News Intel in Talks To Buy GlobalFoundries for $30 Billion: Report

ezst036

Reputable
Oct 5, 2018
168
53
4,660
0
Intel's rumored buy-out plan also comes as it lobbies the US government for subsidies
I really enjoy buying fabs for wealthy people. That made my day.

Intel made what, 40 billion profit in 2019? That makes it clear that I needed to buy them a fab as they just couldn't afford it. This was a top priority.
 
Reactions: JoBalz

spongiemaster

Estimable
Dec 12, 2019
1,598
792
3,560
0
I really enjoy buying fabs for wealthy people. That made my day.

Intel made what, 40 billion profit in 2019? That makes it clear that I needed to buy them a fab as they just couldn't afford it. This was a top priority.
The likely alternative if they don't get subsidized is to build overseas or not at all. Both of which are worse scenarios.
 
And earlier this year they spent $20B on 2 fabs in Arizona.

And in Ireland ...


New Mexico expansion ...

https://www.intel.com/content/www/us/en/newsroom/news/new-mexico-manufacturing.html#gs.6kai6a

And soon Bavaria ...

 
Reactions: gg83 and JoBalz

escksu

Respectable
Aug 8, 2019
520
192
2,060
0
Seems a bit strange since they already have some factories of their own. It's really ironic considering where GF came from.
Intel is currently facing severe capacity issues. Their biggest problem is not losing out to AMD, its not making enough chips because demand is way too high.

Buying GF is an easer and faster way than building an entirely new plant.
 

watzupken

Commendable
Mar 16, 2020
545
238
1,270
1
I feel Intel very highly geared to have all these big acquisition plans. Anyway, while most companies don't want the added responsibility of running a fab along with their chip business, Intel is the only one doubling down on it. At this point, I am not too sure if this is the right way ahead, but given that there were idle capacity reported for some fabs prior to COVID hitting us, I am not sure an aggressive expansion of fab by Intel makes sense. Also, Intel is not exactly the fab that most chip makers will turn to because Intel is also a competitor when it comes to chip design.
 

Howardohyea

Proper
May 13, 2021
229
56
190
13
I feel Intel very highly geared to have all these big acquisition plans. Anyway, while most companies don't want the added responsibility of running a fab along with their chip business, Intel is the only one doubling down on it. At this point, I am not too sure if this is the right way ahead, but given that there were idle capacity reported for some fabs prior to COVID hitting us, I am not sure an aggressive expansion of fab by Intel makes sense. Also, Intel is not exactly the fab that most chip makers will turn to because Intel is also a competitor when it comes to chip design.
I'm not sure, but I think a lot of these companies using IFS would be ARM devices and not X86/64? ARM isn't in desktop yet so I guess they won't have much overlap?

Oh also after all Intel already have the largest fabs in the world by capacity/revenue by far, acquiring GloFo would make Intel have the fab power to drown AMD with cheap CPUs
 
Reactions: Why_Me

spongiemaster

Estimable
Dec 12, 2019
1,598
792
3,560
0
The EU or the US should just buy Global Foundries.

It's cheaper than buying fabs for publicly-traded companies like Intel and TSMC.
The purpose of subsidizing new FAB construction is to keep the country at the technology forefront. Buying existing non-leading edge fabs doesn't accomplish that. Nor does it create new jobs, which in turn generates more tax revenue.
 
Reactions: everettfsargent

Kamen Rider Blade

Distinguished
Dec 2, 2013
300
138
18,860
0
If I was the US Government, I'd have Anti-Trust watching over this with a fine tooth comb.

I'd want Intel & it's Foundary Services to be Permanently severed into two companies if they were to allow a Global Foundaries Buy-out.

Force Intel to be a Purely Chip R&D/Design firm.

Force the newly formed (Intel Foundaries + Global Foundaries = Intelligent Global Foundaries) into a pure Foundary Service to directly compete against TSMC.
 
This doesn't make sense to me... I give Intel a lot of grief over whatever it does, but buying GF doesn't make sense to me given on how they're investing currently.

I'll think this could be the usual "let's inflate the potential value for the IPO and get some cash from dumb speculation". That makes more sense to me, really.

Regards.
 

Nick_C

Distinguished
Apr 20, 2007
100
13
18,695
3
Not sure that the purchase would be permitted by the relevant competition authorities around the world - as it may be seen as an attempt to monopolise Foundry provision.
 
Last edited:

Co BIY

Honorable
Jun 18, 2015
629
120
11,190
8
Maybe it's just a case of seeing the assets for sale relatively cheap rather than a grand strategic move in support of the company's long term strategy.

In a market with limited players they not only have to calculate what it's like if they buy GF but what it's like if one of their competitors does so (or does so cheap).

Also Intel can make a good argument for relaxed anti-trust scrutiny based on their relative weakness right now. Also the national champions concept may be trending.

Not sure that the purchase would be permitted by the relevant competition authorities around the world - as it may be seen as an attempt to monopolise Foundry provision.
Intel getting fully into the foundry business would make it (foundry market) more competitive rather than less as it is currently totally dominated by TSMC.
 
Last edited:

spongiemaster

Estimable
Dec 12, 2019
1,598
792
3,560
0
Not sure that the purchase would be permitted by the relevant competition authorities around the world - as it may be seen as an attempt to monopolise Foundry provision.
Intel is only 6th in 300mm wafer capacity accounting for 6% of the market. Global Foundries is 7th with 4%. If you add GloFo's capacity to Intel's, the combined total would still leave Intel in 6th place, 1% behind Kioxia. Intel wouldn't be monopolizing anything with only 10% market share. #1 Samsung has 21% of the market, more than twice the capacity of Intel and GloFo combined.
 
Last edited:
Reactions: Groveling_Wyrm

kjfatl

Prominent
Apr 15, 2020
7
9
515
0
Seems a bit strange since they already have some factories of their own. It's really ironic considering where GF came from.
Ironic, Yes, but it makes a a lot of sense. Global foundries will fade away if they don't get a huge infusion of cash. Everything they are doing today is likely to be available in China in a few years at a lower cost. They have a huge number of skilled engineers who know how to make high end semiconductors, but they don't have the capital. Intel screwed up royally in forcing many expensive process engineers to retire, and they blew it when it came to building parts for other companies. The key to making this work is for Intel to keep Global Foundries as a wholly owned, but separate subsidiary.
Semiconductor production is an insanely expensive business. If Intel can't double it's market share in semiconductor production, they will become another on of the BUNCH companies.
 
Reactions: vern72

Groveling_Wyrm

Distinguished
If I was the US Government, I'd have Anti-Trust watching over this with a fine tooth comb.

I'd want Intel & it's Foundary Services to be Permanently severed into two companies if they were to allow a Global Foundaries Buy-out.

Force Intel to be a Purely Chip R&D/Design firm.

Force the newly formed (Intel Foundaries + Global Foundaries = Intelligent Global Foundaries) into a pure Foundary Service to directly compete against TSMC.
This isn't one of the AT&T type of monopolies, splitting into the Baby Bells, where Intel will get split. Intel has no monopoly in any form, whether design, or foundry services. As Spongie said...

Intel is only 6th in 300mm wafer capacity accounting for 6% of the market. Global Foundries is 7th with 4%. If you add GloFo's capacity to Intel's, the combined total would still leave Intel in 6th place, 1% behind Kioxia. Intel wouldn't be monopolizing anything with only 10% market share. #1 Samsung has 21% of the market, more than twice the capacity of Intel and GloFlo combined.
There probably won't even be much of a glance at Intel buying GloFlo, in terms of government oversight. They may actually encourage it, as it ensures greater taxes.
 

Howardohyea

Proper
May 13, 2021
229
56
190
13
I thought about it for a few days and I'm wondering what Intel will gain except fab capacity?

Intel's 14nm/10nm is already competitive with TSMC 7nm in terms of transistor density, and with GloFo abandoning leading edge nodes I'm guessing Intel and GloFo is roughly on-par technology wise. Acquiring them would make Intel expand their fab power immediately without making their own fabs.
 

waltc3

Commendable
Aug 4, 2019
222
110
1,760
0
If true, the rumor would certainly surprise me....;) AMD unloaded Glo Flo because it no longer wanted the FABs--seems pretty clear to me. Intel would be doing AMD a favor in picking up GloFlo, seems to me, as GLO Flo hasn't had any more success in going to 7nm than Intel has had getting there. Believe me when I say that if Intel was shipping 14nm identical to TSMC's 7nm, Intel would never let that PR stone go unattended...! Instead, Intel has been promising to catch TSMCs 7nm for the last 2-3 years, and in a very public fashion, IIRC.

I thought about it for a few days and I'm wondering what Intel will gain except fab capacity?

Intel's 14nm/10nm is already competitive with TSMC 7nm in terms of transistor density, and with GloFo abandoning leading edge nodes I'm guessing Intel and GloFo is roughly on-par technology wise. Acquiring them would make Intel expand their fab power immediately without making their own fabs.
A sillier notion I have not heard...it is pure bunk, imo...;) Just looking at Intel's real 14nm TDPs is proof enough that Intel is a few years behind TSMC--I mean, aside from the fact that Intel admitted it just a few months ago to investors and staff--it was very public and Intel made it more than once in the last couple of years. If you look at the size of Intel's and AMD's packages, but you don't compare their real TDPs--not the ones Intel puts out for public consumption, of course--it's easy to see how far ahead of Intel TSMC is atm.

You know, after Intel has publicly admitted the same--more than once in the last few years as already mentioned--it boggles the mind where people get off thinking that AMD is cheating in its node sizes while Intel continues to lie about being behind TSMC. Can you picture Intel lying to the effect of deliberately making itself look worse than it has to?

The real yields cannot hide the TDP picture, nor would it do Intel any good at all to keep telling investors and customers how far behind TSMC Intel currently is. In light of the facts, Intel has no choice but to stick to the facts lest the company is seeking more ridicule.
 

spongiemaster

Estimable
Dec 12, 2019
1,598
792
3,560
0
If true, the rumor would certainly surprise me....;) AMD unloaded Glo Flo because it no longer wanted the FABs--seems pretty clear to me. Intel would be doing AMD a favor in picking up GloFlo, seems to me, as GLO Flo hasn't had any more success in going to 7nm than Intel has had getting there.
AMD spun off their manufacturing arm because they were so broke as a company they couldn't afford to stay in the business. It takes huge amounts of capital to build and upgrade fabs. Before they sold it off, AMD had over $5 billion in debt and they actually had to stop the upgrade of one of their facilities and cancel planned new ones because they couldn't secure the necessary capital to pay for them. As part of the deal to offload the fabs, AMD also offloaded over a billion dollars in debt to the new company. It didn't matter what AMD wanted, they only had one financial option which was to sell it off if they wanted the remaining company to stay afloat.
 
If true, the rumor would certainly surprise me....;) AMD unloaded Glo Flo because it no longer wanted the FABs--seems pretty clear to me. Intel would be doing AMD a favor in picking up GloFlo, seems to me, as GLO Flo hasn't had any more success in going to 7nm than Intel has had getting there.
Getting to 7nm is the most useless thing ever if you can't make enough of them to sell one to everybody that wants one.
Even if AMD could get 100% of TSMCs output of 7nm they would still be a small player compared to intel.

Maybe intel wants the fabspace to upgrade it to 7nm so they can provide all the chips to the whole world.
Maybe intel just wants the patents and the fabs themselves are just a bonus.
Maybe the whole thing is just fake news.
Believe me when I say that if Intel was shipping 14nm identical to TSMC's 7nm, Intel would never let that PR stone go unattended...! Instead, Intel has been promising to catch TSMCs 7nm for the last 2-3 years, and in a very public fashion, IIRC.
Every single announcement intel makes includes the "we are industry leaders" "we have the cutting edge" and so on.
 

PCWarrior

Distinguished
May 20, 2013
169
54
18,670
3
Global foundries has a history of licencing nodes from other foundries, particularly Samsung, and tweaking them. Intel could buy Global foundries and not merge it with their own fabs. Instead have it license leading-edge (or almost leading edge) nodes from Samsung and TSMC (like Samsung's 8nm node or TSMC's 7nm node) and use it for customers that would otherwise go to Samsung or TSMC. It will also act as a vehicle for Intel to license nodes from TSMC and Samsung in case their own process gets delayed again, like it happened with 10nm, without losing face as it won’t be Intel's fabs buying/licensing the nodes but Global foundries.
 
Reactions: Howardohyea

Howardohyea

Proper
May 13, 2021
229
56
190
13
Global foundries has a history of licencing nodes from other foundries, particularly Samsung, and tweaking them. Intel could buy Global foundries and not merge it with their own fabs. Instead have it license leading-edge (or almost leading edge) nodes from Samsung and TSMC (like Samsung's 8nm node or TSMC's 7nm node) and use it for customers that would otherwise go to Samsung or TSMC. It will also act as a vehicle for Intel to license nodes from TSMC and Samsung in case their own process gets delayed again, like it happened with 10nm, without losing face as it won’t be Intel's fabs buying/licensing the nodes but Global foundries.
hm I haven't paid attention to this area, but you sure do have a point. To some extend I'd also argue that IFS is already getting some of their customers, didn't Intel say a few weeks/month back that they're already in talks with around 40 companies about using IFS?
 

ASK THE COMMUNITY

TRENDING THREADS