• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

News [Bloomberg] Samsung Considers $10 Billion Texas Chipmaking Plant, Sources Say

NTMBK

Diamond Member
Nov 14, 2011
9,141
2,367
136
Samsung Electronics Co. is considering spending more than $10 billion building its most advanced logic chipmaking plant in the U.S., a major investment it hopes will win more American clients and help it catch up with industry leader Taiwan Semiconductor Manufacturing Co.

The world’s largest memory chip and smartphone maker is in discussions to locate a facility in Austin, Texas, capable of fabricating chips as advanced as 3 nanometers in the future, people familiar with the matter said. Plans are preliminary and subject to change but for now the aim is to kick off construction this year, install major equipment from 2022, then begin operations as early as 2023, they said. While the investment amount could fluctuate, Samsung’s plans would mean upwards of $10 billion to bankroll the project, one of the people said.

If it starts cranking out chips in 2023, it'll be about 2 years behind leading edge- but could probably still pick up a lot of volume. Interesting implications for security-sensitive chips for the US government, too...
 

Ajay

Diamond Member
Jan 8, 2001
8,158
3,103
136
Damn paywall. What customers are they targeting? Screwing up 7nm EUV is a pretty big black eye for them - hard to see how they would get any large volume orders anyway.
 

ao_ika_red

Golden Member
Aug 11, 2016
1,674
712
136
So I guess TSMC's America expansion was not just lip service then.
Still feel gutted when GlobalFoundries (IBM) shelved their 7nm effort. I imagine those execs are now salivating over huge order TSMC and Samsung enjoy.
 

Ajay

Diamond Member
Jan 8, 2001
8,158
3,103
136
From the con call:

Pat Gelsinger
: Majority of our [client] products in 2023 will be on Intel’s 7nm, but we will have increasing use of external foundry. Intel is committed to building a competitive product with leadership in the marketplace. This means packaging, software, external and internal fabs, and we are confident in delivering a product leadership across all categories. Long term innovations are coming out of research as we look to close gaps and leap ahead of external foundries. Unquestioned leadership in process is our goal.
Samsung will be on their 3N GAA, so IDK WTF is going on. Sounds like, process wise, Intel is still stuck in a quagmire.
 
Mar 11, 2004
20,846
3,009
126
Considering sounds like fishing for subsidy or tax breaks from the US government. Which...wouldn't be awful really, although I'd prefer if they poured money into advanced research and then licensed it or offer different things to spur it in the US (thinking stuff like education subsidies, maybe offer subsidies for foundries in exchange for job goals or something).

It would've been nice if we'd had something like that in place already, it might've saved GF's advancement and maybe even helped Intel's struggles.

So I guess TSMC's America expansion was not just lip service then.
Still feel gutted when GlobalFoundries (IBM) shelved their 7nm effort. I imagine those execs are now salivating over huge order TSMC and Samsung enjoy.
I think it was less lip service and more was a light start to gauge the overall response (US and China specifically).

Doubt it. Samsung seems to be all but floundering and I wouldn't be surprised if they have been losing billions on manufacturing.

I think we're getting to where the next stages of chip manufacturing is going to require nation state investment. The EU announced pretty hefty investment (wasn't it almost $150billion?) for getting to 2nm.

If the US were to announce some program for that as well, possibly tied to different programs, we might would see some big changes. It could even get some players with deep pockets that wouldn't typically be delving into this, like Apple, Microsoft, Amazon, and others (most of them already have chip design teams). Plus auto companies might get involved as well as they're currently experiencing chip crunch limiting their production, and that's going to be even more important in the car industry moving forward.
 
  • Like
Reactions: Ajay

gdansk

Senior member
Feb 8, 2011
590
252
136
Samsung Electronics foundries are doing OK in 2020. Their revenue grew each quarter (up 25% YoY in Q4, for example) but market share declined. I'm not sure that is permanent as their 7nm was too late so it had few designs and was not a in good position to take advantage of the pandemic orders. Their 5nm seems to be a decent improvement (+20% ish) delivered on time and TSMC has capacity problems so they have some large orders.

Main topic: it isn't even new for Samsung to have their state of the art in Austin, wasn't that the case a mere 6-7 years ago?
(Edit: Yes, Austin was Samsung's first fab to produce 14nm chips, in January of 2015)
 
Last edited:
  • Like
Reactions: lightmanek and Ajay

Ajay

Diamond Member
Jan 8, 2001
8,158
3,103
136
Come to think of it that might be why Intel isn't officially giving up - they are looking for that sweet Gov $$$.
Yes. Sort of some not so subtle subtext here:

Pat Gelsinger: It is incredible to come back home. This is my dream job. There are four areas I will focus on.
...

Finally it is also a point of Intel’s culture. I trained at the feet of Grove, and we will have the Grovean attitude to execution. This will be data driven as we rebuild the company. The best days for Intel are in front of us. This is a priority for Intel, the industry, and our nation.
 
  • Like
Reactions: Thunder 57

Thunder 57

Golden Member
Aug 19, 2007
1,605
1,609
136
Come to think of it that might be why Intel isn't officially giving up - they are looking for that sweet Gov $$$.
Money is fine and good, but I think Intel's problems have more to do with management and/or lack of talent. I'd go into it further but that would start this thread down a bad road, so that's all I'll say about that.
 

moinmoin

Platinum Member
Jun 1, 2017
2,286
2,742
106
Yes. Sort of some not so subtle subtext here:
I expect Gelsinger to talk about nation (and faith) more often.

The funny part is, AMD is already executing on that in a way Intel isn't. DARPA's Fast Forward initiative running since 2012 is what eventually led to both Frontier and El Capitan, and Zen and CDNA could be considered offspring of that whole long term development.
Intel meanwhile has Aurora in that very same initiative which was originally based around the Xeon Phi product line and was planned to be done in 2018 before it was restarted from scratch to be finished this year. Let's see how that goes.
 
  • Like
Reactions: Tarkin77

DrMrLordX

Lifer
Apr 27, 2000
16,878
5,841
136
Back on the subject of Samsung:

It's odd that it's Samsung attempting to expand their footprint in the United States and not TSMC. Samsung's position is fairly geopolitically stable, but TSMC . . .
 

Failnaught

Junior Member
Aug 4, 2008
15
9
81
Samsung's position is fairly geopolitically stable, but TSMC . . .
Taiwan will be fine.

It's hopeless to militarily capture Taiwan. China will face serious diplomatic isolation afterwards. I can think of only Russia which openly annexed Crimea, and there was a huge uproar. Among people who actually live in Crimea, they overwhelmingly supported it. This won't be the case in Taiwan. Also, try running a fab (and an internationally integrated, export oriented economy) after every country around the world shut their doors. It just won't happen. TSMC will be worthless to the Chinese if they tried something so clumsy.

Also, in the ancient text Art of War, it is stated that war is the absolute last choice. It is much better to politically integrate an adversary, or have foreign powers fight one another, than to directly engage in hostilities.
 

DrMrLordX

Lifer
Apr 27, 2000
16,878
5,841
136
Taiwan will be fine.

It's hopeless to militarily capture Taiwan. China will face serious diplomatic isolation afterwards. I can think of only Russia which openly annexed Crimea, and there was a huge uproar. Among people who actually live in Crimea, they overwhelmingly supported it. This won't be the case in Taiwan. Also, try running a fab (and an internationally integrated, export oriented economy) after every country around the world shut their doors. It just won't happen. TSMC will be worthless to the Chinese if they tried something so clumsy.

Also, in the ancient text Art of War, it is stated that war is the absolute last choice. It is much better to politically integrate an adversary, or have foreign powers fight one another, than to directly engage in hostilities.
That would assume that logic were the only force in play.
 

ThatBuzzkiller

Golden Member
Nov 14, 2014
1,010
156
106
Taiwan will be fine.

It's hopeless to militarily capture Taiwan. China will face serious diplomatic isolation afterwards. I can think of only Russia which openly annexed Crimea, and there was a huge uproar. Among people who actually live in Crimea, they overwhelmingly supported it. This won't be the case in Taiwan. Also, try running a fab (and an internationally integrated, export oriented economy) after every country around the world shut their doors. It just won't happen. TSMC will be worthless to the Chinese if they tried something so clumsy.
The CCP will try to prove you otherwise since they're still ideologically flirting with Chinese reunification ... ;)

Also, mainland China is changing from an export oriented economy to a domestic consumer oriented economy so capturing TSMC's trade secrets by force would still prove to be valuable. The only nations that would blockade China are liberal democracies but that'll become far less effective in the future once CCP successfully transitions the country to it's desired economic model since they aren't interested in just being content with the middle income trap ...

I think your conceptions about China are highly outdated ...

Also, in the ancient text Art of War, it is stated that war is the absolute last choice. It is much better to politically integrate an adversary, or have foreign powers fight one another, than to directly engage in hostilities.
Taiwan won't be able to float anymore by itself in a few years once they've milked out the maximum potential behind silicon based MOSFETs since they can't exactly as easily expand into other strategically important goods/industries like NAND, DRAM, telecommunications, chip design, biotechnology, and aerospace technology ...

Chip manufacturing is literally all Taiwan has and if it's competitors like China (SMIC), South Korea (Samsung), or the US (Intel) obtain parity in this field then it's pretty much lights out for Taiwan since trading with them offers nearly no advantages. The day TSMC reaches the limits of silicon MOSFETs is likely going to be the day they stop growing and will mark the day of the start of their decline. If TSMC goes down then so too does Taiwan since they probably can't diversify their business or economy ...
 

NTMBK

Diamond Member
Nov 14, 2011
9,141
2,367
136

DrMrLordX

Lifer
Apr 27, 2000
16,878
5,841
136
A big new US fab for Samsung is a smart move. If Intel are serious about integrating 3rd party "tiles" in their SoCs, then shorter supply chains between Oregon and those 3rd parties will help.
Right, but currently Intel is contracting with TSMC. nVidia is using Samsung.

edit: also, @ThatBuzzkiller has a good point that is sort-of on-topic, wrt Samsung et. al leaving companies like TSMC behind in the future once they've reach the limits of silicon. Expanding into other markets could be a winner for Samsung, but it'll be a matter of life-and-death for TSMC.
 
Last edited:

Kenmitch

Diamond Member
Oct 10, 1999
8,381
2,081
136
Well that's certainly not going to help any of the problems with shortages.
Are these the fabs that can't even produce enough gpus?

My parents, brother and sister most likely won't freeze to death now! Nasty weather in Texas currently!

I had no idea a whole state could be off the grid. I guess they're going to have to rethink about it, maybe come up with some kind of a backup plan.

Wouldn't the fabs have backup generators? Would be silly to not plan ahead, but then again Texas did it so who knows.
 

MrTeal

Diamond Member
Dec 7, 2003
3,057
705
136
Are these the fabs that can't even produce enough gpus?

My parents, brother and sister most likely won't freeze to death now! Nasty weather in Texas currently!

I had no idea a whole state could be off the grid. I guess they're going to have to rethink about it, maybe come up with some kind of a backup plan.

Wouldn't the fabs have backup generators? Would be silly to not plan ahead, but then again Texas did it so who knows.
I'm not sure they would be able to run even with generators. Those plants must use a massive amount of power, and if you want to run for the days the grid has been down you're going to need a tonne of diesel. Given the state of the roads and priorities for hospitals and other areas, I don't imagine bulk fuel is too easy to come by.
 

gdansk

Senior member
Feb 8, 2011
590
252
136
Are these the fabs that can't even produce enough gpus?

My parents, brother and sister most likely won't freeze to death now! Nasty weather in Texas currently!

I had no idea a whole state could be off the grid. I guess they're going to have to rethink about it, maybe come up with some kind of a backup plan.

Wouldn't the fabs have backup generators? Would be silly to not plan ahead, but then again Texas did it so who knows.
I think Austin S2 is still 14/11nm class products. Not sure what that'd be but not Ampere GPUs.
 

Mopetar

Diamond Member
Jan 31, 2011
5,400
1,982
136
Are these the fabs that can't even produce enough gpus?
It's not just GPUs, but all manner of chips. The US automakers are all scrambling to get chips right now because they canceled a lot of their orders (which someone else was happy to snap up) and are stuck without any.

The larger issue is that this could represent a disruption in the supply chain. A lot of wafers are going to need to be tossed because you can't just disrupt the fabrication process and restart it days or weeks later as though nothing happened. Depending on the process a wafer can take weeks to go from the starting phase to finished product and if some critical phase near the end was disrupted, that's weeks of wasted effort.
 

NTMBK

Diamond Member
Nov 14, 2011
9,141
2,367
136
I'm glad to hear they are shutting down non essential factories. The grid is collapsing, the energy should be diverted to keeping homes lit and heated.
 
  • Like
Reactions: moinmoin

ASK THE COMMUNITY