Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
It depends on who is designing the chip (for TSMC) or fabricating (for IBM). It also depends on if IBM's partner fabs can mass produce any chips at 2 nm.

Comparisons here for transistor density: https://www.anandtech.com/show/16656/ibm-creates-first-2nm-chip

IBM 2 nm should have about 333 MTr/mm2 versus TSMC's 171 MTr/mm2 at 5 nm.

I've followed and invested in foundries for a long time. These 'lab' chips with blistering transistor densities are nothing new. There's a huge difference between what's achievable in a lab and what can be massed produced economically (i.e. yields) in vast quantities. You see these stories every couple of years.

TSMC will be pouring biblical sums of money into R&D for their N2 with GAA and beyond processes. The big breakthrough has been EUV.
 
Amazing!! this should drive alot of technology in the US
why is that?

The U.S. already has a bunch of 2- and 3-generation old fabs. Why will another one make a difference 3 years from now.

well if technology taught us anything is to never say never..

IBM doesn't have fabs. It doesn't transfer its processes to commercial fabs. It doesn't have a full 2-nm process to transfer.

I'll stick with "never."

There must be 100 "chips" in the typical cell phone and 1,000 in the typical computer. But all of them are not the CPU chip.

Take even a $10 generic mouse apart and look inside, there are a lot of "chips" and each one of them is made in a "fab". Most of these sell for a few cents. and don't make the news. Less then 1% are CPU chips. You don't waste a <5 nm process on a 5 cent part.

yes, I'm aware of that. And there are lots of fabs in the U.S. that are already 2 or 3 generations behind the cutting edge node. So why would this one make a difference?
 
  • Like
Reactions: Captain Trips
No way they build the plant that quickly. The intel plant, also in Phoenix, took almost 10 years to complete.
 
People are talking about how this plant will be producing outdate product by the time it comes online. That may very well be the case, but I have to suspect these plants don't hit the ground running making the absolute cutting edge stuff right out of the gate. It takes time to tweak the operation and get it just perfect. Once the basics are in place, they will be able to make the latest stuff here.

That’s not how it works. If you tool a fab for 5nm, then to get to the next smaller process node you spend billions of dollars to re-tool it (replacing most of the machinery, etc). You can’t build a fab and then “tweak it” to get to the next process node.
 
I love seeing how much AZ is growing into a major tech hub for the US. The housing market here is insane. Sucks for those who want to buy a home but its great for those of us that own homes here.
 
Last edited by a moderator:
5nm in 2024? So chips for old stuff.
Maybe it's less risck to TSMC to dip their toes with proven technology... It's not like $12 billion is an insignificant 'dip' into the US supply chain. TSMC wouldn't do this without forecasted demand, no?
 
That’s not how it works. If you tool a fab for 5nm, then to get to the next smaller process node you spend billions of dollars to re-tool it (replacing most of the machinery, etc). You can’t build a fab and then “tweak it” to get to the next process node.

Sounds like I don't know crap, then....
 
Apple keeping things home in the US. Love it
Nowhere does TSMC even say that these chips will be for Apple. They're probably for other customers who don't demand the latest and smallest die sizes and performance the way Apple does.
 
Maybe it's less risck to TSMC to dip their toes with proven technology... It's not like $12 billion is an insignificant 'dip' into the US supply chain. TSMC wouldn't do this without forecasted demand, no?

Not really relevant to my point, is it? The reason for the bad thing doesn’t mean it is not a bad thing.
 
It's relevant in that the opinion of 'bad' is in the eye of the beholder.

But it was not an “opinion.” By the time the fab starts producing chips, the process node in this fab will be at least 2 generations behind the lead node. By that time, other U.S. fabs will also be running on an equivalent node. Responding to “that will be old technology” with “but…but…reasons!” is facially irrelevant.
 
why is that?

The U.S. already has a bunch of 2- and 3-generation old fabs. Why will another one make a difference 3 years from now.
Do you mean intel Fabs? having this production should drive ARM adoption in PCs and allow more devices access to faster ships aka linksys and others
 
Don’t know that guy.

How about this one?

Looking good Cliff!

Posted my Ryzen setup on craiglists after reading it, thanks for the heads up.

1626373043646.png
 
New jobs in US! This is wonderful news. Who’s ready to apply with me?

From Apple's perspective - its smart in case relations with China deteriorate. They have the design and engineering facilities in the US - now they have manufacturing capability here as well.

I see an investment into TSMC eventually making this relationship closer.
 
From Apple's perspective - its smart in case relations with China deteriorate. They have the design and engineering facilities in the US - now they have manufacturing capability here as well.

I see an investment into TSMC eventually making this relationship closer.

Apple’s chips aren’t manufactured in china, so how does the future ability to manufacture some chips in the US help in the event relations with china deteriorate?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.