- cross-posted to:
- technology@lemmy.zip
- pcmasterrace@lemmy.world
- cross-posted to:
- technology@lemmy.zip
- pcmasterrace@lemmy.world
Memory-maker Micron has found a way to keep prices for its products sky-high for another five years, by signing 16 “strategic customer agreements” (SCAs) that include a floor price the company says comes with “a very robust gross margin for Micron, well above our peak quarterly margins in any past cycle.”
Micron CEO, president and chairman Sanjay Mehrotra explained the SCAs in prepared remarks delivered during the company’s Q3 earnings call. He explained that Micron has signed 16 SCAs, most of them covering 2026 to 2030, and that they involve a commitment to buy a certain quantity of product and pay for it in a pricing band that has a floor and a ceiling price. The floor price covers the historically high gross margins mentioned above, and the ceiling price means those who commit to an SCA are insulated if memory prices go even higher.
You must log in or register to comment.
The same Micron that plead guilty to price fixing of memory 20 years ago.
https://www.theregister.com/off-prem/2004/02/27/memory-makers-hit-by-price-fixing-claims/1070959
not the same as price fixing if their entire output is being shoveled into long term agreed-upon contracts, is it?
gross margin for Micron
Gross indeed. Fucking greedy scumbags
Remember they are traitors to actual customers when the bubble burtsts.
They should fail, totally, and vanish from the world.
Traitors would mean they were ever on your side. Welcome to capitalism, bud. They’ve always been on the side of maximum profit, like all other corporations.
What alternatives would you suggest?
There sadly are none, now, outside of the “Big Three” who all did the same.
I’d suggest we start actively targeting smaller businesses for our suport, to prevent a monopol… Or triopoly in this case, from allowing everyone to be screwed by collusion and manipulation of markets and pricing.
Not a lot you can do with RAM, or SSDs, or GPUs now, but if we get competitors crop up we should be active in our suport of them and avoiding the ones who want to charge you $600 for an SD card now.
Plus a lot of this is part of the Bezos plan to force everyone into cloud computiong so all your data belongs to corporations and they can do whatever their constantly changing EULAs allow while overriding legal protecations like how a Health Exchange bypasses HIPPA laws and a corporate owned messenger is excluded from wiretapping laws if they sell their data to the agencies.
They’re like a movie actor who was in obscurity for decades and is suddenly popular. Is it being a traitor when everyone is offering larger and larger amounts for your time?
If you applied for two jobs, both said they wanted to hire you and started budding higher and higher so you’d work for them, are you a traitor to the customers who ultimately pay your inflated salary?
I don’t blame Micron. I blame the people who invested billions into AI which allowed this to happen.
The big 3 are known for being a price fixing cartel.
Any individual company could make more money by scaling up production. But they don’t because they already know the other 2 won’t.
So in your analogy the actors have all agreed to only work 10 hours a week and it’s nearly impossible to become an actor without having tens of billions to start
and it’s nearly impossible to become an actor without having tens of billions to start
There are over 300 companies with over a $50B market cap. The money is there.
Micron doesn’t expand production not because they don’t want more money but because they think AI is a fad like everyone else.
There are over 300 companies with over a $50B market cap. The money is there.
That doesn’t mean they all have $20B in cash to spare, OR the expertise to pull it off. Even China is still playing catch-up and nobody else is even trying because to get started in 2026 would mean poaching a bunch of people with nasty non-compete clauses to even get started planning.
There are documented cases of these same 3 companies colluding to constrain supply, even openly discussing what the target price of DRAM should be. Some of the people who were found guilty got their fines or even prison time and then came back and got promoted.
That doesn’t mean they all have $20B in cash
Apple has $68B in cash on hand. Not stocks or other investments, cash.
There are documented cases of these same 3 companies colluding to constrain supply,
Yes, but that doesn’t change that there are dozens of companies that could make ram but don’t because they pass off the price increases and blame ram manufacturers while not risking their own money.
Apple has $68B in cash on hand. Not stocks or other investments, cash.
And when they put all of that into their new RAM business, their investors are going to have a bunch of questions.
Yes, but that doesn’t change that there are dozens of companies that could make ram but don’t because they pass off the price increases and blame ram manufacturers while not risking their own money.
Even with tens of billions in cash it’s not that simple. Like I said, you first start off by convincing a bunch of people their non-competes and give up their massive annual bonuses, so that’s going to be an investment that the existing companies don’t have to make. Then you’ll have caught up with RAM manufacturing in about 10 years or so once you’ve ironed out all the details.
China, the biggest economic powerhouse in the world and the epicenter of modern electronics manufacturing, is still not making HBM3 or 4 or whatever nvidia uses now for Rubin at a commercially viable scale at CXMT, they’re only barely making enough DDR5 to get contracts.
That’s after starting the company in 2016 and getting DDR4 manufacturing going in 2020 and getting people from Samsung to steal trade secrets (and come work for them).
Just because you can build a fab for 20 billion doesn’t mean it’s only 20 billion to get started, or a quick 2-3 year endeavor to do so. Publicly traded companies can’t really do this because it’s a bunch of expenses over multiple years without a single cent of revenue for many years. This needs to be properly private cash or government backing.
And if one was based on seven companies selling debt to each other creating a bubble that will pop and destroy the economy?
I would take the one that paid less, but would not be an active plague on the planet and its people.
I wonder if anybody has told Micron about what happens when customers sign a contract but then declare bankruptcy shortly after.
Or what happens to the stock price once the temporary surge in demand fades or new companies enter the market and disrupt it.
That’s the point of the agreement. It locks in that demand by contractually requiring the companies agreeing to it to buy a certain amount of product regardless of whether they actually want it.
Personally, I view this as a sign that Micron believes the AI bubble is going to burst within 5 years, so they’re locking people in at bubble prices now.
Likely they’re just trying to hoard all of the compute in the US despite not producing many of the components, to try to create a substantive gap on AI. DARPA is not used to just being 6 months ahead. But eventually, something’s got to give. If nothing else, China is catching up on chip making.
“Locks in”…if all of a sudden there was no demand you can be assured they would “lock out”. Micron likes to put the boot to the throat when they have an advantage. Not someone I’d do business with.
Yeah, everyone was paying to back out of their contracts as soon as the prices went through the roof. The customers will do the same when they come back down if they are stuck in these contracts.
deleted by creator
Assuming these customers are the reason for the price hikes, their backing out is the demand loss needed to bring prices back down.
Software developers: more Electron and bloated frameworks are what the people want! Running 10 independent browser instances for simple chat apps is a great idea!
I never understood the scale of this bullshit until I had a user request his script packaged. Now, Intune doesn’t like executing random python scripts, so I gave him several options for 8 KB, 50-lines-of-code script.
He disregarded them and wrapped it all in Electron.
I received a 380 MB zip file.
But, well, it had an .exe inside so I could work with that.
¯\_(ツ)_/¯
Five years is too long for the buyers. The AI bubble will burst before then and then the market price will drop as the inflated demand disappears, especially if this continues long enough for more production capacity to come online.
They might not have had much of a choice in making the deal, though. Micron has been extracting the absolute maximum they can out of this situation. Make a deal or get nothing. Their clients will remember, though, and flag them as an unreliable supplier. Once this ends—and these always end—they’ll likely have a lower market share and end up having to cut prices.
Micron is optimistic in saying the demand won’t start easing until 2028. A lot of the rest of the technology manufacturing industry is about to grind to a crawl if not a halt because it’s nearly impossible to get components. Some companies are already delaying product launches and I think a lot more are about to this summer as they realize what’s happening. If non-AI businesses start to slow, the whole economy starts to slow, the AI demand will falter and that’s when the bubble bursts. I’m thinking maybe by the end of this year, more likely next year.
When the bubble bursts I’m guessing at least a couple of the companies Micron signed SCAs with will fold and Micron won’t get anything.
Interesting analysis. I was thinking the same, their customers might not make it.
About this point:
They might not have had much of a choice in making the deal, though. Micron has been extracting the absolute maximum they can out of this situation. Make a deal or get nothing. Their clients will remember, though, and flag them as an unreliable supplier.
Are the other two any better? If not Micron might get away with it. It doesn’t strike me as a very competitive market.
When the bubble bursts I’m guessing at least a couple of the companies Micron signed SCAs with will fold and Micron won’t get anything.
This is the key. The plan for a lot of these companies is that only two outcomes exist, unimaginable success where being gouged hardly matters or just utter failure and the obligations go away in bankruptcy.
Alternatively, they just break the SCA and maybe pay some penalty less than their obligation otherwise would have been. I have seen companies sign agreements knowing up front they will break the agreements, but the contract penalties still make business sense.
I’m still waiting to see what happens when OpenAI decides to back out of some of their purchasing obligations. It’s bound to happen, even if OpenAI does great. If folks think the tech sector is a bit wobbly the past few days, it pales in comparison to what such an announcement would do to the industry.
Isn’t that called price fixing, and is generally illegal?
its price fixing if the agreement is among ‘competitors’ - this is price fixing for a customer(s)
Yes, but the fine is far lower than their profits here… so, it’s only illegal if you can’t afford it.
“Cost of doing business”
If there’s some collusion, sure, but you’d have to find a government body with the will and teeth to prosecute.
Nothing really against charging whatever you feel like outside of things like certain supplies during disasters. It’s shitty
Only if regulatory bodies do something about it.
no?
Agreed, it’s obviously not related to that concept. Confusing suggestion.
Wouldn’t this just be selling security?
You would enter the SCA if you want to secure your supply chain against the risk of inflated pricing. The risk would now be overspending if the market drops. Comparing the two risk profiles, an organization might decide that they have more stomach for overpaying a set amount over 5 years. As opposed, of course, to the risk of paying an arbitrarily expensive amount indefinitely as the market remains volatile.
So now, while planning out the next 5 years of business objectives, you can plan against a much more solid best/worst case scenario. That minimized uncertainty, which lets the business keep moving even if at a more expense pace.
This is just like selling / buying options in any other industry, e.g farming to buy something at a specified price in the future.
They’ve agreed to buy XYZ product at 123 time at a price between $100 and $200.
If prices plummet, they’ve still agreed to buy it at $100 (similar to selling an options $100 Strike PUT), but if prices skyrocket, they don’t need to pay more than $200 (similar to buying an options $200 Strike CALL)
Anything in between is just the price, so if it’s $150 then its just $150
It’s not exactly the same as using options, but the rough idea.
I’ve never signed a contract to buy something 5 years from now at a certain price.
It’s a real thing.
You’re not a company dealing with massive inventories :P
Yes, but it’s AI, and Peter Thiel have bought all the politicians he could, so they will let them get away to “gain an advantage against China in AI”.
Unfortunately yes, but the current government is not going to enforce the laws
Padme: “…and then they’ll drop, right?”
Soooo if you raise the price sooo high then it lowers the price of barriers to entry. Congrats on maybe a year of monopoly before memory becomes a commodity like corn. Or AI crashes so hard that your customers refuse to come back. I feel like the later since using Micron memory for the past 30~ years.
Thus the 5 year contract - even if someone else starts adding supply they still get today’s high prices.
I’m not buying shit unless I have to. And then I’ll likely buy used.
Well, I know who’s gonna take a beating when the bubble pops and the market falls out from under them. What a stupid decision.
“Hey guys, this AI thing is gonna be like this forever. We’ll never lack for insane demand ever again.”
I think the logic for the customers is that either: A) It will work out exactly as predicted and we can afford whatever the hell we want, so it’s worth it to have secured supply
B) Declare bankruptcy, the purchasing obligations no longer matter.
They’re all positioning for government bailouts anyway. Whether Trump or a neoliberal lickspittle, they know they’re pretty much covered. All of the important c-suite people will be perfectly fine.
Does it go something like this? Those companies who buy the NAND chips to make AI accelerators, SSDs and RAM for data centers are the Microns customers, like Nvidia. Even if they are in trouble and know that the datacenters are not being built, they can’t cancel the deals because that would call the bubble into question. They will have to take any markup deal that Micron gives, them because if they don’t there will be nothing for them nor Micron anyway.
is there any good reason to charge more? I remember a long time ago, the people who run youtube had to have new technology invented to be able to store the videos on there.
is anything like that happening with ram chips?
Simple supply and demand problem
There are only a small number of RAM chips that are not already bought. If you make computers and want at one of those chips you can get priority if you sign a 5 year contract. They are betting that the supply/demand problems ease and so the price they can get will go down in a year - but you are locked into the high price for 4 more years. If prices go up even more you get a great deal, if prices go down you lose - make your bets, but remember this bet also comes with supply today and so you may not have a choice.
the barrier to entry for ddr can’t be so high that someone can’t buy a fab machine and undercut them can it?
It can.
China seems to be working on it.
They need to hurry up, I need some third shift Sunnlecs branded ddr6 STAT
Hold on, let me just solder together several billion MOSFETs and I’ll sell you some homemade RAM on the cheap!
It’s not just the price. If you get excavators to your new chip factory plot today to start building foundations it’ll take several years until you get first chips out of the line after everything is calibrated and ready to go. By then you’ve thrown few hundred millions on the building, machines and all the physical stuff. Hired and trained workers, managed supply chains and built a system which is pretty expensive to keep running.
So, you’re betting quite a lot of money and time against that the market stays like it is for the next 10 years (give or take) to just break even. If the bubble bursts in 5 years you have incomplete factory without potential market and a metric shitload of debt on your company. And that’s the same odd you’re betting against when trying to raise funding. Venture capital understands this risk too pretty well and that’s why everyone and their dogs aren’t building chip factories right now.
Saw a great documentary on the company that makes all the memory fab devices. They aren’t quite as bad a processors, There is only one company left that makes the house sized machines, but they’re largely self contained. Still need a clean room, but you probably don’t need as huge of a factory.
