Live data from Hacker News

Samsung's 60% DRAM price hike signals a new phase of global memory tightening

buysellram.com

411–420 of 449 posts

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#411

Earlier quoted context omitted.

It is a weird form of centralized planning. Except there's no election to get on to the central committee, it's like in the Soviet era where you had to run in the right circles and have sway in them. There's too much group-think in the executive class. Too much forced adoption of AI, too much bandwagon hopping. The return-to-office fad is similar, a bunch of executives following the mandates of their board, all becau…

Every corporation is a (not so) little pocket of centrally planned economy. The only saving grace is that it can die and others will scoop up released resources. When country level planned economy dies, people die and resources get destroyed.

> The only saving grace is that it can die and others will scoop up released resources.

Ideally. Realistically in market with only few companies around it makes it even less competitive.

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#412

I'm especially annoyed that this is most likely intentional. (Not at all)openAI saw they are getting behind their competitors (gpt 5 and 5.1 were progressively worse for my use case - actual problem solving and tweaking existing scripts) are getting better. (Claude and sonnet were miles ahead and I used gpt only due to lower price). Now not only open weights models like Qwen3 and kimik2 exceeded their capability and…

Small correction, Strix has at most 96GB available for GPU. But that's still a plenty

But yeah, both AMD and Intel are also pushing NPU builtin into the higher offerings so there is a very good chance that a good portion of AI will be happening closer and closer to users

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#413
post #374

Earlier quoted context omitted.

Nope, second hand is already evaporating. 2 months ago there were a load of second gen xeon scalable servers on offer. Now every one of them has had the ram stripped out and its just the chassis on offer. Fabs are not wasting their time on DDR4 now.

Nice that people are recognizing that they don't need DDR5 for most workloads. I have some DDR4 that I'm not currently using, so I just posted it for sale. :)

Not really matter of choice, just budget

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#415

I pre-ordered and picked up a framework desktop with 128GB of DDR5-8000 inside of it. This is the type of system that is the a indirect byproduct of the change towards AI - it may not have been what AM was originally intending with the AI Max 395+ line - but it definitly is the kind of optimized thinking that will drive AI into the hands of consumers. That's part of the reason I think this boom-bust cycle might be a…

> This is the type of system that is the a indirect byproduct of the change towards AI - it may not have been what AM was originally intending with the AI Max 395+ line - but it definitly is the kind of optimized thinking that will drive AI into the hands of consumers.

it literally was intended for exactly that, it has AI in the name of the cpu, and it was from the get go targeted at AI and GPU heavy workloads (3D rendering etc)

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#416

Earlier quoted context omitted.

I'm not going to disagree outright, but you're going to pay quite a bit for such a combination of single-thread peak performance and high power efficiency. It's not clear why we should be regarding that as our "default" of sorts, given that practical workloads increasingly benefit from good multicore performance. Even gaming is now more reliant on GPU performance (which in principle ought to benefit from the high PCI…

I said "single-thread performance or power efficiency", not "single-thread performance and power efficiency". Though at the moment, the best single-thread performance does happen to go along with the best power efficiency. Old server CPUs offer neither. > Even gaming is now more reliant on GPU performance (which in principle ought to benefit from the high PCIe bandwidth of server parts) A gaming GPU doesn't need all…

You have a processor with more than eight threads, at same bus bandwidth, what do you choose, dual channeled or four channeled processor.

That number of threads will hit a bottleneck accessing only through to channels of memory.

I don't understand why you brought up the topic of single-threading in your response to the user, given that processors reached a frequency limit of 4 GHz, and 5 GHz with overclocking, a decade ago. This is why they increased the number of threads, but if they reduce the number of memory channels for consumer/desktop...

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#417
post #379

Earlier quoted context omitted.

To make sure we’re on the same page… In economics, “central planning” refers to a system where a central authority (typically the state) makes comprehensive decisions about production, investment, and resource allocation across an entire economy, replacing market mechanisms. This is associated with command economies like the Soviet Union’s Gosplan system. And of course I will grant firms use hierarchical coordination…

There almost never are any markets within any single company. Which makes internals of any company a planned "economy".

I’ve worked at many companies with market mechanisms of sorts in play enough to matter. As one example, product lines that do better (as evidenced by market feedback) get more clout and revenue and command more influence. As another, there are a wide variety of ad-hoc teams that form not because of hierarchy but because people have learned who they like working with and have seen results with. These are closer to markets than C&C or hierarchical decision making.

Next, to avoid a definitional stalemate and talking past each other, let me try a different angle. If you want to predict what a company will do next, what mix of models do you use upon?

If I were to predict a company’s production choices, I would model them as being very sensitive to costs and changing dynamics to maximize profit . I would not do that when modeling a centrally planned economy. For that I would use longer time scales and do long supply chain analysis in the service of a social welfare function. See what I mean? What am I missing from your POV?

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#418

Earlier quoted context omitted.

> I think a better solution is exponential tax on a company size. I.e. once a company starts to earn above, say, 1 billion, it will be taxed by income by ever increasing amount. This is in the right spirit but you want two things to be different about it. The first is that the threshold for a given industry doesn't make sense as a dollar amount, it makes sense as a market share percentage. Having more than 15% market…

> And the second is that taxes create a perverse incentive for the government. You absolutely do not want the government to have even more of a financial incentive to sustain and create more of the companies of that size. What you want is to have fewer of them. That's not really a convincing argument. The government is the body for setting up the economic rules, it is not bound by it. The government doesn't have reve…

> The government is the body for setting up the economic rules, it is not bound by it.

When a new law is proposed, the Congressional Budget Office prepares a report on the impact it will have on the budget.

Now suppose a new law is proposed that will remove an existing unfair advantage of large companies over small ones, causing more small companies to form and take market share from incumbent larger ones. If large companies pay a 50% tax rate and small companies pay a 10% tax rate, the CBO analysis will show tax revenue going down. Then in order to make up the shortfall at a given level of deficit spending, the government would have to raise taxes or reduce spending, both of which are unpopular, so instead the bill gets tabled and the huge companies retain their unfair advantage. That's the perverse incentive we don't want to see.

> Money is created by the government, it doesn't have a value yet.

If the government can create an unlimited amount of money with no drawbacks, why don't they just send everyone a check for a trillion dollars? If they can't then whatever they want to spend in excess of what they can get away with printing or borrowing has to come from tax revenues, and then what happens when you set up an incentive structure where the government gets more money to spend the bigger they allow companies to get?

> Sure, there are a lot of idiots now-a-days, that think a state should work like a business and make profits, but they are misled.

This is a straw man. The only people who think the government should make a net profit are the people trying to build some kind of sovereign wealth fund. The US government isn't doing anything even resembling that -- it has been running massive deficits for decades. It's to the point that interest on the debt is now a major component of the budget -- we're now spending about as much on interest as on Medicare.

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#419
post #324
post #300

Earlier quoted context omitted.

I am so glad both top rated and majority of comments on HN finally understands DRAM industry instead of constant DRAM is a cartel that is why things are expensive. Also worth mentioning DRAM and NAND's profit from Samsung is what keep the Samsung Foundry fighting TSMC. Especially for those who thinks TSMC is somehow a monopoly. Another things to point out which is not mentioned yet, China is working on both DRAM and…

DRAM manufacturers have literally been convicted of price fixing in the past why do you have to white knight for them?

And I am 100% sure a lot of other industries in commodities would have been convicted of price fixing if we look into it. And I say this as someone who have witnessed it first hand.

Unfortunately commodity business is not sexy, it doesn't get the press, nor does it get told even in business schools. But a lot of the times these call called price fixing is a natural phenomenon.

I wont even go into what get decided in court doesn't always mean it is right.

I will also add we absolutely want the DRAM and NAND or in fact any industries to make profits, or as much profits as it could. What is far more important is where do they spend not those profits. I didn't look into SK Hynix but both Samsung and Micron spends significant amount of R&D at least try to lower the total production cost of DRAM per GB. We want them to make healthy margin selling DRAM at $1/GB, not losing money and then go bankrupt.

Re: Samsung's 60% DRAM price hike signals a new phase of global memory tightening

#420

Earlier quoted context omitted.

I'm not going to disagree outright, but you're going to pay quite a bit for such a combination of single-thread peak performance and high power efficiency. It's not clear why we should be regarding that as our "default" of sorts, given that practical workloads increasingly benefit from good multicore performance. Even gaming is now more reliant on GPU performance (which in principle ought to benefit from the high PCI…

I said "single-thread performance or power efficiency", not "single-thread performance and power efficiency". Though at the moment, the best single-thread performance does happen to go along with the best power efficiency. Old server CPUs offer neither. > Even gaming is now more reliant on GPU performance (which in principle ought to benefit from the high PCIe bandwidth of server parts) A gaming GPU doesn't need all…

What is the best single thread performance possible right now? With over locked fast ram.
Post reply on HN