Earlier quoted context omitted.
That is not even close to correct.
I'm very interested to hear what exactly I said was incorrect.
Intel made bad engineering decisions. There was not under-investment in fabs or R&D.
Memory has grown to nearly two-thirds of AI chip component costs
521–524 of 524 posts
Re: Memory has grown to nearly two-thirds of AI chip component costs
#522Earlier quoted context omitted.
Bankruptcy risks. When Samsung had to sell memory at a loss after COVID, no one came to save them. They buffered their memory division using profits from their other businesses. That’s how Samsung survives memory downturns. According to some stories, this is how Samsung convinced TSMC to not enter the memory business - that you need a nation or other lines of business to prevent bankruptcies. The market has stabilize…
...And why do they go bankrupt? Because it's an incredibly capital intensive process, involving billions of dollars of investment into manufacturing infrastructure. That is to say, making memory is quite hard.
Making memory is easy in the sense that any company that makes JEDEC compliant DDR5 can sell you their product and you can put it into your mainboard (soldered or not).
You can't do that with Intel or AMD CPUs. There is no spec. Each vendor does CPU socketing and chipsets in their own unique way.
Not to mention that your design needs to be good. For DRAM manufacturers, they just need to design one really good cell design and spam it a billion times.
The problem is that DRAM is a specialised product that has a specialized process. Your fab is a one trick pony that can't take advantage of other markets. Since you're stuck with fixed costs on a single design, you're cranking up the volume to make your money back, but everyone does that and if everyone does it, then the market is flooded with cheap memory that won't recoup the costs.
Now that AI memory is something the companies can specialise in to increase the margins, they can exit the contested market with low differentiation. If anything, they will prefer it of a single company is left doing consumer DRAM, because it is one less company in the AI memory space.
Re: Memory has grown to nearly two-thirds of AI chip component costs
#523Earlier quoted context omitted.
why not? i'm sure they can jump into the hustle. Increasing the availability doesn't mean decreasing the price ... people think those are intrinsically related - not so much. You can get a prada shirt for $2,000 ... as many as you'd like, for $2,000 a piece. No problem. They'll make the factories go burr all night long. Still $2,000.sweeping There's a bunch of things like this. $100 bills for instance ... a new entra…
> why not? i'm sure they can jump into the hustle. Not so quick. Critical difference is the relationship between enterprises and the state. In China, the state owns the enterprise, in one way or another. High costs of memory is a threat to the established Chinese electronics manufacturers. The Chinese state can optimize returns at a higher level than the one some petty chip manufacturer operates at, especially if doi…
I mean kind of? Industry and politics are always cohorts. China has structural differences but new entrance to commodities and defensives are almost exclusively price sensitive offers. If say, Singapore or even a firm in Ohio tried to enter the market for global play, they'd undercut as a strategy
Except... Maybe they don't need to. Demand is outstripping supply right now. Competing on availability may be sufficient
You can easily see this on a mini-macro scale with popular restaurants. Often a restaurant with an equivalent menu and prices will open nearly adjacent to a very popular place and can sustain itself simply because you don't have to wait an unacceptable length of time to be seated
Re: Memory has grown to nearly two-thirds of AI chip component costs
#524Earlier quoted context omitted.
I'm very interested to hear what exactly I said was incorrect.
Intel made bad engineering decisions. There was not under-investment in fabs or R&D.
Yes, their primary "bad" decision was to try to squeeze as much as possible out of DUV so that they didn't have to spend a ton of money upgrading to EUV.
This was a business decision, not an engineering decision.