Valve's Steam Machine Has Been Delayed, and the RAM Crisis Will Impact Pricing (theverge.com) 40
Valve has pushed back the launch of its Steam Machine, Steam Frame and Steam Controller hardware from its original Q1 2026 window to a vaguer "first half of the year" target, blaming the ongoing memory and storage shortage that has been squeezing the tech industry.
The company said in a post today that rising component prices and limited availability forced it to revisit both its shipping schedule and pricing plans. Valve had previously indicated the Steam Machine would be priced at the entry level of the PC space.
The company said in a post today that rising component prices and limited availability forced it to revisit both its shipping schedule and pricing plans. Valve had previously indicated the Steam Machine would be priced at the entry level of the PC space.
Boooooo (Score:5, Funny)
Re: (Score:2)
Thanks, Altman!
(Sure, there are others but he's the poster boy)
Re: Boooooo (Score:2)
Re: (Score:3)
To throw darts at.
Manufactured crises (Score:4, Insightful)
Where LLM actors are using investor money to suck up everything in order to keep it out of the hands of the competition.
Cannot wait for this bubble to pop and we get to see those responsible hanging from street lights. (Hahahahahaha - not)
Re: Manufactured crises (Score:2)
it'll still be entry pc price.... (Score:2)
but entry pc price is going up too.
Re: (Score:2)
excuse me?
entry PC prices (Score:2)
Valve: "The Steam Machine will be priced comparative to an entry level PC..."
AI Market: Price of entry PCs now jumps up to "both your kidneys and a testicle"
Re: (Score:1)
Prices: how long? (Score:2)
can litterly pick up a 5060 pc at a big box for that.
Yes, but for how long? That's the problem.
If the current trends of rising prices of core component keep up (i.e.: unless the AI bubble bursts already), EVERY newly built machine is going to be expensive.
Re: (Score:1)
upgrades delayed (Score:2)
Re: (Score:3)
Well, the AI tech-bros still haven't found anything close to the revenue needed to justify their current insane spending.
I'm holding my breath for a big explosion that sinks 99% of them. Leaving the remaining hanging on by threads that never recover.
The sooner it bursts the less damage the dopy LLMs will do.
Re: (Score:2)
The problem is that even when the bubble pops there won't be a flood of hardware for people to buy for pennies on the dollar. The memory isn't in the correct form factor for desktop use.
So manufacturing needs to switch what they are supply, so there was be a multi month lag before the supply hits the market as new products.
Re: (Score:3)
The problem is that even when the bubble pops there won't be a flood of hardware for people to buy for pennies on the dollar. The memory isn't in the correct form factor for desktop use.
So manufacturing needs to switch what they are supply, so there was be a multi month lag before the supply hits the market as new products.
I'm looking forward to the fire-sale on giant racks of GPUs. Sure, our houses will sound like jet engines, but we'll be able to render our Blender projects in real time as we go!
Memory (Score:3)
The memory isn't in the correct form factor for desktop use.
The latest crop of server indeed use LPDDR5X which is indeed incompatible with DDR5 DIMMs and SODIMMs (it's not possible to de-solder the LPDDRs off recycled mainboard and re-solder them on new DDR carriers PCB and plug those into conventional motherboards' DIMM slots).
So yeah, you won't see the market flooded with rebuilt DIMM and SODIMM sticks.
BUT it's a format used (soldered) also by SBC, mobile devices (tablets and smartphones), and some laptops.
So if the hardware is cleared "for pennies on the dollar"
Re: (Score:2)
The big issues with the GPU is many don't have active cooling, instead expecting a front to back airflow to exhaust fresh air across it.
So you need some fairly forceful directed air movement. You can rig up things, but it won't be just plug this desktop GPU into the motherboard and connect PCIe power and you're all set.
GPU cooler (Score:2)
The big issues with the GPU is many don't have active cooling, instead expecting a front to back airflow to exhaust fresh air across it.
But out there in the real world: coolers are probably the most often modded thing on graphic cards, so...
So you need some fairly forceful directed air movement. You can rig up things, but it won't be just plug this desktop GPU into the motherboard and connect PCIe power and you're all set.
...so yeah: replacing the cooler with something more fitting for a workstation tower or a compact gaming case is probably going to be extremely common.
At worst, just some new 3D-printed shroud with a couple of Noctuas on it would probably work in a pinch,
but you can bet that AliExpress is going to be filled with cheap custom coolers for the 2-3 different most common PCIe GPU boards found in data centers.
Re: (Score:1)
Re: (Score:1)
Let's hope China saves us (Score:4, Interesting)
Re: (Score:2)
They have DRAM chips available for a while now.
Hope they ramp up the production soon.
Re:Let's hope China saves us (Score:5, Insightful)
Rumor has it some Chinese chip manufacturers are developing DRAM chips, which should be cheaper.
Do they come with optional bit-flipping technology? But QA jokes asside China has had DRAM manufacturing capability. CXMT started production of DRAM in 2018. But it takes many years to get a new facility up and running for DRAM production. If there are any facilities coming online right now they have nothing to do with AI or the current shortage.
AI is economically disastrous (Score:5, Insightful)
AI doesn't destroy jobs because of replacement, but simply because of malinvestment.
The amount of companies getting destroyed here for a bit of boom and bust in datacentre construction is ridiculous.
Re: (Score:2)
They should target AI at people who want/need them, and are willing to pay for things like a Dell GB10 which has a petaflop capability. In other words, instead of building all those datacenters and running them at public expense, let those AI consumers buy those toys that will do their needed compute for them at home, where they alone get to foot the power bill. Then they can do the cost benefit analysis and determine whether they need to continue it
Re: (Score:2)
99% sure there's already something out there that does this, but I don't have the hardware at home to run it.
Poetry for the modern age (Score:1)
The promise of AI was once
diffusion of infinite cunts.
Then they took our RAM,
so now it's a scam,
and I'm just a mem-paging dunce.
Re: (Score:2)
AI arrives
feed the beast
bubble's sound
Re: (Score:2)
I see no seasonal reference, sir.
Re: (Score:1)
may be, but if situation stay too long, all the rest of manufacturing will be halted anyway so we'll be staying with whatever hardware we have today as it will make all the parts goes up. And it is impacting lots of consumer thing. Let's hope there is solution found because the same who dont want to invest adding capacity are also selling other things which will be impacted so they may not want to keep the situation forever....
Valve keekaped themselves in the design phase (Score:3)
They went with an split design with 16GB of DDR5 on the computer (Via DIMMs) and 8 GB VRAM on a RDNA 3 dGPU (soldered) instead of going with a Unified design of 24 GB of LP-DDR5 (via CAMM2) on a SoC with RDNA 3.5.
LP-DDR5 is showing signs of getting more expensive, but not as fast as GDDR and DDR5, so, over all, less of a problem for Valve. Also, in not the same negotiation power when you buy a bunch of GDDR and another bunch of DDR5, than buying a mega-bunch of LP-DDR5
So, the GPU is hamstrung both in the Ammount of VRAM (8GB) and the Feature Set (RDNA 3 Vs RDNA 3.5). Not only that, had went CAMM2, they could in the worst case scenario, pivot from 24GB total (12 CPU and 12 GPU) pivot to a 16GB config (10GPU and 6CPU) Both more balanced for a primarily gaming machine.
All those technologies (LP-DDR5 CAMM2 and iGPUs with RDNA 3.5) were known to Valve during the design phase of the machine (Valve is a somewhat large customer of AMD to they get advanced access to roadmaps and prototypes)...
But alas, it is what it is....
But, a
Re: (Score:1)
Re: (Score:2)
yea dunno why they used such a old setup.a amd strix halo apu can win a fight with a 4060 even a 4070.
strix halo was way too expensive, even before the RAMpocalypse.
valve was aiming for "slightly above average steam hardware survey", not balls to the wall performance. That's why i has restrained in my recomended hypotetical setting.