One good thing about the AI boom is that it’s flooding capital into ram development and ram production. Once the bubble pops all these ram companies will have a glut of cheap high performance memory to sell to anyone who can buy it in cash. The question is if the bubble popping triggers a full blown recession will anyone have any cash to buy.
Only $1,000,000 per gigabyte, think of the performance!
when will they go back to optimizing software to run in the available ram, rather than continuing to increase ram size?
When the devs run out of available RAM.
Or if you’re 100 Rabbits and you live on a boat powered by solar so you’re actively limited by available electricity so you write tools based on your home-made tiny VM.
Who are “they”?
who builds software?
Governments, corporations, military, startups, communities, hobbyists…
The Stonecutters, obviously. https://youtu.be/fI4bRqjobbw
Who rigs every Oscar night?
If you’re able to optimize AI tools to run larger models in less RAM, you’ll have a lucrative future ahead of you.
Electron apps will still be slow.
I hope that one day I can afford a ticket to an exhibition where we all gather around a security-glass display case with a stick of one of those in it and bask in its presence.
8GB is only 2,762.89 before taxes. Act now, limited quantities.
Are there heat issues to deal with?
Since limited waver space (because HBM takes so much of it at the same capacity) is one of the main driving factors for the Ram-apocalyse, this might actually help. And the performance looks actually great (the comparison to the DDR4 i am running shows some parameters being 10x better, this is actually a meaningful upgrade), meaning that some applications that currently use HBM might be able to just use DDR6 instead, causing a greater allocation of wavers towards that.
Wafer
Generative ML inference isn’t so dramatically bandwidth limited anymore, either, so the “demand” for HBM isn’t as clear.
I hear superyachts are pretty nice too.
DDR6 has no official launch date. Many believe 2029 is a realistic target for commercialization. However, in the current state of things, it would be a miracle if DDR6 had no setbacks.
One can only hope the bubble pops before 2029.
It will, it’s coming next summer/fall unless it comes earlier. If the AI industry continues on in its current state until 2029 without popping, then it won’t pop. It means they stabilized their debt and financing.
The bubble will pop, unless it doesn’t. Noted.
One can only hope the bubble pops before 2029.
I mean a TiB of RAM sounds awesome, until you realize what a nightmare that would make setting up the swap partition if you ever use hibernate.
Plus I never even come close to using all of my whopping 64GiB, and even if I did I could still (theoretically) upgrade to 128GiB using the same slots (if I had a million dollars, at least).
What even is the point of using hibernate anymore? Used to be that it started up slightly faster on Windows since it’d skip a lot of the startup bloat, but in the year of our lord 2026, even Windows starts pretty fast on modern hardware, let alone Linux
Swap for hibernation does not need to be the same size as the RAM. It only needs to be big enough to hold what is currently loaded at that time. So in many cases you can get by with only a couple of GB or less.
Ha, amateur! I opened 16 new browser tabs while writing this comment!
Oh, I see. Good to know.
Great how real RAM development starts the moment nobody but a handful of billionaires can afford it.
Ddr6 is nothing new tho. Look at your gpu . It had it for years at this point.
GDDR versioning doesn’t necessarily line up like that. For instance, GDDR7 is technologically akin to LPDDR5X, just with a different set of configuration tradeoffs.
Cool the world is going back to ddr4 and ddr3. Who cares at this point? If it’s not going to be affordable and we already have all these pre existing computers that are working just fine anyway.
The progress must go on, supply will catch up eventually - hopefully, at least.
Nobody is going back to ddr3 except maybe the 3rd world and a few home servers. DDR4 maybe, but I imagine most people aren’t buying new machines just for cheaper RAM, can you really say “going back to DDR4” if you were never on DDR5?
DDR6 will get you much better performance for models that don’t fit entirely in your VRAM so it’s going to be very helpful for people running local models.
And when the memory market normalises, which could be pretty soon if nvidia stops ordering a ton of HBM it doesn’t even need, we should get back to the price per GB going down over time. CXMT is now apparently getting 90% yield on DDR5 which is excellent and the big 3 should be getting their first new foundries up next year since they started building them in around 2022 when the LLM boom started and they take about 4-5 years to finish.
So any time you see a headline saying the memory crisis is only getting worse, it’s because the members of the memory cartel REALLY want you to buy right now. Because they can’t keep up the grift for much longer.
Raises hand
Two of my Pc’s run ddr5, and the rest of the Armada are on ddr4. DDR3 still exists in plenty of Appliance devices and reliably running Legacy systems, and I keep a handful of DDR3 sticks around just in case I get to revive something older.
DDR6 will get you much better performance for models that don’t fit entirely in your VRAM so it’s going to be very helpful for people running local models.
For those of us who are not running after AI shit, DDR4 is fine.
At this point it doesn’t matter anymore if it were DDR99 with 578644.7 GT/s modules as we, consumers, won’t see them again in real life
$5000/stick or something.
I swear covid made the dumbasses in charge of major corporations realize that bigger numbers exist and that they could be put on price tags.














