News TSMC Arizona allegedly now producing AMD's Ryzen 9000 and Apple's S9 processors: Report

Isn't Intel doing 3nm? Thus, Intel's product is far more power efficient; going into 18A, no one can compete, not even ARM's product, like Apple's
 
This is one reason I think AMD might've opted to use N4 for Zen 5 CCDs and maybe even RDNA4 GPUs. I wouldn't go so far as to presume it's why Nvidia stuck with an N4 node for Blackwell, but maybe.
Does this have anything to do with evading tariffs? Or just companies investing in a node that has lots of (domestic and otherwise) availability?
 
This is one reason I think AMD might've opted to use N4 for Zen 5 CCDs and maybe even RDNA4 GPUs. I wouldn't go so far as to presume it's why Nvidia stuck with an N4 node for Blackwell, but maybe.
Nvidia did it to save money. Same reason they are rumored to be looking at Samsung for their next architecture.

 
Isn't Intel doing 3nm? Thus, Intel's product is far more power efficient;
Intel is believed to be using N3B, which is a first-gen 3nm TSMC node. The efficiency data between Intel's Lunar Lake & Arrow Lake vs. others is out there, if you look for it.

going into 18A, no one can compete,
That's going to be Panther Lake, initially. We shall see how it compares.
 
Does this have anything to do with evading tariffs?
I hope Taiwan isn't going to get hit with tariffs, but we'll see. I think the main benefit is securing availability, in the event that Taiwan goes offline.

Based on what's been reported about TSMC charging more for wafers from their AZ plant, I think they'd still be more expensive, even if the N4 wafers from Taiwan got a 10% markup added to them.
 
ok there are a fab in usa, one in japan, the prices should go down now, right ?
Assuming it's producing supply everyone wants, maybe. However, non-Taiwanese plants are all n-1 (meaning at least 1 generation old node) and the cost of the doing business in the US and Japan is typically higher than Taiwan.

Given those factors and the fact that TSMC is unlikely to create a supply glut against itself, I wouldn't expect prices to change simply because of those two plants (remember TSMC was planning on adding more capacity simply to handle their current order flow). Now if Intel and Samsung can start competing on leading edge nodes, then there is a chance of getting reduced prices.
 
Isn't Intel doing 3nm? Thus, Intel's product is far more power efficient; going into 18A, no one can compete, not even ARM's product, like Apple's
Intel is using TSMC for their consumer CPU's and GPU's. Intel 3 is for their server chips, so you'd have to compare power/efficiency to AMD's EPYC chips.

18A: "no one can compete" could be a bit of a stretch. Even if that ends up being true, it's temporary; a lot of Intel's success will depend on 1) how quickly they can ramp up 18A and have decent yields and 2) how quickly Intel can move on to 14A and do the same thing. Apple will be able to compete simply on the basis of having their closed, proprietary, and optimized hardware-software ecosystem; they don't necessarily even not to have the most advanced node of the day. ARM arch is part of that equation -- I wouldn't discount that aspect.
 
I hope Taiwan isn't going to get hit with tariffs, but we'll see. I think the main benefit is securing availability, in the event that Taiwan goes offline.

Based on what's been reported about TSMC charging more for wafers from their AZ plant, I think they'd still be more expensive, even if the N4 wafers from Taiwan got a 10% markup added to them.
I was referring to the universal 10% tariffs, but I guess its just business.