The problem with Intel is really that they cannot keep their existing fabs at or near capacity.
That somewhat incomplete picture. The pool of shrinking fab vendors has be happening for multiple decades at this point. The plants and tech to make every smaller stuff keeps getting more and more expensive.
The problem with Intel was that the board and the executives where in denial for way too long that Intel needed to get serious about growing a foundry business. And spending too much time listening to Wall St folks who kept chirping about the AMD-spin-out model as a insanely great idea. That very nearly killed AMD in the process; in major part because that foundry had no customers at the time of the spin out (so AMD had lock-in, no choice contracts for years. And a supplier who didn't have to work hard to maintain the business. ).
Furthermore, Intel's chip designs on TSMC still have major competitive problems with AMD x86 design on same process. Even if went external fab they are still eyeball deep in problems. Again the board and upper execs more interested in stock buy-back and doing distracting hocus pocus is more so the root cause problem.
X86 has been and will continue to decline in both desktop/mobile as well as server/cloud, thus they have overcapacity today.
And then they have expansion plans (Ohio), and at the same time they are outsourcing their own chips to TSMC.
The Ohio site keeps getting assigned the next fab process 'just around the future corner'. That isn't so much capacity as it is an excuse to defer costs ( slow construction and spend less) .
The other major problem at Intel is that they are the board rowing in different directions at the same time.
" ... Intel board chairman Frank Yeary attempted to spin off Intel Foundry or even sell it to TSMC earlier this year, but faced strong opposition from Lip-Bu Tan, ...
... TSMC never wanted to buy Intel's fabs or manufacturing assets
due to both technical and business reasons. ..."
Let the internal clashes begin?
www.tomshardware.com
The Chairman is spending tons of time trying to sell "ice to Eskimos in Winter". The 'strong opposition' from Tan isn't as major a problem as the supposed buyer being largely uninterested. They would want it even less if the Ohio site was still on the original build schedule. ( The faction myopically pursuing sell off wants a smaller , cheaper deal because already at the point that nobody with chip business sense would want to buy it even at a lower price. )
Chairman
"Frank D. Yeary joined Intel Corporation’s board of directors in March 2009 ..."
2009 - 2017 track record of dubious investments and management is at the core of problems Intel has now. Botched 14nm and even more bllow 10nm was all his watch. Under investing in EUV for the sake of saving dividend payout levels .. ditto. Same board that signed off on putting Kraznich as CEO 2013-2018 ...
They do not (yet) have any sizeable customer for their new process, they are kinda cancelling 18A and betting on 14A
Not really 'kinda' cancelling 18A. They have very real product rolling out out 18A (Pather Lake ... although that likely is a 'paper launch' in 2025... 2026 product). There has been some 'doom and gloom' reporting about 18A recrently
The production process that Intel hoped would pave the way to winning manufacturing deals and restore its edge in churning out high-end, high-margin chips is facing a big hurdle on quality as it puts newer technologies to the test.
www.reuters.com
But the 'cancel' part is more so applying marketing and sales effort into trying to get customers to buy 18A. It was 20A that largely got dumped ( internally and externally). In part, that was to suppose to help free up resources so 18A could roll out more smoothly. That didn't happen. Of course firing the CEO , reorg parts of company , multiple rounds of layoffs all in the middle of that probably didn't help either.
If the yields were 10% in 2024, then yeah it was probably as hard as 'selling ice to Eskimos in the middle of Winter'.
Intel isn't 'betting on 14A' . The current CEO said they would not do 14A if they don't have a substantive external customers. If Intel cancel's 14A they would seriously need 18A+ or 18A-p or some enhancement to work.
Intel really found any very substantive customers for their older processes either. Not sure who the 'bigger fool' is going to be buy up the fab part when there is no viable runway for that business. Intel would need to go the non bleeding edge vendor to keep fab going (until it was viable to sell).
A, which, assuming it yields, would be great, but again, need volume customers to keep the fabs full …
Any funding, whether government or private, will not do anything until that customer problem is solved …
It really isn't about volume as much as it the bleeding edge is becoming more excpensive. Going down to just one bleeding edge vendor is still going to end up with expensive chips. That company will just have an monopoly (which isn't likely to get reduced chips. They will just be the last player with enough money to even try. That isn't 'volume' as much as herding all the possible customers into just one 'store' to create a largge enough potential cash pile. Those increasingly expensive chips won't be 'volume' for long. )
Even if AMD had failed to revive x86 fortunes and x86 intel was still higher demand, Intel would still be in trouble for not finding more folks to direct share the R&D costs for the most leading edge process.