It’s pretty accurate, AI and every other nonsensical idea in the IT world is making everything worse because all the big bosses want to see AI take off.
Baby Boomers destroyed the world and hated the Internet Gen X and the very oldest of millennials want to see AI take over… They’re now aware that AI will take their jobs too.
Gen Z and younger millennials are using AI daily.
I am 37 years old, I work in IT support… I know now to Google things, I know how to talk to people… So I don’t need AI.
Yes yes I’m very very aware how great AI is for science and medicine… I’m not in either of those fields so I don’t need AI.
I have some news for you, here is a real story.
I’m a programmer, a pretty good one, I have built real products worth millions of dollars and I’m doing alright. I’m no rockstar, I’m no pundit.
I have a raspberry pi connected to a touchscreen I use for home automation, it’s a nice little home IT project, I don’t really spend a lot of time on it. It’s a hobby and something to walkthrough/demo/make available for the kids to hack on.
I wanted to tighten up the install, get automatic updates running, get it in full kiosk mode and add a couple features. Nice little Saturday project, figured it take a couple hours on the outside, hardest part would be mapping the touchscreen again.
Claude did it for me in 20 minutes. So I got to spend an hour and a half thinking of features I wanted to include, pretty great for me! Pretty great for IT! Need doesn’t enter into it.
Maybe not great if your boss is a pinhead.
That’s actually really cool. All I can think of with AI is how many of the users I support are going to start throwing critical, private customer data into it asking for a recap… Completely unaware of how it’s going to store that data
I’ve wondered about ai in research, is it actually ai or s it more like machine learning?
If capitalism works, we’ll start to see competitors enter the market and prices would go down.
If capitalism worked.
Edit> In case it was lost, I changed tense in my second sentence because I know capitalism doesn’t work and only has ever served to make the rich even richer.
Prices going down from competitors is a capitalism lie which occurs sometimes.
The problem is it’s not quite so easy.
If this was making things like cars or toilet paper, where the design and manufacturing methods are well understood, that’s exactly what would happen.
Making computer chips was well understood. You take a slice of single-crystal silicon, dope its surface with something, use a photo mask and a light to etch off the doping where you don’t want it. Repeat that process a few dozen or a few hundred times to etch transistors into the silicon, then slice it up into chips and package them.
Machines that do this are readily available.
The problem is, we made those transistors smaller. And smaller. And smaller. And smaller. And smaller.
Then we ran into problems like how to etch features on the silicon that are smaller than the wavelength of light. And we found solutions to that. And so we made the transistors smaller. And smaller. And smaller.
So now to create a current-gen computer chip, you need a process called EUV- Extreme Ultra Violet. Red light has a longer wavelength than blue light, so when you want smaller wavelength to etch smaller features into the chip you eventually go from red to blue to ultraviolet and eventually to extreme ultraviolet. Problem THERE is there’s no EUV bulbs available, and if there were they’d be useless because atmospheric air absorbs EUV, and even if you do it in a vacuum glass absorbs EUV too so lenses don’t work. So you can only manipulate this light with mirrors, which have to be ground to an absolutely insane level of precision.
The resulting machine is quite impressive. You have a giant cavity kept in perfect vacuum. In one side you have an EUV source- that’s a little machine that dispenses a tiny 3-micron droplet of molten tin. As the droplet falls it’s hit by a laser to blast it into a pancake-like shape, then by another much bigger laser that vaporizes it. In the process of vaporizing it releases EUV light. So by blasting 50,000 droplets per second, you have a mostly continuous EUV light source. This light is reflected by shaped mirrors carved into the surfaces of the vacuum chamber that reflect and focus the EUV into a linear beam. Below that (also in vacuum) you have the mask and the silicon wafer, and by moving them back and forth under the beam you etch features ‘smaller than light’ into the chip.
The result is a chip with current pathways less than 100 silicon atoms wide. And if you want to make current-gen computer memory, that’s the only way we’ve got to build it.
EUV machines are pretty much only made by one company, ASML. They’re the size of a double decker bus, they cost a fortune ($200MM+), and they’re in insanely high demand. Like I’ve heard of a guy getting hired by a company for over million a year simply because he’s friends with a purchasing manager at ASML and might be able to get the company that hired him a build slot.
And it’s not just buy the machine and hit ‘start’, there’s a ton of other stuff involved. You need machines to make the photo masks, you need an ultra-clean cleanroom, you need a robotic facility where cassettes of wafers can be whisked from machine to machine with no exposure to even clean room air. And this is all highly specialized stuff, you don’t just call 1800-fab-4you and place an order.
Bottom line- even for a company that already has experience in current-gen chipmaking, setting up a fab like this costs $15-20 billion. And it isn’t just ‘sign a check and come back tomorrow’, the process of building a fab from the project being approved to the first wafer coming off the line is 2-3 years minimum.
Now here’s the bigger problem- semiconductors are always a cyclical market, or at least always have been. Demand (and thus prices) goes up, demand/prices come down. So if you invest $20 billion when prices are high, then in 2-3 years when the cycle is at its low nobody’s gonna be buying your output. And of course, adding more chips to the market will affect market prices (supply and demand). So companies that are building fabs have to look 5-10 years ahead to determine if they’ll get ROI on a fab before they build.
And that brings us to the next issue- with AI, we’re in uncharted territory. The computing market has been pretty well understood since the early 90s. There’s demand for PCs and laptops and servers and gadgets, and it goes up and down and new products come out that changes the mix of what’s ordered, but the cycle more or less continues. Up and down.
Then AI happens. And suddenly we have near-instant, unheard-of levels of demand. And it’s all for current-gen top-shelf stuff- HBM (high bandwidth memory) and GPUs and specialty silicon like NPUs.
Now there are more fabs being built. But it’s also starting to be better understood that AI is a bubble, which almost certainly will pop. So if you’re a DRAM maker and you spend $40 billion building a fleet of new fabs and then the bubble pops, you’re gonna be fucked. That’s why you don’t see everybody+dog diving into the DRAM market face first.That’s probably the best comment Ive read so far on all of Lemmy, thanks!
Most welcome :)
This was a very interesting read, thank you!
Most welcome :)
Just to add, since I didn’t see you mention it: even when you build a fab from scratch and know what to fabricate you need to find and hire people with the “know-how” on how to set the processes up.
All true but it also assumes valve isn’t thinking post gaben
Building a fab now and getting the experience and expertise it needs to build next gen, such as abandoning silicon like the one article I saw said might be the direction things go in might be in valves orbit. Froml all accounts valve is highly profitable and could potentially weather the storm of upfront capital and growing/learning pains.
Also as I understand it valve doesn’t hire developers that don’t have a minimum of 15 years in industry related employment so if valve maintains its current hiring standards (and I’m not completely wrong or it has already changed), they VERY easily could poach all the experience they would need to since they would have the time. Might be a very exciting and enthralling opportunity to work for a gamer/pc focused fab.
I mean I doubt it but if they were thinking long term it would given them a comfortable time frame to do it in.
If turns out they already have the capital to self finance it all that wouldn’t entirely surprise me they could.
So, you’re saying Valve should make RAM? I don’t think that would happen, but it would be an interesting thing to see unfold
Would DDR4 or DDR3 be easier to produce?
In some ways yes. Older designs on older nodes use older equipment- less precise, easier to build. There’s actually a lot of PC builders who are building machines on DDR4 mainboards because the memory is more affordable. A couple of DDR4 mainboards have been ‘un-discontinued’ as a result.
For a desktop or laptop PC, the difference of DDR4 to DDR5 often doesn’t make a huge difference. So you get a last-gen chipset and RAM and you get a decent machine without paying $thousands extra.
CXMT is our best bet. But it takes time to ramp up production. Some article I saw said they’re building things at a quicker pace than other manufacturers typically do which is a good sign.
But when demand slumps back down then now you’re stuck with all this manufacturing capacity that’s not useful anymore. So you have to balance dealing with short term spikes vs long term demand.
Even if it worked it wouldn’t. Neoliberals and free-market advocates will tell you that competitors would rise up if there were no regulations.
But some industries (like silicon and microchips and other advanced tech) are just impossible to breakthrough without years or decades of setting up and operating without profit vs current players.
And those current players, given no market regulations, will still use their money to do anything possible to stifle upcoming competitors or buy them out.
The dream of capitalism was born pre industrial age, under models of “apples sold at the farmers’ market”.
The reality is we have advanced a lot and need mixed economies to keep things fair for most people.
I’m sorry, but capitalism didn’t really exist up until industrialism. Before then, it was mainly feudalism.
Googling is easy. We had mostly capitalism before, the industrial evolution just pumped it even more.
Plus the IE wasn’t a single date but a period. And my point is that previously held ideas of how markets work affected People’s hopes of what competition in a capitalist market would be like.
…in spite of how the industrial evolution created industries (lol) and markets so vast that they became larger than any human or any family trade ever was, intergenerational behemoths that are simply too resistant to competition because of how it takes literally multiple lifetimes to reach their level without already being a competitor.
At this p;oint if I go to a site and they have an “AI Chat Assistant!” pop up I send them a “support” message telling them that I will not be purchasing from people who contributes to the bubble that is driving up costs for components.
I just did that for an MP3 player, why the everloving frak they need an AI assistant to give me a “compare” feature that’s been on commercial sites for decades…
A bit off topic, but what sort of mp3 player? It crossed my mind a few times that I should get one again.
Why would you want the mp3 player they are talking about? If they cant even shell out for decent web dev, why would the hardware be any less cheap…
They could be looking for what features the other poster is looking for specifically that they might not have thought about. Take advantage of the time someone else has put into researching options. No need to do the same research again.
They also could want to know the brand doing that to avoid them.
Exactly this, thanks.
Imagine if valve put it’s money into making components like ram and storage just because the guys doing it already are hurting gamers
Making RAM requires billions of dollars in investment. That’s why there’s so few manufacturers worldwide. I highly doubt that Valve could afford it.
Let’s see what China does then
They’re trying, but they have a lot of catching up to do.
Though fingers crossed that they do catch up and they did actually release their first GPU this year, The Lisuan LX 7G100, costing around 500$. Performance should rival rtx 3060, but there aren’t any third party reviews. https://www.pcmag.com/news/china-just-made-a-gpu-thats-powerful-enough-for-gaming-but-theres-a-catch
It’s unlikely for sure but not impossible as we have no idea what they actually haul in. They have been said to be worth 20-40 billion but they may be worth much more unbeknownst to us.
I also highly doubt it and said as much in a different comment
At what point do we start making the factories ourselves? No really if ddr3 is significantly going up would it make sence for all these smaller companies to pool thier resources and start fabbing thier own chips?
unfortunately chip foundries take years to set up (just filtering the air alone takes months), so it’s not really a solution - although china is attempting to do exactly this.
"It’s hard work and it takes too long. China is trying to do it "
2 years later and after China has done it, “why are the chinese taking the lead on all of this? we should be doing things more like China.”
Sorry, I was perhaps unclear - China is and has been setting up chip foundaries regardless of the global memory prices, simply to have domestic production of those components. Production of memory modules during this period of AI ratfucking is just a happy bonus.
Yes because they have an autonomy incentive, but “our” (as in Western capitalism) incentives is solely on profit and profit is being done, therefore we will never solve this because in fact there’s no problem. We are being mocked by all these PC parts suppliers.
The fact it runs on profit is not the deciding factor in this case. If the RAM manufacturers had more capacity, they would profit even more. The main problem is that those factories are huge investments that you would need to commit until the finish-line successfully without going bankrupt before you will be able to produce anything. China can take this risk because state funds and actual geopolitical justifications. Western countries can say to themselves too easily that they can keep buying it from Asia, which causes no countries or investors to want to commit to a similar project.
I don’t think you’re wrong, I’m just not sure how it relates to what I said.
But they really wanted to shit on capitalism regardless of context.
What’s weird is that capitalism should be driving others to set up fabs because it is so profitable.
I mean fair enough I suppose. It’s almost like capitalism doesn’t quite work like how it says it does in the advertising… sigh.
Haven’t they been drooling over Taiwan’s tech and foundries for years now?
More or less, though they recently bought their way in to a domestic EUV foundry so their coveting of TSMC will likely reduce in the near future as that comes online.
How did they get an EUV foundry? I thought ASML is the only company in the world that can make the machines and they aren’t allowed by the US to sell EUV machines to China.
iirc they just bribed a bunch of engineers to move there and build them one
Why filter the air at all? Why not just turn the clean room into a vacuum? No air contaminants since there’s no air. All the stuff in there looks to be automated anyway.
I have no idea how silicon fabs work so this probably isn’t realistic.
Many reasons! The enormous building imploding is probably the big one, but risks from cold welding, use of atmospheric cooling, relying on exposure to atmosphere for many steps in the process, etc. are all additional factors. It’s a good question though.
Just put them in space, that way they’ll be right next to all the datacenters that Musk will be building there. It’s much more convenient!
Interestingly they’re also doing that, too. At least, if you uncritically accept the popsci reporting about it like musk probably did
I can just see makers starting to do small scale projects such as https://www.tomshardware.com/pc-components/ddr5/russian-enthusiasts-are-building-their-own-ddr5-ram-amidst-the-worldwide-shortage-as-easy-as-sourcing-your-own-memory-modules-and-soldering-them-on-empty-pcbs
And then naturally expanding. I mean we talk about all these hypotheticals but at a certain price range, impractical turns practical. If you cannot convince any big chip makers to make your thing and people will pay 2x for your thing, then getting together with smaller manufacturing starts to make sense.
I dunno im just a small maker person. Made my own laptop from parts and all that. I see this as a potential opportunity.
You still need the chipmaker to make you the memory module, assembly is not the problem
There is a guy on YouTube doing it and making videos about the process. He turned his shed into a clean room lol.
Sure. A hobbyist with the right equipment can make technically functional memory as a scientific curiosity. But making any memory module that is as powerful as anything that’s come out in the last twenty years is going to require millions of dollars in investment and years of effort combining the work of thousands.
Oh nice ill take a look! We have a 10k pick and place machine over at a local makerspace and ive thought about prefab chips before. Just for fun, but its still a bit of money.
Nobody (aside from China, for purely geopolitical reasons) really wants to invest billions of dollars in a multiple years long process that could be instantly undermined by the popping of the AI bubble.
This level of demand is not sustainable, and the big tech companies know it.
This isn’t really accurate, there’s been literally hundreds of billions invested over the past several years. The only reason prices are predicted to normalize after 2028 is because that’s when the new fabs come online that those investments are paying for.
Who do you think is going to be buying DRAM at the current prices in 2027?
Before the bubble, there used to be oversupply of DRAM leading to price fixing scandals, that’s what will eventually happen again.
Unfortunately, we’re not going to get an over supply of DRAM once the bubble pops.
Most of the current production is HBM, which is juuust architecturally different enough from DRAM to not be useful for desktop computers
I was under the impression that the underlying memory chips are the ones we have a shortage of, and HBM is just one way to package them?
Probably anyone who wants RAM. 2027 production is mostly already sold, the shortage isn’t going away until there’s more fabs. 2027 is more likely to see higher prices than lower.
Production being sold out doesn’t mean anything when there are no commercial entities left to actually pay the bill.
You should look into what is actually going on in the industry, the bills are getting paid. Companies are even pre-paying for future allocations, and there is a transition towards 5 year contracts.
I thought ireland and a couple of other countries are stepping up quite a bit recently? I know they have been at the forfront of the EUs efforts to start building chips and have been doing so last 10 years to start scaling.
Just did a search, looks like its going to happen: https://wallstreettimes.com/intel-foundry-5-billion-ireland-chip-manufacturing-2026/
This level of demand is not sustainable Totally agree!
Seems like also for geopolitical reasons
Making RAM at home:
That is so awesome! NGL, some of that equipment looks more sophisticated than what I use in my university’s semiconductor lab. But the spin coater with the pot lid cracks me up a bit.
Nice profile pic. I’ve been wanting to find a maker/tinkerer community on lemmy but I haven’t come across one yet. Am I just not looking hard enough? I love seeing and discussing people’s DIY electronics / 3D printed projects.
Your client shows profile pics?
Yours does not?
Piefed works on firefox moble well, so thats what i use 😁.
I used to use sync, but it’s been abandoned and I heard from other former sync users that summit had a comparable feature set, so I switched to that. Summit shows profile pics next to each user.
Theres some. Honestly more people are on mastodon than lemmy. It hasnt taken off like the mastodon crowd. If you do #maker on mastodon.social for example, its much better.
I wish the threadiverse supported hashtags and more mastodon stuff but it is what it is.
Thanks!
Really? I’ve always had fewer interactions on Mastodon but then I’m fairly new here and I was fairly early to Mastodon.
I had to really look. But once i started following it was an avalanche. Lots of indie creators and such.
It takes 4-5 years to build a new memory fab from scratch for companies in the industry.
They’re talking about no price reductions for 1.5 years.
(And the reason they’re saying that is because the first significant new production capacity isn’t expected to be up until mid-2027 and isn’t expected to be scaled up to full production until 2028.)
And they already moved that forward by something like six months, probably by spending more on stuff like construction, so I doubt that there’s more slack to eliminate. Can’t just speed it up a little more.
It upfront investment cost is simply too much to just start manufacturing your own chips. Just the fab alone would take around 20-30 billion and 4-6 years to build. Then add another couple of billions and 3-4 years to design a DDR5 chip that would give comparable performance to what the big 3 manufacture.
ASML has an order book running to 2028, the problem is it’s almost impossible to create a competitor to ASML - Demand for fabs is not the issue.
At scale sure. But for like a small controller or power supply?
It’s decades away. Making dumb chips is possible, but we lack the supply chains, infrastructure, and learned knowledge of making advanced tech.
If you doubt this look at how intel and nvidia are having a hard time in making usable chips on US soil.
Earlier this month, research showed that memory prices are predicted to rise sharply between now and 2027, with no reductions expected until 2028. Now, Valve has shared its own view of the hardware market, and it’s equally glum.
That’s not really a position new to this month.
TBF, from how they’ve described the situation it sounds like Valve has as much sway and insider knowledge as me or you. So I don’t think I’d trust their opinion here
I saw in an interview of their engineers that memory suppliers basically just say take it or leave it, at the risk of paying more later or having none at all.
On the scale of all tech companies, Valve is very low down. Even Big Players like Amazon have trouble getting the best deals; if you’re not Apple or nVidia, you get to just go fuck yourself.
last i heard even nvidia having problem making gaming gpu that they’d slashed 50’s card production down. now im not sure how credible this source is but its pretty telling how bad the situation is rn. even if its a lie to slash unprofitable product line.
Samsung is one of the Big Three memory manufacturers. Samsung’s phone division had some news articles talkimg about how they couldn’t get Samsung’s memory division to give them privileged access, even.
(Though for context, about two years earlier, when there was a glut of memory and Samsung’s memory division was losing money, they apparently went to the Samsung phone division and tried to get them to buy a bunch of the excess, which I suppose would have meant putting more memory in their phones than would have been optimal. The phone division told them no way.)
And now it’s the phone division losing money and the memory division making a ton of money.
https://www.megamobilecontent.com/news/2026/07/17/samsung-phone-division-first-loss-memory-prices/
Samsung’s mobile division may have posted its first-ever quarterly loss in Q2 2026, despite strong Galaxy S26 sales and record-breaking profits for Samsung Electronics overall.
The culprit is memory. DRAM prices have reportedly surged roughly 850% year-over-year, driven by insatiable demand from AI server infrastructure. Samsung’s semiconductor division is reaping the benefits, but its mobile division is paying the bill.
That’s not too surprising. Samsung is one of those mega companies that its own divisions fight for contracts all the time. I remember hearing about, for example, the Samsung division that produces touch screens isn’t guaranteed a spot in the latest Galaxy smartphone but has to bid on contracts for that spot just like everyone else.
Well, I don’t think Nvidia is “having trouble” making 50 series cards. I think they have fab allocations that they are choosing to prioritize to AI products and its current ROI with blank check hyperscalers, compared to retail consumer products.
Lmao what makes you think even Apple has any special treatment?
Industry experience.
Apple has been overall getting the sweetest deals from chip vendors for quite a while now.
I mean, historically, absolutely, everyone knows that. But do you know for a fact that, now, in 2026 they’re still getting first pick at normal “sweetest deals”? I would assume maybe better than some, or at least still getting supplied at all. But I thought I recently (in the last 2-3 months’ish) heard something about them also getting fucked: by the big AI datacenter players; Apple, to my knowledge, is not, at least not on that scale.
If you have insider/industry experience, though, that’s that… I no longer am close friends with people at Apple, and to dispute confident, current inside knowledge becomes a different discussion for a different place with different people.
My information is probably less out-of-date than yours, but yeah, they’re probably getting hit as well, just not as hard.
I’m curious what their next iPhone is going to cost
They did have to bump up the prices.
Yeah, the specific RAM shortage is affecting everyone.
Also it’s Apple. Even if they got the chips at the same price before they would never NOT raise the price if given the opportunity.
Why do people think Valve has some godlike influence over everything? Sony has sold over 90 million PS5-s, Sony reduced its Playstation 5 manufacturing in part because they struggled to buy RAM. Valve has sold less than 5 million Steam Decks. If a hardware powerhouse like Sony can’t get all the RAM they want what makes you think Valve, who barely has presence in the global hardware market, has any sway in the global ram production?
deleted by creator
Maybe about the same sway but they are purchasing memory on a different market that is about future goods rather than goods currently sitting on shelves. So they will have more day to day visibility on RAM futures than your average consumer will. Price increases will happen there before they affect store shelves.
They certainly have a lot of motive to place blame for their own price increases, justified or not.
Okay Valve. Can YOU make memory?
No they can’t. Building a fab costs $10-20 billion and takes 2-4 years. And that’s assuming you have the people, the expertise, and the IP to be able to produce memory in your new fab.
Now Valve is privately held, 50+% owned by GabeN, so real numbers aren’t public. But it’s estimated to be worth around $20-40 billion. That’s why he can afford to spend his time driving around on yachts and submarines. That means if Valve wanted to, they might be able to get into the business.
But from a business POV, it’d be a HUGE risk. You’re betting basically half the company that you can go in with ZERO experience in semiconductor manufacturing and start making top tier memory, AND that by the time you can bring it to market there’ll still be any demand for it. That’s a BIG risk. And given that they’re making money hand over first selling hardware built by others, it’s not a risk they have any need to take.
So there’s been a massive spike in demand for a product, that demand isn’t forecast to let up for the foreseeable future, and yet it’s outright impossible to get in on the action.
RAM is Lostech.
If Valve said the crisis is going away soon people would then wait for the prices to come down. This is very basic manipulation to convince people to buy now.
Yeah, Valve is really having trouble moving those units. Never mind that they sold out in a couple days and have waitlists till next year
To buy what? They are out of stock
I’d believe this if these devices weren’t sold out and this information contradicted everything else we know about the situation on the market.
Jesus Christ, who hurt you?
I mean that question rhetorically; I know who it was, it was everybody. So, I don’t blame you. But I do think you’re off base a bit on this one, specifically.
Especially considering the congruence with all of the evidence pointing to them telling the truth.
