I think the epic mistake was made by investors and executives who bought tales about imminent super-intelligence about to displace millions of knowledge workers from entire industries replacing them with AI datacentres.
These people are the main reason why AI companies have unlimited funding, and can afford to buy global RAM supply for years in the future despite their expenses exceed revenue by billions.
antitrust law can be a great tool if the government actually uses it for enforcement, which it hasn’t been because it causes parties to lose elections.
How does buying long term put options require perfect timing? The whole point of a long term put option is that you only have to be right at some point between when you buy it and when it expires.
No, put options suffer from time decay and IV crush. A $1,000 MU put for December 2027 was at about $200 today. So if Micron dropped tomorrow, you'd probably start making a bit of money, although the delta is only -0.32. But if you held all the way to December 2027, Micron would then have to drop all the way to $800 before your position is profitable.
When and how that transition happens is subject to a number of complex factors, and it's not even necessarily the case that incremental drops in the stock will produce incremental gains for your put option.
This is a really bizare argument for anyone who actually knows about options and trades them. If your thesis is that RAM is in a massive bubble and Micron is going to crash when it bursts, you don't express that thesis by buying a put struck around Micron's current bubble price. The fact that you chose a $1000 strike as your example is weird because that's basically the most expensive way to make the argument you're supposedly making.
"IV crush" is an especially strange objection in this context. IV crush matters when you buy options at elevated implied volatility and that volatility collapses. If Micron suddenly drops hundreds of dollars because the alleged bubble is bursting then the implied volatility would sharply rise, which makes your put more valuable, not less. Invoking "IV crush" here mostly makes it sound like you've heard the terminology without thinking through how it actually applies to the scenario you're describing.
If you genuinely think Micron is going to collapse sometime over the next two or three years because this entire RAM shortage is an overhyped bubble, then the obvious trade is to buy puts around where you think the stock should return to once that bubble disappears. Micron wasn't remotely a $1000 stock before this run. We can be generous and use a $300 strike since even though that's still 100% higher than Micron's price prior to this run-up, it gets the point across.
A long dated $300 put is currently around $7 per share, so one contract costs roughly $700. If Micron eventually falls to $200, that contract is worth $10000 at expiry. At $100, it's worth $20000. If the crash happens well before expiry, it can be worth even more than its intrinsic value because there's still time value left.
If you're claiming to be certain that a gigantic bubble is going to burst and wipe hundreds of dollars off the stock price, there are long dated far out of the money puts specifically capable of expressing that position. Pointing at an expensive $1000 strike put and saying "look, options are complicated" is just a weird or rather superficial misunderstanding of some financial concepts.
I personally use options for hedging, which I think is generally the responsible purpose of them.
It's true that you can get something like a directional bet by going far enough out of the money with strong enough conviction never to exit the position early. But this also exposes you to a lot of risk that it might not pay off even though the original idea was correct. Microsoft crashed in 2000, but it never returned to its pre-1998 price, because there was some real value in the dotcom bubble and they were able to capture a disproportionate fraction of it.
None of that helps you get memory now. You're not going to see your payoff for 2-3 years or whenever the bubble bursts. At which point the bubble has burst and you can simply buy ram at a normal price again. And the thesis of this discussion is that we can't buy memory at a reasonable price now, not 3 years from now.
This isn't an abstract discussion about the usefulness of options trading or other hedging strategies. You want to pay for ram today by betting on returns years in the future? Risky play! Hope you can stay solvent.
The point of LEAPS is you don’t have to perfect the timing. You buy far enough out to avoid theta decay, and far enough out of the money to minimize risk.
Regardless of timing, for shorts to pay out requires the market to actually correct itself. You won't be able to get your magical shorts money until the price of ram goes back down anyway. The market will remain irrational longer than you can remain solvent.
The price of RAM does not need to come down in order for a way-out-of-the-money January 2027 put on NVDA to increase in value from its current purchase price.
Google (which is diversified more and whose performance is less likely to track AI outcomes closely) aside, the two other members of the AI Big 3 are non-public, so, no, you really can’t.
You can make a lot of money with shorts if you are correct about when the market will move. If you think that the market is overestimating the real value you have not even half of the puzzle.
The same thing was said 4 years ago about NVIDIA on HN, that it's stock it's outrageously overpriced, given it's $20 bln revenue, that it should have at least 10 times more revenue to justify that stock price, which is fantasy, that there is no plausible way for such demand no matter what you think about GPT-2.
nVidia earning is way above zero by a healthy margin. This was also true 4 years ago. How much their shares should cost is debatable but still, nVidia is obviously a profitable business.
Anthropic net loss in 2025 was $42 bln, OpenAI $38.5 bln. Both are spending enormous amounts with no obvious path to profitability.
I'd come at this from a different angle: we still want this to be market system, so we need to make this priced into the market. How can we price this in?
I would try to solve this by making the market structure reflect the underlying difficulty: we have to decide what capacity to produce years in advance, to construct the memory fabs. So this should be a futures market, and a capacity crunch would affect short-term-futures, but leave full term futures at the same price. Because the companies supplying the memory can just construct more capacity to fill those futures at the same cost regardless of the AI demand.
ah yes, adding a futures market to a sector heavily invested in AI certainly won't lead to catastrophic over-speculation that will collapse the industry entire
> How should the memory companies have acted differently?
Same way nVidia did through the crypto insanity - make sure they're supplying enough to the consumer market so it doesn't get completely destroyed and pulls down the other parts of the consumer market they're reliant for long term success.
> Should we blame this on memory companies or the AI companies bidding for memory?
Blame doesn't change the outcomes, neither does it improve the negative consequences. Think in terms of "what does destruction of our consumer market mean for my prosperity?" not "oh, how do defend poor companies again?"
For one, they could have could have not massively scaled back consumer memory manufacturing as a matter of duty to customers. But money and greed must prevail.
Isn't it the case that public companies owe a primary legal/fiduciary duty to shareholders over customers?
The question of choice between a profitable vs unprofitable venture is easy. But yeah, to what extent is choosing profitable instead of VERY profitable a breach of duty to shareholders?
As far as I am aware, the answer to your question is ‘no’[0].
> Contrary to what many believe, U.S. corporate law does not impose any enforceable legal duty on corporate directors or executives of public corporations to maximize profits or share price. The economic case for shareholder-value maximization similarly rests on incorrect factual claims about the structure of corporations, including the mistaken claims that shareholders “own” corporations, that they have the only residual claim on the firm’s profits, and that they are principals who hire and control directors to act as their agents.
This started long ago, and is one of the best examples of “if you repeat a lie often enough eventually it’ll be widely believed”. It’s a useful tool to manipulate public opinion over a few generations. But no one would ever do such a thing, right?
Mm not really. This has been argued so many times on HN. It’s imbued in de facto case law. You’re not going to find it in legislation.
Willingly getting sued by hostile shareholders is in itself not acting in the best interest of shareholders so while it’s not in code, it’s buried in many decades of case law and in reality it reflects the maligned incentives for companies.
>Isn't it the case that public companies owe a primary legal/fiduciary duty to shareholders
The shareholders own the company. It is their property. They paid for it, they own it, and likewise they can do whatever they want with it.
It would be crazy if I proposed you let me drive your car to work everyday. Why? Because you own your car, it's your property, and it works for your own interest. Basically every human agrees with this logic, but somehow "the company is just focused on pleasing shareholders" escapes this.
Shareholders functionally own the company. Yeah you can't stroll inside and take a ream of copy paper because you own a few shares, but the company also cannot be sold (ownership transfered) without you getting paid for your slice.
The shareholders, via the board, hold final say over the company, it's direction, and it's alignment...because they own it.
This is like asking how many times can a politician can break his word before people vote for the other guy.
It's more like voting for public officials. Shareholders can vote to fire a CEO if they feel he's not acting in their best interests no mattter if that's the case or not.
Raid your own assets, cut costs manically so you can have a few splendid quarters with stock buyback bonanza or focus on long term value creation, which a lot of times involves giving at least a passing tought to other stakeholders such as client and employees?
The getting was great for some time for HP shareholders under Fiorina, or GE shareholders under Welch. Lots of them left the sinking ships at the right moment, but I bet that lot of the others left holding the bag, would have preferred having bought AAPL.
The overwhelming majority of incentives for executives is and always has been aligned with squeezing more value out of customers, in the form of increased margins, lowest possible costs with the highest prices the market will tolerate.
There are a few, I mean very few executives who stand firm on defending good faith toward customer happiness and quality. One of the only things stopping the spiral is competition. That’s why there’s an incentive to consolidate into a few massive conglomerates.
An aside, this is a poorly-structured article. It's mostly screenshots of citations from other articles with no alt or any other way to actually search through what they're citing.
People have the option of just not buying stuff. I sold my Tesla in 2021 for a stupid price, drove a beater for two years, and bought another new car a couple years later. Now, certainly not everyone has that luxury, but this is computers, not real estate. Prices will crash.
Unless you know, your main and only PC/notebook just died and need a replacement ASAP, they you don't really have an option to not buy. And no, the used market (where I live) is just as fucked, proportionally reflecting the retail gouging, with people wanting quite a lot of cash for tower PCs and notebooks from the Pentium 4 - Core 2 Duo era.
> Prices will crash.
Does your crystal ball also say when exactly? If my laptop broke, then I need a replacement now in the overpriced market, not in X years when prices will crash.
If there were such a crystal ball, the same market doing the gouging would be pricing it in already ;) does suck that "wait and see" is the optimal answer to the average consumer but I guess that's the capitalist hellscape we have wrought
I think it's more likely the industry would rather rent us our compute and storage than let us own it ourselves.
We can have our Raspberry Pis and our ESP32s, but meaningful amounts of compute and especially storage will be rental-only.
Requiring proof-of-humanity using device-bound cryptographic attestations (which the "age verification" camp will rally behind) will insure the toy computers we are allowed to own won't be able to conduct meaningful online discourse. Only phones and other locked-down platforms will be trusted to make the attestations necessary to interact with banks, merchants, and governments.
Edit (now that I'm not on mobile):
I think storage ownership, more than RAM or compute, is the cornerstone of digital sovereignty. Big tech will be happy to let us store our data on their servers for a fraction of the cost of self-hosting, so long as we let them train their models on our data. Any storage, be it self-hosted or rented, where the data is inaccessible for model training, mass surveillance, or censorship will command top dollar.
It'll be just like "dumb TVs" going extinct. "Dumb storage" will be impossible to get for any reasonable cost for consumers.
They're coming for business computing too, BTW, but consumers are the low-hanging fruit.
I hope so. I feel so, so, so terribly guilty for not building my wife a computer for her in 2023. She is rocking a 1080ti, and I really wish I had at least upgraded her system, but now I just straight cannot afford to even consider it, given the price of rent, food, gas, energy, all on top of the actual hardware prices.
And I am a SWE.I am not making bad money. I can't imagine what others are going through.
I was so lucky to buy a high-end miniPC to replace my home PC for about $1,500 USD in 2025. The same model now sells for more than $8,000 USD, without disk or RAM. Triple that if you match my original disk and RAM buy. Everything got incredibly expensive, not only storage and RAM.
Way to miss the whole point of the article and fail to understand the point. Do you think if significant part of US industry goes bankrupt because they can't source parts for their products, work machines and consumers disappear... you'll still have luxury of running around in a Tesla?
The AI companies must be regulated. They are already destroying the world and the economy before they can even make an attempt at destroying humanity.
Either AI will get cheap enough and blow up and destroy the economy by eating everyone's lunch and killing off sectors left and right. Or it will become an uncontrolable weapon destroying the internet and killing off people on the battlefield. Or it will tank the economy because there may not be a way to monopolize it making the billions of investments moot leading them to their only leverage of hoarding up all the resources pricing out the competition but also destroying the industry. And maybe all of those at once.
Yes, but zoom out more. Wealth inequality and the k-shaped economy are the real culprit.
When a few people have so much wealth that they are no longer price-sensitive, they bid up the price of everything and anything of value be it stocks, real estate, computer hardware, fine arts, sports teams, etc.
The result is that a business which tries to make quality products at reasonable prices will fail. They aren’t luxurious enough for the people with money. They aren’t cheap enough for the people without. Customers in the middle hardly exist. Any valuable good that is genuinely scarce will inevitably become a luxury.
This comes at a perfect time. The public doesn't need computers, they can just use their phones.
If the consumer had access to this RAM, they might all just run local or semi-local AI. It's important to outbid them so you can rent AI to them, and extract money from them in a million other ways while they use it.
Make RAM half the price it was before and you'd still have very few people trying to do local AI on it because it's just so damn slow for the task. Having 12 channels of DDR5 in an Epyc system with NVMe storage is still one of the slowest ways I can run LLMs locally.
The M4 (soldered wide LPDDR5X + unified architecture + GPU with neural cores) is also closer in design to a dedicated GPU than an Epyc server with DDR5 or the components in the article.
But yes, a dedicated consumer GPU can still typically outpace the M4 quite well (if things fit in VRAM) and both will have their socks blown clean off by a hyperscaler GPU cluster. For some hard numbers I can run a ~24 GB model on my 5090 (about as fast as you can get on a single consumer class device) about 3x-4x faster than on my M4 mini and I'd still consider that pretty slow to running models 10x the size in the cloud.
nvidia investing in every aspect of the hype cycle is terrifying. as an ai skeptic i almost don’t want the bubble to burst anymore, i don’t know if the economy can recover at all. perhaps a new economic paradigm will emerge, but if so, the transition is going to be rough.
The economy always recovers because the value in society is create by people existing and wanting things. The people won't stop existing or wanting things and eventually they just will believe something new is valuable and the economy will recover and grow. Because nothing material was lost just numbers in the spreadsheet. The recovery will be rough though your right about that.
2/ where is your capital and your means of production? Your machines? Your factories? Your raw materials, your tools? How long do you survive tomorrow when you quit your job and you have no other income?
You're not a capitalist. You're just a consumer for them to suck dry. The faster you learn it, the better off you will be.
You are right. People are misusing "capitalist" to mean someone who favours a capitalist system, where as it actually means someone who posses capital; someone whose relies on possession of assets rather than earned income.
> We are all Capitalists, until the Market comes after the stuff we love.
Anyone with half a brain isn't a "Capitalist" (with a big C, like in your example) which is why every single prosperous nation (including USA) regulates markets to avoid destruction of their prosperity and society.
Colloquial "capitalism" today is just cronyism that in no way resembles the capitalism that Adam Smith wrote about. We really should move towards Rhine Capitalism, the only system that has stood the test of time and truly benefits everyone.
He only used that metaphor a few times, and he did not attribute the sort of mystical power to the metaphor that modern neoliberal economists do. He never proposed the modern extreme interpretation of the concept, nor did he suggest that government should release all control over the economy.
Marx treated Smith rather favorably, and credits him for first understanding that labor in general is what creates value. Smith is best understood as an early pioneer in understanding political economy and trade. He likely wouldn’t recognize or understand the capitalism of today (nor would Marx).
Should we blame this on memory companies or the AI companies bidding for memory? How should the memory companies have acted differently?
I think the epic mistake was made by investors and executives who bought tales about imminent super-intelligence about to displace millions of knowledge workers from entire industries replacing them with AI datacentres.
These people are the main reason why AI companies have unlimited funding, and can afford to buy global RAM supply for years in the future despite their expenses exceed revenue by billions.
The main tool to fix this is to forbid companies to sell below cost price. It is part of antitrust law.
antitrust law can be a great tool if the government actually uses it for enforcement, which it hasn’t been because it causes parties to lose elections.
You can short them and make a lot of money (to buy RAM with?) if you feel this way and are correct.
Not having the money to buy a new computer often means not having the money to put where ones mouth is.
Can't afford thing? Gamble on stock market instead.
Ridiculous argument that already rich people make.
No?
Shorting doesn’t only require you to be right. It requires perfectly timing when the market will realize you’re right.
You can buy long term put options.
That still requires perfect timing. Getting the timing right on a long-term contract is even harder than with a short-term contract!
How does buying long term put options require perfect timing? The whole point of a long term put option is that you only have to be right at some point between when you buy it and when it expires.
No, put options suffer from time decay and IV crush. A $1,000 MU put for December 2027 was at about $200 today. So if Micron dropped tomorrow, you'd probably start making a bit of money, although the delta is only -0.32. But if you held all the way to December 2027, Micron would then have to drop all the way to $800 before your position is profitable.
When and how that transition happens is subject to a number of complex factors, and it's not even necessarily the case that incremental drops in the stock will produce incremental gains for your put option.
This is a really bizare argument for anyone who actually knows about options and trades them. If your thesis is that RAM is in a massive bubble and Micron is going to crash when it bursts, you don't express that thesis by buying a put struck around Micron's current bubble price. The fact that you chose a $1000 strike as your example is weird because that's basically the most expensive way to make the argument you're supposedly making.
"IV crush" is an especially strange objection in this context. IV crush matters when you buy options at elevated implied volatility and that volatility collapses. If Micron suddenly drops hundreds of dollars because the alleged bubble is bursting then the implied volatility would sharply rise, which makes your put more valuable, not less. Invoking "IV crush" here mostly makes it sound like you've heard the terminology without thinking through how it actually applies to the scenario you're describing.
If you genuinely think Micron is going to collapse sometime over the next two or three years because this entire RAM shortage is an overhyped bubble, then the obvious trade is to buy puts around where you think the stock should return to once that bubble disappears. Micron wasn't remotely a $1000 stock before this run. We can be generous and use a $300 strike since even though that's still 100% higher than Micron's price prior to this run-up, it gets the point across.
A long dated $300 put is currently around $7 per share, so one contract costs roughly $700. If Micron eventually falls to $200, that contract is worth $10000 at expiry. At $100, it's worth $20000. If the crash happens well before expiry, it can be worth even more than its intrinsic value because there's still time value left.
If you're claiming to be certain that a gigantic bubble is going to burst and wipe hundreds of dollars off the stock price, there are long dated far out of the money puts specifically capable of expressing that position. Pointing at an expensive $1000 strike put and saying "look, options are complicated" is just a weird or rather superficial misunderstanding of some financial concepts.
I personally use options for hedging, which I think is generally the responsible purpose of them.
It's true that you can get something like a directional bet by going far enough out of the money with strong enough conviction never to exit the position early. But this also exposes you to a lot of risk that it might not pay off even though the original idea was correct. Microsoft crashed in 2000, but it never returned to its pre-1998 price, because there was some real value in the dotcom bubble and they were able to capture a disproportionate fraction of it.
None of that helps you get memory now. You're not going to see your payoff for 2-3 years or whenever the bubble bursts. At which point the bubble has burst and you can simply buy ram at a normal price again. And the thesis of this discussion is that we can't buy memory at a reasonable price now, not 3 years from now.
This isn't an abstract discussion about the usefulness of options trading or other hedging strategies. You want to pay for ram today by betting on returns years in the future? Risky play! Hope you can stay solvent.
The point of LEAPS is you don’t have to perfect the timing. You buy far enough out to avoid theta decay, and far enough out of the money to minimize risk.
Regardless of timing, for shorts to pay out requires the market to actually correct itself. You won't be able to get your magical shorts money until the price of ram goes back down anyway. The market will remain irrational longer than you can remain solvent.
The price of RAM does not need to come down in order for a way-out-of-the-money January 2027 put on NVDA to increase in value from its current purchase price.
You're suggesting gambling that Nvidia will start to fall within the next 3 months? That sounds like requiring perfect timing to me.
Ack, typo! I meant January 2029. That’s how far out the options chain goes.
Google (which is diversified more and whose performance is less likely to track AI outcomes closely) aside, the two other members of the AI Big 3 are non-public, so, no, you really can’t.
Can I? Neither OpenAI nor Anthropic are publicly traded.
You can make a lot of money with shorts if you are correct about when the market will move. If you think that the market is overestimating the real value you have not even half of the puzzle.
With non-public companies?
The same thing was said 4 years ago about NVIDIA on HN, that it's stock it's outrageously overpriced, given it's $20 bln revenue, that it should have at least 10 times more revenue to justify that stock price, which is fantasy, that there is no plausible way for such demand no matter what you think about GPT-2.
nVidia earning is way above zero by a healthy margin. This was also true 4 years ago. How much their shares should cost is debatable but still, nVidia is obviously a profitable business.
Anthropic net loss in 2025 was $42 bln, OpenAI $38.5 bln. Both are spending enormous amounts with no obvious path to profitability.
The market can remain irrational far longer than you can stay solvent.
I'd come at this from a different angle: we still want this to be market system, so we need to make this priced into the market. How can we price this in?
I would try to solve this by making the market structure reflect the underlying difficulty: we have to decide what capacity to produce years in advance, to construct the memory fabs. So this should be a futures market, and a capacity crunch would affect short-term-futures, but leave full term futures at the same price. Because the companies supplying the memory can just construct more capacity to fill those futures at the same cost regardless of the AI demand.
Well, the best way to solve this problem keeping it a market system is if consumers band together and create a fab that will supply them.
Since almost everybody is a consumer of those companies, I do suggest we reuse the governance system we have that solves other "everybody problems".
ah yes, adding a futures market to a sector heavily invested in AI certainly won't lead to catastrophic over-speculation that will collapse the industry entire
> How should the memory companies have acted differently?
Same way nVidia did through the crypto insanity - make sure they're supplying enough to the consumer market so it doesn't get completely destroyed and pulls down the other parts of the consumer market they're reliant for long term success.
> Should we blame this on memory companies or the AI companies bidding for memory?
Blame doesn't change the outcomes, neither does it improve the negative consequences. Think in terms of "what does destruction of our consumer market mean for my prosperity?" not "oh, how do defend poor companies again?"
For one, they could have could have not massively scaled back consumer memory manufacturing as a matter of duty to customers. But money and greed must prevail.
Isn't it the case that public companies owe a primary legal/fiduciary duty to shareholders over customers?
The question of choice between a profitable vs unprofitable venture is easy. But yeah, to what extent is choosing profitable instead of VERY profitable a breach of duty to shareholders?
As far as I am aware, the answer to your question is ‘no’[0].
> Contrary to what many believe, U.S. corporate law does not impose any enforceable legal duty on corporate directors or executives of public corporations to maximize profits or share price. The economic case for shareholder-value maximization similarly rests on incorrect factual claims about the structure of corporations, including the mistaken claims that shareholders “own” corporations, that they have the only residual claim on the firm’s profits, and that they are principals who hire and control directors to act as their agents.
[0] https://corpgov.law.harvard.edu/2012/06/26/the-shareholder-v...
This started long ago, and is one of the best examples of “if you repeat a lie often enough eventually it’ll be widely believed”. It’s a useful tool to manipulate public opinion over a few generations. But no one would ever do such a thing, right?
Mm not really. This has been argued so many times on HN. It’s imbued in de facto case law. You’re not going to find it in legislation.
Willingly getting sued by hostile shareholders is in itself not acting in the best interest of shareholders so while it’s not in code, it’s buried in many decades of case law and in reality it reflects the maligned incentives for companies.
>Isn't it the case that public companies owe a primary legal/fiduciary duty to shareholders
The shareholders own the company. It is their property. They paid for it, they own it, and likewise they can do whatever they want with it.
It would be crazy if I proposed you let me drive your car to work everyday. Why? Because you own your car, it's your property, and it works for your own interest. Basically every human agrees with this logic, but somehow "the company is just focused on pleasing shareholders" escapes this.
This is not at all how share ownership works. You don’t have partial title to a corporation by owning shares.
Shareholders functionally own the company. Yeah you can't stroll inside and take a ream of copy paper because you own a few shares, but the company also cannot be sold (ownership transfered) without you getting paid for your slice.
The shareholders, via the board, hold final say over the company, it's direction, and it's alignment...because they own it.
This is like asking how many times can a politician can break his word before people vote for the other guy.
It's more like voting for public officials. Shareholders can vote to fire a CEO if they feel he's not acting in their best interests no mattter if that's the case or not.
No, and I don't know who keeps teaching you that nonsense.
And what does that mean?
Raid your own assets, cut costs manically so you can have a few splendid quarters with stock buyback bonanza or focus on long term value creation, which a lot of times involves giving at least a passing tought to other stakeholders such as client and employees?
The getting was great for some time for HP shareholders under Fiorina, or GE shareholders under Welch. Lots of them left the sinking ships at the right moment, but I bet that lot of the others left holding the bag, would have preferred having bought AAPL.
fiduciary duty doesn't mean you always have to do what makes the maximum profit
avoiding the destruction of good faith with consumers is a legitimate business interest
The overwhelming majority of incentives for executives is and always has been aligned with squeezing more value out of customers, in the form of increased margins, lowest possible costs with the highest prices the market will tolerate.
There are a few, I mean very few executives who stand firm on defending good faith toward customer happiness and quality. One of the only things stopping the spiral is competition. That’s why there’s an incentive to consolidate into a few massive conglomerates.
Good faith with customers also has a monetary value, even if it is hard to calculate. This is still maximizing profits.
both of them
memory companies could have attempted to protect consumers, at least a little, but the AI money machine goes brrrr
> How should the memory companies have acted differently?
Well if your thesis is that they should have acted differently, then we should blame the laissez-faire capitalists.
[dead]
An aside, this is a poorly-structured article. It's mostly screenshots of citations from other articles with no alt or any other way to actually search through what they're citing.
I gave up on the article and watched the video
People have the option of just not buying stuff. I sold my Tesla in 2021 for a stupid price, drove a beater for two years, and bought another new car a couple years later. Now, certainly not everyone has that luxury, but this is computers, not real estate. Prices will crash.
>People have the option of just not buying stuff.
Unless you know, your main and only PC/notebook just died and need a replacement ASAP, they you don't really have an option to not buy. And no, the used market (where I live) is just as fucked, proportionally reflecting the retail gouging, with people wanting quite a lot of cash for tower PCs and notebooks from the Pentium 4 - Core 2 Duo era.
> Prices will crash.
Does your crystal ball also say when exactly? If my laptop broke, then I need a replacement now in the overpriced market, not in X years when prices will crash.
If there were such a crystal ball, the same market doing the gouging would be pricing it in already ;) does suck that "wait and see" is the optimal answer to the average consumer but I guess that's the capitalist hellscape we have wrought
If this goes on long enough the prices crashing won't matter because the manufacturers of all the other PC hardware will be gone.
I think we're seeing the end of the PC as we know it, personally.
I think this is the goal. They too powerful for the plebs, enable too many things. Like thinking and learning. Can’t have that.
/s if you need it
I think it's more likely the industry would rather rent us our compute and storage than let us own it ourselves.
We can have our Raspberry Pis and our ESP32s, but meaningful amounts of compute and especially storage will be rental-only.
Requiring proof-of-humanity using device-bound cryptographic attestations (which the "age verification" camp will rally behind) will insure the toy computers we are allowed to own won't be able to conduct meaningful online discourse. Only phones and other locked-down platforms will be trusted to make the attestations necessary to interact with banks, merchants, and governments.
Edit (now that I'm not on mobile):
I think storage ownership, more than RAM or compute, is the cornerstone of digital sovereignty. Big tech will be happy to let us store our data on their servers for a fraction of the cost of self-hosting, so long as we let them train their models on our data. Any storage, be it self-hosted or rented, where the data is inaccessible for model training, mass surveillance, or censorship will command top dollar.
It'll be just like "dumb TVs" going extinct. "Dumb storage" will be impossible to get for any reasonable cost for consumers.
They're coming for business computing too, BTW, but consumers are the low-hanging fruit.
I hope so. I feel so, so, so terribly guilty for not building my wife a computer for her in 2023. She is rocking a 1080ti, and I really wish I had at least upgraded her system, but now I just straight cannot afford to even consider it, given the price of rent, food, gas, energy, all on top of the actual hardware prices.
And I am a SWE.I am not making bad money. I can't imagine what others are going through.
I was so lucky to buy a high-end miniPC to replace my home PC for about $1,500 USD in 2025. The same model now sells for more than $8,000 USD, without disk or RAM. Triple that if you match my original disk and RAM buy. Everything got incredibly expensive, not only storage and RAM.
Way to miss the whole point of the article and fail to understand the point. Do you think if significant part of US industry goes bankrupt because they can't source parts for their products, work machines and consumers disappear... you'll still have luxury of running around in a Tesla?
Could this be the opening China needs to get their foot in the door? This feels short sighted.
yeah CMXT is likely to come online in the not too distant future
All(?) ASUS mainboards got a week ago a new BIOS:
> Enhanced memory performance, system stability, and compatibility with CXMT memory chips for improved low-latency operation.
A year ago I'd hardly know what CXMT is.
This is how prices crash.
The AI companies must be regulated. They are already destroying the world and the economy before they can even make an attempt at destroying humanity.
Either AI will get cheap enough and blow up and destroy the economy by eating everyone's lunch and killing off sectors left and right. Or it will become an uncontrolable weapon destroying the internet and killing off people on the battlefield. Or it will tank the economy because there may not be a way to monopolize it making the billions of investments moot leading them to their only leverage of hoarding up all the resources pricing out the competition but also destroying the industry. And maybe all of those at once.
Yes, but zoom out more. Wealth inequality and the k-shaped economy are the real culprit.
When a few people have so much wealth that they are no longer price-sensitive, they bid up the price of everything and anything of value be it stocks, real estate, computer hardware, fine arts, sports teams, etc.
The result is that a business which tries to make quality products at reasonable prices will fail. They aren’t luxurious enough for the people with money. They aren’t cheap enough for the people without. Customers in the middle hardly exist. Any valuable good that is genuinely scarce will inevitably become a luxury.
This comes at a perfect time. The public doesn't need computers, they can just use their phones.
If the consumer had access to this RAM, they might all just run local or semi-local AI. It's important to outbid them so you can rent AI to them, and extract money from them in a million other ways while they use it.
Make RAM half the price it was before and you'd still have very few people trying to do local AI on it because it's just so damn slow for the task. Having 12 channels of DDR5 in an Epyc system with NVMe storage is still one of the slowest ways I can run LLMs locally.
Funny, I'm running local LLMs in a modest iMac M4, they are slow indeed but not utterly slow. A dedicated system should be way faster I'd guess.
The M4 (soldered wide LPDDR5X + unified architecture + GPU with neural cores) is also closer in design to a dedicated GPU than an Epyc server with DDR5 or the components in the article.
But yes, a dedicated consumer GPU can still typically outpace the M4 quite well (if things fit in VRAM) and both will have their socks blown clean off by a hyperscaler GPU cluster. For some hard numbers I can run a ~24 GB model on my 5090 (about as fast as you can get on a single consumer class device) about 3x-4x faster than on my M4 mini and I'd still consider that pretty slow to running models 10x the size in the cloud.
Did you dodge the phone price increase somehow? Phones also need RAM (and ambitions of running local models)
nvidia investing in every aspect of the hype cycle is terrifying. as an ai skeptic i almost don’t want the bubble to burst anymore, i don’t know if the economy can recover at all. perhaps a new economic paradigm will emerge, but if so, the transition is going to be rough.
The economy always recovers because the value in society is create by people existing and wanting things. The people won't stop existing or wanting things and eventually they just will believe something new is valuable and the economy will recover and grow. Because nothing material was lost just numbers in the spreadsheet. The recovery will be rough though your right about that.
Same for small business or any small buyer.
just stop buying. you people buy way too much crap anyway
> The DRAM and NAND flash manufacturers at fault for the ongoing RAM, SSD, and GPU price increases
Yes, it's the manufacturers fault, not the unbelievable market demand.
We are all Capitalists, until the Market comes after the stuff we love.
>We are all Capitalists
1/ speak for yourself
2/ where is your capital and your means of production? Your machines? Your factories? Your raw materials, your tools? How long do you survive tomorrow when you quit your job and you have no other income?
You're not a capitalist. You're just a consumer for them to suck dry. The faster you learn it, the better off you will be.
You are right. People are misusing "capitalist" to mean someone who favours a capitalist system, where as it actually means someone who posses capital; someone whose relies on possession of assets rather than earned income.
> We are all Capitalists, until the Market comes after the stuff we love.
Anyone with half a brain isn't a "Capitalist" (with a big C, like in your example) which is why every single prosperous nation (including USA) regulates markets to avoid destruction of their prosperity and society.
Colloquial "capitalism" today is just cronyism that in no way resembles the capitalism that Adam Smith wrote about. We really should move towards Rhine Capitalism, the only system that has stood the test of time and truly benefits everyone.
Adam Smith wrote about the invisible hand and let the free market run. What ended up happening?
It turns out the hand is invisible because most of the time, it's not really there.
He also wrote about the problem with monopolies, but that was later on in his book so we forgot to do that part so the market isn't properly free.
He only used that metaphor a few times, and he did not attribute the sort of mystical power to the metaphor that modern neoliberal economists do. He never proposed the modern extreme interpretation of the concept, nor did he suggest that government should release all control over the economy.
Marx treated Smith rather favorably, and credits him for first understanding that labor in general is what creates value. Smith is best understood as an early pioneer in understanding political economy and trade. He likely wouldn’t recognize or understand the capitalism of today (nor would Marx).
In the past people just starved to death when someone needed all the resources for their thing. Sounds like improvement.
> Memory companies
Someone misspelled "Altman", https://www.mooreslawisdead.com/post/sam-altman-s-dirty-dram...
Someone misspelled let's tax ai companies until we get back to balance.