Do you think Intel could jump into the memory business with the new gubmint sponsored fabs and help out so the PC environment doesn't hit the skids?
15 replies (most recent on top)
@zh Remember that time when VP Anne K. tried to leapfrog?
Once seen, some things can not be unseen.
@OP Sorry, no jumping is allowed in the fabs.
Safety First.
No bonus money 4 u after failed CES out takes.
Any business need good execution and deep technology. While Intel lacks both, it will fail in any business.
Intel managers are so low in technology, they only use PPT or excel at most, they are driving blindly down the cliff.
Intel wasted over $10 billion on emerging memory over more than 10 years mostly in Lehi Utah. No please don't bring this up again.
Nope.
@dm Samsung has managed to keep going with memory, but they run it as a separate business, not as something to fill in volume at fabs which are mostly producing compute devices.
Read about how the ramp issues on the latest nodes have nearly driven Samsung into a capital breakdown situation as has happened to Intel. It doesn't take more than a couple of failed nodes for the entire economic model to fall apart.
@db The poor decision was to keep trying to do memory.
There is a reason why memory is only sustainably produced by companies where that is the only product.
Memory is a big fat R&D pipe which must lead to high volume production in order for the capital cycle to be sustained.
It can not be done as a side line.
Packaging is a similar situation, where the capital needed for R&D can only be recouped if the company is one of the top package foundries.
The compute business is not so extreme, although it too requires large scale production over a relatively short period of time in order to capture the highest ASP possible in order to be profitable.
Since 14nm, Intel has demonstrated what happens when the ramp is too slow, and so misses the highest ASP. It then becomes very difficult to make enough money back in order to fund the next node.
This is referred to as a capital cycle breakdown. In the dictionary, a picture of BK is shown next to the description.
As node progression accelerates, the inability to rapidly scale the ramp results in very poor gross margin and eventual failure of the business model. Memory has had rapid node progression for some time now, and it is speeding up for compute devices as well.
Intel has the memory fab in Dalian Fab68. It was built for many years (10 years), but was sold in 2025 to SK Hynix. This is a very stupid decision by the stupid Intel manager again. So many single-point-of-failure decisions by Intel mangers.
@a7 actually that’s a good question. Let’s say you’re Samsung. You already have fabs with a leading edge DDR5 DIMM process and a new greenfield site that has some tools in it (but not necessarily the right mix). How soon can greenfield crank out memory in high volume?
My swag is two years minimum. Before you ask where Intel’s DRAM process is.
The gubmint couldn't find its own technological behind with 2 hands and a map, so with Inbed being the same way I don't see the memory thing as being viable. Maybe snow cones on a good day.
Yes. Intel could, but they also can't.
Intel is already poised to make a ki-ling in memory by Fabbing Optane Persistent Memory out of Fab 68 in Dalian China
@a2 Please do note what asset class wildly outperformed all others in 2025: precious metals.
The only more vertical chart is $jpt10Y, and that is not a coincidence.
Precious Metals could be referred to as non-digital Bitcoin, or think of it as land which can be moved around as needed.
As BTO used to say, 'You Ain't Seen Nothin Yet'
This question was answered in some detail on another thread.
Short answer is that Intel could potentially pick up some contract business, if one or more memory companies run out of internal capacity.
But see how long the excess demand remains, if or when the economy slows down.
The gubmint is still force feeding the economy with trillions of new debt, but the effectiveness of that is slowly declining. That is really the only question which needs to be answered, as all other economic issues flow from the rate of growth in the national debt.
Everything else is a derivative of that root factor.