate in cloud services and this could easily be sold as an alternative to owning "expensive" desktop or laptop hardware.

They’re going to pivot to you renting desktop cloud compute instead of owning anything.

“Bubble”

I have an alternative take.

If hyperscalers are using more RAM, and that RAM is not available for consumers, it means all the heavy stuff will happen in the cloud. Why would we want both the hyperscalers and consumers to have RAM simultaneously? Consumers would want more RAM to run local models but then hyperscalers capacity will be unused.

Because RAM isn’t in PCs only. It’s in tablets, phones, laptops, DIY computers like the Raspberry, mini PCs, watches, smart TVs, game consoles, cars, routers, cameras, all smart appliances from refrigerators to washing machines, fitness trackers, printers etc. Cloud services are irrelevant to most of these categories.

A chip that produces refrigerator ram is also capable of producing hbm3 ? Don't they require retooling? Won't the same problems surface as required to establish new fabs?

They do require retooling and that's what's happening here. RAM manufacturers decided that it's way more lucrative to focus on HMB production than DDR 4/5 production. Capacity is the issue and that's capped unless you build new fabs but they won't do it because there's no guarantee that the demand will keep the same in the next years.

Memory manufactures sit on a war chest of IP. So even if someone has excess fab capacity and wants to get into memory manufacturing, they will have to fight an uphill battle of about a zillion patents.

Most memory companies have backroom deals to exchange tit-for-tat patent violations against each other.

Not sure how a new memory manufacture comes into being without getting sunk from licensing costs?

china?

Here’s the thing, what if memory manufacturers take this opportunity to collude and basically never reduce the price of memory below the current levels since it’s too hard for a new competitor to just rise up and undercut them? Everything I hear about is how hard and risky it is to spin up a new fab.

And by doing this, they ensure local LLMs never become feasible for the vast majority of people and AI companies solidify subscriptions forever.

Keeping prices at this level is precisely how one or more competitor will rise up. Making memory isn’t super hard. That’s why it is a commodity. The problem with the memory market is that up and down cycles have bankrupted the vast majority of players in the past. Now we only have 3 players left except for a few smaller ones in China.

The reason memory prices can stay high for years in this mega cycle is because the 3 players will be very cautious on overbuilding. They’d rather under build, make great profit (not maximum) and reduce the risk of going bust if this suddenly ends.

Same for TSMC in chips.

Great opportunity for Chinese companies though. This shortage is exactly what Chinese companies need to scale.

> Making memory isn’t super hard.

Then why do only 3 companies make it?

Bankruptcy risks.

When Samsung had to sell memory at a loss after COVID, no one came to save them. They buffered their memory division using profits from their other businesses. That’s how Samsung survives memory downturns.

According to some stories, this is how Samsung convinced TSMC to not enter the memory business - that you need a nation or other lines of business to prevent bankruptcies.

The market has stabilized to 3 players.

...And why do they go bankrupt?

Because it's an incredibly capital intensive process, involving billions of dollars of investment into manufacturing infrastructure.

That is to say, making memory is quite hard.

Making the memory can be much easier than predicting future demand.

Placing the bet isn't as hard as making an accurate prediction.

//Making memory isn’t super hard. That’s why it is a commodity.

These two aren't related.

Dram is a commodity because the you can replace a chip from hynix with a chip from micron, the have the same behaviour.

And a price competitive Dram isn't easy manufacture, or China would have made it already.

> up and down cycles have bankrupted the vast majority of players in the past

Exactly, so what’s the incentive for anyone to sink half a billy into building out more capacity.

The existing players get to rest on their laurels and succeed whether or not the AI bubble busts.

The incentive is that your 2 competitors will build more than you and gain market share on you if you are too conservative.

Samsung, SK Hynix, and Micron all have to balance between capex spending, making as much profit as possible, and risk of bankruptcy.

So are the new competitors currently in progress of starting up? Because it takes at least several years.

Only Chinese companies have a chance. Problem is that China can’t buy EUV machines and the most advanced memory chips now use EUV.

Heck, the US is now pressuring ASML to not sell even DUV machines to China, period.

When costs are high enough, you can recoup that, if you have an appetite for risking the downturn.

If the collude to say make the price $1000 for a component that costs them $100(including opportunity costs), then either a new company or a greedy company in the collusion can make their price secretly $900 and get massively more profit.

Right now their opportunity cost is too high.

> risky it is to spin up a new fab

You don't need a new fab. You can build memory in 20 years old fab.

Then that's a cartel and hopefully regulators will act.

They won’t.

They will. DOJ prosecuted memory makers in the late 90s and 2000s for collusion.

This boom is magnitudes higher than before. The attention will be endless.

Current DOJ is corrupt as fuck, it will not happen. Get back to reality.

That was a very different DOJ. They no longer work for us. They act as Trump's personal law firm.

Then China will come and eat their lunch. I for one will only buy Chinese RAM from now on, no matter the prices.

>I for one will only buy Chinese RAM from now on, no matter the prices.

Memory is a commodity, so I think you will be very lonely in your quest.

I really don’t want to give anyone ideas, but doesn’t this make the Nvidia 5090 an unbelievably good deal right now?

The VRAM in the 5090 is only made by one country in the world.

The 50xx series is special, because its ram is so dependent on a single commodity. It’s not like a 4090 or a 3090; their VRAM chips have been around for years.

If there’s a shortage or interruption in DDR7 VRAM, it seems like every GPU that requires it would explode in value.

I hope I don’t regret posting this because I’d really like to buy one myself…

An unbelievably good deal at $4000 plus?

Possibly the best deal there is

I really need to shut up, or bite the bullet and by one.

If you graph the tokens per second on the 5090, your jaw will hit the floor at how cheap it is

With only 32gb of vram, you can only run small/quantized models, in which case what's the point? At $4000, that gets you 20 months of 10x claude or chagpt subscriptions, which provide far better models. You'd need some use case where you can tolerate worse models, and use a steady supply of them. That doesn't match most people's usage patterns.

The 5090 is crap for inference. Unless you like dummy models, sure they will run at light speed. All the rage is MoE with 500B-1T weights nowadays.

It's gone up like 300% in cost in the last year.

Which surely is the highest it'll ever be! You're suggesting that the price will go down in the future? Would love to hear more about your thought process!

Are you saying we're entering a period where tech increases in price instead of decreases? I guess it depends upon time horizon, but your statement isn't very specific.

There was only a very brief time it was selling for MSRP (last fall for $2000). Even if you use that as the previous data point, it's only 200% increased.

> it's only 200% increased.

If it's 4k instead of 2k msrp, that's a 100% increase.

I believe msrp is $2000 right?

if you can buy one!

The RTX 5090 is faster than an H200. It just has less ram (32 vs 141), doesn't have NVLink, and technically isn't allowed to be used in a datacenter.

The datacenter GPUs sell at an 80% margin. They're incredibly overpriced. But the laws of supply and demand are undefeated and so here we all are.

> The RTX 5090 is faster than an H200. It just has less ram

H200 has HBM and much more 64-bit compute

Let me try again.

RTX 5090 has more CUDA cores that run at a higher clock speed. H200 has more RAM and significantly more RAM bandwidth.

Which one is net faster depends on your use case. But you may be very surprised that many workflows are faster on an RTX 5090!

I recently built a system at insane ddr4 prices ($2000 for 256gb). But that’s only after seeing how ddr5 prices were 3-4x that!

Yeah I upgraded all of my systems to DDR5 last year, so now I have to buy for ddr5 memory upgrades.

Had to fork over almost $1k for a 64G DDR5 kit a few weeks back. At least AMD chips large L3 cache allows folks to get away with lower grade udimms.

Also had to do an Intel build, and there was no way we were going cudimm at current prices. =3

I find it deeply ironic, that iran has blocked helium supply- while it relies on AI created slopaganda to subvert its advesary. Its one of those afterwits of history.

> iran...slopaganda

A US soldier i know commented that the iranian ai slop is "scary and powerful".

With how things are going, I'm really wondering how we are gonna tackle the consumer market for things like gaming and machine learning.

No doubt Cloud Gaming is in the cards for the future, only purists like myself with an RTX 5090 will pay premium for offline gaming

In the long run cloud gaming is inevitable, it’s just more economically efficient for the cost of the hardware required to render graphics to be amortized across consumers and not sit idle when being unused by collocating them with game assets in POPs.

Once enough gaming compute runs at the edge it also allows for more technically advanced games than would currently be economically feasible (but aren’t made mostly for lack of a market/adoption of cloud gaming and the resulting lack of technical know-how). So I think it will stick and probably end up winning over the holdouts, once the cost of rendering the games they want to play with consumer hardware becomes too large to stomach.

You could make the same economic argument for any SaaS, but the margins SaaS providers look for make it so that the only time it isn't cheaper to run your own software/hardware stack in place of SaaS is when the hardware requirements are very low, not high. SaaS makes sense economically when you take into account the admin, compliance, etc. costs... and the admin costs of a Nintendo Switch are pretty low.

Economic efficiency does not win the day because the free market is a myth. Cloud gaming is a technically worse solution because the latency floor is higher. It's a microeconomic disaster (rent vs buy, buy wins). The only reason it would become a thing is if the multinationals succeed in concentrating more wealth and power, which consumers aren't interested in supporting. It's a bad deal and consumers know it. They would have to be forced into it by having the consumer hardware market taken off the table (which is happening and the only possible avenue for a technical regression like cloud gaming to have a market).

I assume that memory manufacturers don’t really care where the money is coming from, as long as the "numbers go up" game is working.

NVIDIA in their recent quarterly report stopped categorizing "Geforce" as a single category, and merged it into "Edge-Computing".

If you are a PC Gamer or PC Enthusiast as I am, then we have some dark times ahead.

Do we though? DLSS 5 changes that somewhat from a “we need powah” to “we need models”. I think the future consumer GPU market will be tuned for image and world inference while workstation cards will be tuned for image and video inference. The old way of thinking about this will come to an end when we stop looking at the render loop as the be-all-end-all…

Or, we could be fucked.

If DLSS 5 becomes the norm it's possible that just makes things worse. The DLSS 5 demos required an entire separate card to run the model, though IIRC NVIDIA did claim it would eventually work on a single card. Given what the model is doing (yassifying the whole scene instead of just upscaling/reconstructing) it makes sense to me that it would increase compute demand instead of reduce it like previous versions of DLSS.

The demos did, but look how far we have come in just two years? Running local LLMs, running local diffusion models, running local world models (albeit, barely a scene at this point). I do believe that in 10 years time, game will be producing latents and not events they way they do now. I also hope this means that VR can finally get the fidelity it needs to really take off.

From my point of view, I suppose we will enter a "Let AI generate entertainment" era. In which you just might rent everything, including games. No need for a beefy computer at home, you just need a slim endpoint:

"Order yours now, for just $99.99 per month, hardware included! Order today, and you will get three months of 'Office Suite' for free, with a small additional cost of $49.99 after month 4. On a tight budget? Switch to the yearly subscription, and pay comfortably in 18 installments."

On your Karna card…

Why did this happen so suddenly?

Why were tech savy investors unable to figure this out when the datacenter craze had already started?

How to explain this lag between quickly rising demand for all datacenter components besides memory?

RAM is a boom-and-bust industry, so memory manufacturers were reluctant to invest. Here's a good blog post on the economics:

https://davidoks.blog/p/ai-is-killing-the-cheap-smartphone

Maybe long-term purchase agreements from big buyers might have helped convince them it's okay to build, but apparently it didn't happen.

Nine years after Google's seminal paper lit the fuse on AI, a total lack of manufacturing foresight has trapped over a trillion dollars of incoming capital in a hardware bottleneck.

The entire sector is now facing a critical RAM starvation crisis where memory manufacturers are actively slow-rolling supply just to keep prices high and avoid running out entirely.

This has created an unprecedented supply-and-demand distortion where desperate companies are getting rejected even at a 5x markup, and mission-critical SKUs are skyrocketing to 10x and 20x their baseline value.

It is a macroeconomic squeeze at a staggering scale, and the massive venture scale opportunity lies in capturing the value created by this memory gatekeeper.

From the perspective of an armchair economist, the winners will be the investors who invest in RAM wisely. The losers will likely be cash strapped SAAS companies. They’re almost completely dependent on a fleet of servers in the hyperscalers, and they’re leasing those servers and services. That leaves small SAAS companies exposed to incoming inflation in the cost of hosting.

The same reason they didn’t all sell everything to buy NVIDIA the day chatGPT came out

It's still unclear to me: the shortage is semiconductor boules / wafers? or the shortage is semiconductor fab process step availability?

As long as the discussion seems focused on memory, I'd suspect the latter, but if its really the semiconductor boules/wafers, then I'd expect the boule growers to profit, not the memory makers, who just pass on the cost.

So which is it?

It’s fab capacity. Fwiw dram is different enough that fabs are not transferable between dram memory and other usages. It’s nice to think ‘wow if they made the current 10nm dram on the latest 2nm processes it’d be much faster’ but it doesn’t work that way. The specific size is needed for the capacitance. Sram can be made on fabs that make other circuitry since it’s transistor not capacitor based but is less dense.

Dram is just extremely specialised.

There is a good article (featured on HN a couple of days ago) that explains the issue: https://davidoks.blog/p/ai-is-killing-the-cheap-smartphone

This covers it pretty well https://news.ycombinator.com/item?id=48229319, TLDR -memory for AI uses more wafers from same production line as other memory and is more profitable, building new fab very risky historically for companies. The companies have cut production of other memory to favor memory for AI and the market for memory for AI is still unfulfilled so prices still go up for customers of every type.

Regardless of the specific mechanics of the bottleneck, we know what the proximate source of the problem is: openai locking up 40% of Samsung and SK Hynix wafer capacity for the next few years. That's what triggered the madness.

I wonder if it is reasonable to assume the propagation of shortages further. At first it was GPUs, then RAM, then what?

Fresh water?

Since memory is becoming an expensive commodity, I guess the old ways of being precious on the efficient memory usage of your program (like it running on the constrained 1mb memory back then) are making a comeback.

I only feel sorrow for the electron devs, they will have a hard time.

Good time to focus on more memory efficient means of training and inference.

SeedLM from Apple is an interesting approach for inference memory efficiency. I'd like to see someone try and build that into training so that it's not a post training compression step.

It seems to me the max memory you can buy in a laptop stagnated for the past 3 years or so.

And the max storage in pre-built computers has stagnated at 2010 levels (~1TB). This was first due to the switch to the much more expensive and much faster charge trap flash. In the 2020s it finally started to approach 2010 sizes in pre-builts but then the corporate finance wars re: fab capacity happened.

I have always felt insulted that most laptops even offer a low 4 GB of RAM I rather take 16 GB in previous gen memory

My several years old laptop has 128GB of RAM, is that not enough? I admit that it's a pretty heavy one.

for the most part, unless soldered down, it has been hard to find higher than dual channel (maybe quad for a massive odm gaming laptop). each stick and platform having set maximum memory capacity has put a glass ceiling for those machines.

doesn't matter anyway when things are not reasonably priced. i am stuck at the same memory capacity in my personal system for the better part of two decades, partially due to the above and the current pricing today.

I think the companies that drive up the prices here, need to pay an extra-tax to all of us. I fail to see why I now have to pay more due to the AI monster companies ruining the economy.

Or to put it another way, the prices will only come down the other side of an intense catastrophe.

AI growth is locked in now, only if it were to stop will demand be abated.

Time to let ASML sell to the Chinese memory producers … or not.

Since January, I've been lucky and picking up various used DDR4 memory sticks for cheap-ish. I got a total of 64 GB for $180. I feel like I hit the jackpot!

As models gain efficiency, will the need for ram cool?

They’ll just fill up the ram with bigger models. Demand will INCREASE, not decrease.

Every time we add capacity with almost anything, we find ways to saturate it.

Braess's paradox for roads. When we add capacity to road networks, traffic increases even more than the capacity.

https://en.wikipedia.org/wiki/Braess%27s_paradox

Jevons paradox is at play. Right now frontier AI is very expensive which heavily suppresses demand.

If you made it 10x cheaper right now you would see a truly unimaginable wave of bot slop.

AI is choking the computing economy. Many companies will die. It's already a mass extinction event and will leave behind deserts.

And four-fiths the cost of a consumer PC build.

Built a new machine with 64GB DDR5 and 5TB SSD in January 2025. It's sheer luck that I dodged that bullet.

I wish I had figured that out a year ago. MU up ~10x, SNDK up ~37x. My crystal ball is woefully under performing.

The cost of memory should continue go up as we tend to have the AI to have context and remember lots more.

Memory makes computation universal.

it’s fun and ironic that “having a memory” is what AI appears to lack the most in practice while at the same time it demands more computer memory than anything to run

A commodity rapidly increasing in price. What could go wrong?

How can I use this information to MY advantage? Do I started going into something to do with AI chip memory-stuff? If so, how? But just on a software level cause hardware is hard.

Anyone invested in Micron stock?

Up 700% in a year.

WallstreeetBets has been disturbingly accurate in its predictions - basically anything related to AI.

unified memory architectures are getting more interesting for inference workloads.

if we survive the bubble bursting and there isn't a "too big to fail" bailout with public money manipulation by bought politicians

we are going to have amazing cheap used hardware for a decade

The algorithm advances are going to crash this so hard.

Or will more efficient algorithms just mean we run even more AI models, increasing the demand for AI chips even more?

I heard Greg Brockman on a podcast saying they are limited by computer and memory. They have line of sight in solving many different kinds of problems. But they also have to survive in the meantime. Hence the focus on enterprise recently. They could just ask Government to fund them doing other research areas

Better algos = more demand

Memory squeeze will get worse before it gets better.

I mean, god willing, but it'll be just as likely that we'll blissfully consume 100 million token contexts in that case.

isn't there a law for that? as things become cheaper you consume more?

You're probably thinking about jevons paradox. But you slightly mis-stated. It is the phenomenon that increasing the efficiency of resource consumption can end up increasing total consumption.

As you stated it, it would merely be a property of (nearly) all demand curves. Jevons paradox only happens sometimes. It isn't a law.

Jevons paradox: https://en.wikipedia.org/wiki/Jevons_paradox

Jevons paradox.

https://en.wikipedia.org/wiki/Jevons_paradox

jevons paradox

classic uneducated algo copium talk

Reply on news.ycombinator.com