this post was submitted on 24 Jun 2026
670 points (99.0% liked)
Technology
85739 readers
4889 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I know it'd be expensive, but I wonder if it'd be worth it to valve to start producing ram. They've certainly got the money to get it started, they are getting heavy into hardware that they can use it in, and they could sell it as well.
I don't know if there's a shortage of raw material or if no one wanted to invest in more manufacturing when AI could crash within a short time.
There's a reason why there's only, like, three RAM manufacturers. It's horrifically expensive to start production.
Chips. Where are they made? Right now its in Taiwan. Wafers for these chips are the most expensive part and that requires special factories and incredibly expensive engineers.
Valve cant just make chipa, they have to make it in quantities to satisfy demand while justifying upstart costs.
Yeah, you have to make much more than you need and sell the excess to keep per unit costs down. At that point you're a chip manufacturer. I'm not sure even valve could afford the startup costs involved.
In a nutshell this is impossible because of how the global supply chain works. Specifically how most of the hardware engineers/factories are in Taiwan, and how the technology to make chips is proprietarily owned by a company in Norway.
Like the whole reason China wants Tiawan in the first place is the same reason they can't just bomb them into submission... Their population of highly skilled hardware engineers that fundamentally make the global chips supply chain possible is impossible to replace.
And China also can't really invade because all the facilities that make the silicon are rigged to self destruct if China puts boots on their soil, at least last I heard.
It's a lot of things. But complex tech can involve literally thousands of hardware engineers. Each with very specific skills.
The proximity of these highly skilled workers to cheap chinese labour is another reason why this is difficult to replicate elsewhere.
Would be brilliant
I mean, it would bring global tech to a standstill. It would be a significant problem. Once existing stuff broke, there would be no replacement. I know very little about chip manufacture, except that the lithography machines are fantastically complex and costly. It would probably take years to spin up new production.
This seems like a pretty solid mutually assured destruction deterrent and doesn't even involve nukes.
You have clearly and concisely explained the exact reason the US wouldn't and couldn't allow China to invade Taiwan (well, wouldn't under a rational administration).
From what I can find, it looks like ASML has a software brick they can just drop into the update stream. As cool as physical disabling would be, a remote software trigger is simpler and leaves the machines in tact to spin back up after aggression ends
AsmL is a dutch company...
Also ,i'm not sure if HBM requires the smallest nodes
Manufacturing their own sticks would onlympush the problem to the price of RAM chips.
The resources it takes to start manufacturing modern RAM chips is such that THE ENTIRE FUCKING NATION OF CHINA is finally getting around to it.
I know Valve is a big company, but that's a pretty bite to chew and swallow.
Buying chips from CXMT and manufacturing the sticks is likely to be a solid business for someone. No idea if Valve wants to be the ones doing it, though.
My undiagnosed adhd brain: how difficult can in be
The actual process of creating semiconductors is basically:
In recent years, step 3 has gotten so complicated, based on needing to create radiation of exactly a particular wavelength of extreme ultraviolet light focused exactly on the silicon (and the mask/stencil above it), because that wavelength allows for the smallest possible features on the silicon. So they take purified tin, melt the tin into molten liquid, and ejecting the molten tin in a liquid jet downward into a vacuum at exactly the right speed to where it forms into droplets of the exact size for the machine (about 50 μm), then blasts each droplet, mid-fall, with a 1.6kW laser that heats it up so hot that it vaporizes and ionizes into plasma at the exact position where a system of highly polished and precisely positioned mirrors focuses the UV radiation evenly onto the silicon surface.
Oh, and the machine makes one tin droplet every 1/50,000 of a second, so in any given second it ionizes 50,000 droplets in the stream.
The machine costs something like $300 million, and requires full time experts to make sure that it's working correctly.
Everything else in the fabrication facility is similarly complicated, which is why a fab represents something like $30 billion in total costs over its lifetime.
Joke's on them with their fancy techno-babble. I'll build my own memory
I am on 5 mm technology already, how hard can it be to get smol?
That’s freaking crazy
Now is the time to do it for anyone that can. So much market share available to whoever gets there first.
Even Tsmc tried to make memory and failed. It's difficult.
They'd need to source the components outside of the increasingly monopolistic US-alligned group of hardware manufacturers. The only way you end run the Big Three is to go to... CHINA. And we've layered so many sanctions, tariffs, and putative measures on import of Chinese hardware that it would be a fool's errand to bother.
Even if there's an AI crash, the long-term outlook for chip demand only goes up. The problem isn't with the economic demand, it's with the provisioning of capital. For the most part, you need to spend tens - if not hundreds - of billions of dollars to start producing even the middle tier of nano-computing components in modern use.
I might suggest there's another way to tackle this problem. And it's one that Valve already is heavily invested in.
Lower resolution games. Lower hardware requirements. More efficient software engines. More games focused on the mechanics and story than the raw, realistic visuals.
You can run Doom on a pregnancy test and people still buy that game. Games like "Undertale" and "Vampire Survivors" do incredibly well in part because they are so accessible to anyone with a 15-year-old rig. Rather than trying to build a PS5-killer machine, you can go the Nintendo route and build a novel interface that runs on more basic components. Then you exploit the hell out of your Disney-esque IP without worrying that Halo: Remastered Delux Ultra looks better than the next iteration of Metroid Prime.
My understanding is that it’s the latter. AFAIK it takes something like 3-5 years to get a fab going if you already know what you’re doing, so it would not only be wildly expensive but you’re also gambling that RAM won’t come back down to a reasonable supply/demand in the next 5-10 years to break even on the whole process.
There’s also the fact that it wouldn’t really make sense for Valve unless they wanted to make a huge pivot in their whole business. Entry costs aside, manufacturing RAM is not really something a company can just do as a “side gig”. Valve is only like 400 people, so it wouldn’t be Valve just “starting to produce RAM” but rather Valve turning itself into a RAM manufacturer that also distributes video games.
You know i always forget how small valve actually is.
No need, others vountries ram are emerging. Hope they start to get to techno. Even better if they copied it from US. I think they are at reliable ddr4, testing ddr5.