Here's a wild prediction, take it with many grains of salt. By 2026, foundry becomes independent entity, sort of an American tsmc. It will also change name, not Intel anymore.
The design and product will be sold to Qualcomm. Intel gets dismantled, the brand disappears
Sadly, @d9+1k00kzqpp might be onto something.
Yes. Intel will not exist just like Lucent Technologies.
It’ll be broke up or sold for parts by end of 2026. Need to get headcount down to 50k max, will be quite profitable and good price for shareholders. Employees are just pawns.
This will likely get lots of downvotes from my for peers but we all know we employees at Intel have our heads in the sand (or A-s) and think we are great stilll
@ax During the split they made some choices to push certain expenses to IFS and not ProdCo, and certainly weren't trying to make IFS look better.
The (delulu) thought might have been that a few external customers were bound to happen soon, and so IFS could eat more cost and make ProdCo look more impressive.
Or it it could be that IFS was already deep in the red and they decided to throw everything but the kitchen sink in there (and make ProdCo look more impressive).
As with any corporate decision there is room to argue for or against. Bottom line is that the expense treatment being handled any differently would not have made IFS wafer cost competitive with TSMC.
With 18A and 18A-P, IFS is working through how to build based on industry standards and discard many of the custom design, tool mods and recipes that it tinkered with in an attempt to make CPU more competitive. All that likely never made a difference or was merely to offset other deficiencies.
Moving to industry standards will lower wafer cost, and with the current fab reductions the company is getting a bit more serious about labor cost as well.
AI and robotics are the holy grail, with mass outsourcing as the transition. Then any geo differential disappears or even shifts in favor of fabs in the US.
I would bet more on Broadcom than Qualcomm but other than that I totally agree
In recent years (that I know of), foundry losses only existed because management wanted to make prodco look better. If foundry ever really charged what it cost, prodco would likely look like sh-t. And, I am not saying foundry would look any better
Then, and only then, would we get a real view of where foundry sits against TSMC and can judge accordingly.
@af I agree, the IFS losses are a problem which must be solved.
Breaking out IFS from ProdCo was the best thing the company has done to expose to itself and the market what is really going on at the company, and it is ugly.
The company set up TD and the fabs as a loss leader, with a relentlessly pursued priority of output over cost. They did this because that was what ELT thought was needed to keep market domination going, but that just highlights the issue with the IDM model and why Fabless + TSMC has rapidly proven to be more efficient.
By breaking out the IFS results, Intel was admitting that the IDM model was no longer tenable. It was forced to realize this due to the failure of 10nm, on cost and output. Now that reality-based management has been allowed to prevail, LBT and the Board are in the process of fixing the original mistake, which was how TD and the Fabs are managed.
Expect to see IFS downsized, outsourced and at least partially sold off, as the company focuses on eliminating the losses as quickly as possible. The previous hopium is discarded, thinking that a few big customer wins would prevent the company from having to effectively manage IFS. Reality is that ramping a foundry takes years and it is really hard to get any major customer to switch from another foundry that is performing well.
So expect the hits to jut keep coming in IFS, till it is in the black for the first time..ever.
@ah that doesn’t mean it can’t be used as part of the name of another company. This why Foundries was included in the name.
@a6
"Universal" is already the name of the company which owns NBC Television
The current effort should enable the company to create the option of spinning off (an at least marginally profitable) IFS and/or selling what then would be more profitable Product groups.
Strategy seems to be to clean up the leftover efforts from when the company dominated client and server markets, and become a more focused conglomerate.
Then the option to stop being a conglomerate becomes a real thing. IFS losses make it difficult to get a decent valuation if it were spun off or sold now, and IFS is also not yet set up to operate under industry standard design rules and PDK (until 18A-P), so no other foundry would have much interest in IFS at the moment.
Because of the ongoing IFS losses, the company needs ProdCo profits to remain solvent, so getting IFS profitable is Job #1. Absent external customer growth, the only other option is to reduce (labor) cost, so expect outsourcing in the near future.
Cash burn can also be reduced by selling off the older fabs or some entire sites, and I expect something to happen on that front as well. Maybe something big.
There is also room to make ProdCo more profitable and that is happening, by removing all the unprofitable groups which were providing products that were not considered to be the core business.
These ongoing efforts seem to have LBT aligned with the Board and any ELT opposition solved by pushing them out of the company, so expect a continued, steady stream of changes as the company makes decisions about how much products it will continue to offer, and then acts on those decisions instead of the usual procrastination.
Then in a few years many options become possible, but that is because the business has been right-sized. At that point, there would be little motivation to break up the company. Breaking it up only seems to make sense right now, because of the IFS losses.
TLDR: It can't be broken up right now and once it can, there will be little reason to do so.
good prediction this some variation will be true. late 2026 after company rightsized and more profitable. we are just window dressing right now for the great sale
I predict def met will survive untouched despite various actionable events
I agree. I felt that this was going to happen when they changed our name from AZFSM to OTF a few years back, except that I think that Intel will still exist, but only to design the x86 CPUs that someone else builds.
This has been predicted before. How about Universal Foundries for the new fabco? It is a grander scale than Global Foundries.
@a1 it doesn't need to have Intel's brand to support legacy customers and client market share is slipping fast.
I don't believe the brand will disappear. Intel still dominates the PC market and there are too many customers who still need to be supported. The CPU business would become a subsidiary of another company.