For Intel to succeed as a foundry it needs customers that target the same "high end, power hungry" market segment. I don't see how Qualcomm (low power) fits that niche. More likely big AI accelerators (like, perhaps, Microsoft is planning).
As long as Intel depends on its high end processors for most of its profits, it will be difficult to develop a low power process for other, less profitable, customers.
There's no nearly-conceivable future where the minimum costs to design a 28nm or below chip are tractable for most small companies, but even most capable individuals can design something on 130nm and afford to have that taped out today.
My bet is that TSMC will go down and the talent will either go to the mainland or the US.
Source?
Some sources: Harvards economic complexity US export of computers in 2021: https://atlas.cid.harvard.edu/explore?country=undefined&quer...
Chips Over Oil "Fact checked": https://www.techopedia.com/chinas-strategic-pivot-in-the-glo....
Just check average .... X....
I'm not sure if China, in decline, is worth mentioning. Eg. Average wage, GDP, ...
The west is not the US alone...
I regard US + EU + Canada + Aus + Japan + South Korea as the west about 1 billion people.
I'm probably missing some here, and there are almost certainly some unannounced customers.
TSMC and Samsung can’t meet the demand of those companies.
If Intel can be TSMC’s customer then surely AMD can be Intel’s customer, so is Apple. Intel’s foundry business is structured to be separated from its processor business
The assumption is that there is going to be a lot more demand. Even after cutting China out which is like 1/3 of global demand. The assumption is that will be driven by AI, mostly in data centers.
Could be. Otherwise Intel foundry services will face an uphill battle vs. TSMC.
TSMC’s revenue is like 70 times of Intel’s foundry business. Intel only plans to leapfrog Samsung by the end of this decade. What Intel aim to lead is the process. They plan to introduce 18A by the end of this year, followed by 14A.
China will not be cut off from the global market, only at the very high end. Nvidia just recently showcase their latest chips to Chinese market. They will continue to be a main driving force along with the US.
They will if the alternative is that they lose customers. It’s not like they have a lot of options. Also they can use Intel for less advanced chips.
Intel knows full well that their chip competitors will be skeptical. That’s why they announced today that the company will be split into 2 distinct groups - Foundry and Product.
Microsoft and ARM join Intel on stage at their foundry event. The chips they have Intel to make are definitely competitors to Intel’s x86 and GPU businesses. That’s what AMD will do unless they get all they want from TSMC or Samsung.