| ▲ | piinbinary 12 hours ago |
| I have a suspicion these new prices will stick around, even after the RAM shortage ends. Speaking of which, what's the timeline of the RAM shortage ending? I have no sense for whether it is going to be (for example) 6 months or 3 years. |
|
| ▲ | revolvingthrow 12 hours ago | parent | next [-] |
| >what's the timeline of the RAM shortage ending? Barring unusual market forces like Taiwan invasion the timeline to ending the acute shortage seems to be mid 2028. The AI still has plenty of money to burn and is the biggest driver, but we’re also shortly before gaming consoles ought to release a new gen (although who knows whether they won’t get delayed for a while). There was even going to be a small upgrade cycle for nerds waiting for 2nm fabbed devices, same as pre-ai datacenters looking for power efficiency. Plenty of pent-up demand, too, as many people simply make do with what they have but will upgrade once the silliness stops. If you’re looking for ssd/ram prices to go back to the low of 2024/early 2025 it probably won’t happen before China catches up, which will be a while yet. There is some build up of new capcity happening from current manufacturers but it’s significantly less than what the demand increased by. |
| |
| ▲ | 15155 7 hours ago | parent | next [-] | | > Taiwan invasion Are many DRAM fabs in Taiwan? Does TSMC manufacture DRAM for SK Hynix, Samsung, Micron, or CXMT? | | |
| ▲ | Chris_May 5 hours ago | parent [-] | | If TSMC is blown up, how much RAM is used in new datacenters? | | |
| ▲ | 15155 5 hours ago | parent [-] | | I was thinking this as well - DRAM and NAND demand instantly crashes because the primary chip supply ceases to exist. This "shortage" will pass. |
|
| |
| ▲ | forestingfisher 6 hours ago | parent | prev [-] | | China won’t invade Taiwan. Be realistic | | |
| ▲ | lantry 5 hours ago | parent | next [-] | | If you wanted to be realistic, you wouldn't say for sure that they won't. | |
| ▲ | SXX 4 hours ago | parent | prev | next [-] | | Uh oh 5 years ago nobody would believe there will be a war in Europe with almost 1,000,000 dead. And now China knows that both US and EU are weak and cant even deal with Russia or Iran. And they also have their own semiconductors manufacturing and cut off from what TSMC produces anyway. | |
| ▲ | shepherdjerred 2 hours ago | parent | prev [-] | | And Russia would never invade Ukraine. And Germany could never make it past the Maginot Line |
|
|
|
| ▲ | haunter 12 hours ago | parent | prev | next [-] |
| It’s a permanent price hike Eventually supply and demand will get back in a better balance and we will probably see prices rise slower than inflation until adjusted for inflation prices are close to to where they were before but the actual dollar price isn’t likely to go down. |
|
| ▲ | ErneX 11 hours ago | parent | prev | next [-] |
| The new Xbox CEO said recently they are expecting storage prices to be 5x what they were late 2025 by late 2027. And that RAM should be similar. Anyone making hardware is having a rough time. Like Valve who had to release their new PC at around 40% more expensive than what they originally wanted. |
|
| ▲ | winocm 3 hours ago | parent | prev | next [-] |
| https://www.theregister.com/systems/2026/06/25/micron-locks-... :( |
|
| ▲ | dwa3592 11 hours ago | parent | prev | next [-] |
| Until China floods the market with their memory which is starting |
| |
| ▲ | jauntywundrkind 9 hours ago | parent [-] | | I'm seeing it with NAND. Look at the AWS Prime ssds available, and it's a massive list of strange drives you've never heard of, with very few reviews available, almost all using YMTC. The prices aren't fantastic, but given that five sixths of the drive market is straight up gone, I think we need to start reviewing and hoping for the best here, fast. I really hope endurance is indeed as rated, that we aren't about to all get burned incredibly badly for using YMTC chips. CXMT is indeed starting to get some ram out there. But I am pretty skeptical it's going to make much of a dent. They're currently single digits % of the world ram production. They need to scale a lot to make any dent, and as soon as they do, it feels like there's plenty of people ready to snatch up those supplies. We need massive massive massive growth in availability and there's no sign that current scale up plans will be at all adequate. | | |
| ▲ | rescbr 5 hours ago | parent | next [-] | | The problem with these drives is that you can't ensure the grade of the NAND chips. They could be A+ grade with great endurance but they could also be bottom of the barrel. I had two KingSpec SSDs sourced from AliExpress that failed too soon: one used YMTC and the other used Intel chips. Both failed within 1 to 2 years. I have another one which is 5 years old by now. | |
| ▲ | malshe 5 hours ago | parent | prev | next [-] | | > Look at the AWS Prime ssds available Where? | |
| ▲ | dwa3592 9 hours ago | parent | prev [-] | | right, but that seems to be the only viable path for any reduction in prices unless the bubble suddenly pops which these ultra qualified people (sam, dario, elon, oracle and so many more) won't let happen. |
|
|
|
| ▲ | brandrick 12 hours ago | parent | prev | next [-] |
| Considering what's causing it, I can't imagine it's a particularly short timeline. |
|
| ▲ | jorvi 11 hours ago | parent | prev | next [-] |
| With new fabs built and AI demand shrinking, they will have to. If they don't, considering the last lost price fixing case, they will be absolutely crushed by the EU and probably other governments as well. |
|
| ▲ | mDyJzDPmBdG 12 hours ago | parent | prev | next [-] |
| On supply side 3 years is about right, new plants won't come online faster. Demand might collapse faster if some AI companies go bankrupt or at very least fail next funding round. |
| |
| ▲ | thewebguyd 10 hours ago | parent [-] | | Depends on who goes bankrupt and what happens to their IP when it happens. If OpenAI or Anthropic liquidate, and the IP gets scooped up by MS, Amazon or Google, demand will remain, the public clouds will still want to run them. Maybe some pressure will come off if they lose the appetite to train new models for a while, inference is cheaper, but they'll still finish some of the buildout. |
|
|
| ▲ | justincormack 12 hours ago | parent | prev | next [-] |
| At least 3 years maybe more. |
|
| ▲ | snarfy 5 hours ago | parent | prev [-] |
| When the AI bubble pops. |
