Thread regarding Intel Corp. layoffs

If the High NA EUV rumors are true, then that should accelerate the ramp down of internal non-EUV production.

All that will be moved to other foundries, and more than a few fabs and fab sites will be shut down or sold.

Intel still can not make chips beyond 10nm++++. 7nm, 5nm, 4nm, 3nm programs were all stopped in the middle. Directly jumping to 18A have not shown any result yet after 5 years. 14A is just a boast of the cheating managers in the distant 5+ years in future.

@mb That is not to say the company should turn down any IFS opportunity to produce memory, but the major memory companies have their own fabs.

Might be room to use foundry to help one of those companies to deal with excess demand.

@k4 Every memory cycle, Intel and others see the huge profits and try some new approach to 'do memory'.

Then the price collapses and they say, D'oh!

Intel has done this a number of times and isn't the only company who keeps getting lured in, only to see how much scale is needed to stay in that business.

The company has brought a number of innovations to the memory business over the decades, but none of them last more than 1 business cycle. That is how brutal the memory market is and will likely always be.

Use the fabs to produce memory chips. There's now a large open market for memory to be used in personal computers. When one door closes, another opens. When opportunity presents itself...

signs of a dying company 101 along with its own CEO LBT using the company to profit himself by buying his own start ups.

Intel has no growth, no investment into research, just cost cutting left and right and outsourcing. Any fool with business knowledge knows to stay away from investing their stock money in here. Even Elon invest into research more than LBT.

@gc However this plays out, it seems likely that LBT will choose to shed the older fabs rather than upgrade them.

The exception might be if, as you noted, contract demand somehow explodes. It might be easier to meet that demand by retrofitting some of the newest non-EUV fabs, and that would almost certainly mean building new mods to house EUV.

This is kind of like what happened in the shift from 8" to 12", with the new factor being the possibility of a demand curve that is far steeper than anything the company could ever create with x86.

Short of that, the clock is ticking for any fab which does not already have EUV.

LBT likes to maximize virtualization, by outsourcing as much labor and supply as possible. When done right it creates a lot of value and he seems to know how to make that happen. Lots of people chatter about how contract employees will take over blue badge roles, and all that is true. Few people post about what is already happening, which is the shift to external production for packaging and wafers.

The idea is to maintain a core competency, which can serve a reasonable amount of demand, then contract out everything above that base level of capacity. That enables the core company to continue to get smaller even as demand increases, and makes the whole enterprise more efficient, effective and profitable.

This is about 180 degrees away from IDM 1.0, and is the best path forward by all appearances.

People in this kind of org will increasingly have to produce real value, or one by one, they will be pushed out. The pressure will build over time and be relentless.

@fy The company has been working with UMC on 12nm for some time now.

A company like GF could do a multifab deal, but more likely is some mix of selling fabs where they are not co-located with facilities that Intel still needs and the mothballing of some fabs where it is not feasible to sell them.

So you might see a fab or two sold in the next few years, then others shut down as those ramps end.

It might be possible to add EUV mods to some of the older fabs, but Ohio likely gets built before then, and so the demand would have to be significantly higher than anything expected in the next few years.

When or if IFS is able to attract high volume contract business, then it seems likely that other mega cap tech companies would at least second source with Intel. So you get no-thing, then 1-thing, then a flood.

@fy Tower Semi.

Who can be the buyer of old fabs? GLOBALFOUNDRIES? Texas Instrument?

The company is in the process of moving all older node production to other foundries.

It is true they have always used various foundries for older node production, but they are now more aggressively shifting away from internal production of older nodes.

Watch and see them sell off the older fabs and maybe a few entire fab sites.

Nothing radical about this, but anyone working at the older fabs should make a decision about whether to stay or not, before that decision is made for them.

it's coming..

Please... This is nothing new.
Samsung has been building Intel desktop PCHs for quite a while.

Intel still can not make its own chips.