Intel is reportedly in talks to buy the $30 billion foundry company AMD spun off a decade ago

[Lightwreather]

Summary

The Wall Street Journal is reporting that Intel is looking to snap up AMD’s former fabs as well, in a deal that could value them at $30 billion.

 

Quotes

Quote

Intel Corp. is exploring a deal to buy GlobalFoundries Inc., according to people familiar with the matter, in a move that would turbocharge the semiconductor giant’s plans to make more chips for other tech companies and rate as its largest acquisition ever.

A deal could value GlobalFoundries at around $30 billion, the people said. It isn’t guaranteed one will come together, and GlobalFoundries could proceed with a planned initial public offering. GlobalFoundries is owned by Mubadala Investment Co., an investment arm of the Abu Dhabi government, but based in the U.S.

 

Quote

It’s clear from the WSJ story that the deal isn’t a sure thing, and GlobalFoundries outright denied that it was in talks with Intel. But it’s possible Intel’s negotiating with the investment firm that owns GlobalFoundries instead, as the WSJ points out. It’s also intriguing that the Journal doesn’t have a “no comment” from Intel itself — that’s sometimes a canary to indicate a company did comment, just off the record or on deep background.

 

My thoughts

 This if it indeed goes through (and is approved) could be good for Intel. iirc, Global Foundaries has the tech to outfit a fab with a 7nm node but hasn't done so because of cost concerns. Intel is a large company with a pretty sizeable revenue amount, so this could help intel finally get on 7nm. But as with most everything, we'll have to see how this goes

 

Sources

 WSJ

The Verge

[Somerandomtechyboi]

31 minutes ago, J-from-Nucleon said:

this could help intel finally get on 7nm

"Intel inside, stuck on 14nm till the end of time"

[Mark Kaine]

Its kinda funny to me how Intel effed this up, shouldnt they be making tons of 7nm or even 5nm by themselves for a long time now…?  I know, they cannot, just wondering what went wrong along the way… think they're "stuck" on their own process since a *very* long time now?

 

Otherwise good news for intel, if the fab really can produce 7nm, in adequate numbers…

[Chris Pratt]

34 minutes ago, J-from-Nucleon said:

so this could help intel finally get on 7nm

Intel and AMD node sizes aren't the same. Intel kind of is now on "7nm", at least in mobile and soon on desktop with 12th gen. Intel 10nm is roughly equivalent to TSMC 7nm.

[porina]

45 minutes ago, J-from-Nucleon said:

Global Foundaries has the tech to outfit a fab with a 7nm node but hasn't done so because of cost concerns.

How far did they get with equipment purchase? They might have got some equipment for 7nm development before it was killed off, but I doubt it would be enough to roll out for any significant production volume. Also assuming GoFlo's 7nm could have been comparable to TSMC 7nm if they continued, this is not really worth anything to Intel as they have that level of tech already with 10SF and incoming 10SFE.

 

If we make the big assumption this goes ahead, it would not impact Intel's plans to 7nm at all. It would be no small job to refit GoFlo's assets to align with Intel's internal standards. I'd think they'd largely keep existing capacity "as is" in the short term and refit it to a future Intel node gradually. Maybe some Intel 7nm volume is possible but it would take significant further investment to push it to 5nm and beyond. The main benefit being it is probably cheaper than building fabs from scratch even considering the up front costs.

 

Having more capacity might also play into the foundry business they kicked off again. Not everything needs to be made on leading edge process.

[Levent]

I would not expect this to go through easily. I have been out of the loop on what GlobalFoundries produces when it comes to PC market but isn't AMD still a major customer? I also remember AMD having some sort of binding agreement with them when they sold the company. On that note, Intel has been buying companies like crazy this last decade, I just never expected them to buy a third party fab (IIRC they did have different wafer qualities across their suppliers at the beginning of 2000s and since then they were building their own fabs to keep quality consistent).

[porina]

10 minutes ago, Levent said:

I would not expect this to go through easily. I have been out of the loop on what GlobalFoundries produces when it comes to PC market but isn't AMD still a major customer? I also remember AMD having some sort of binding agreement with them when they sold the company.

AMD have been looking to get away from GoFlo for some time, and apparently have plans to do so.

 

Quote

In a brief Securities and Exchange Commission Form 8-K filing, AMD this afternoon has revealed that it has once again amended its wafer supply agreement with US fab (and AMD fab spin-off) GlobalFoundries. Under the terms of the amended seventh amendment, AMD will see out its existing commitment to use GlobalFoundries through 2024, with the latest amendment setting purchase targets for 2022, 2023, and 2024. Beyond those new targets, however, the agreement releases AMD from all further exclusivity commitments to GlobalFoundries. AMD is now free to use any fab on any process node that it wants.

https://www.anandtech.com/show/16677/amd-and-globalfoundries-update-wafer-share-agreement-through-2024

 

[rcmaehl]

Not as extreme but still . If you can't beat em, buy em.

 

 

[ToboRobot]

34 minutes ago, rcmaehl said:

Not as extreme but still . If you can't beat em, buy em.

 

 

I often think of a real life equivalent to DatAmerica from William Gibson's cyberpunk worlds.

Think of a fully vertically integrated tech company.  From ASML to TSCM to Intel and AMD to ASUS to NVIDIA to MICROSOFT to Apple and GOOGLE/Facebook/Twitter with the old guard of ATT, Oracle and IBM(and the others they swallow)... All functioning as a single monolithic mega corp.   

[pas008]

2 hours ago, Somerandomtechyboi said:

"Intel inside, stuck on 14nm till the end of time"

Yeah funny thing is Intel 14+++++++++++++

Think I forgot some pluses

it's is still competitive with AMD's TSMC 7nm mainstream 

If Intel went chiplet we prolly wouldn't have amd around especially availability playing keyrole for awhile

[StDragon]

16 minutes ago, pas008 said:

Yeah funny thing is Intel 14+++++++++++++

Think I forgot some pluses

it's is still competitive with AMD's TSMC 7nm mainstream 

If Intel went chiplet we prolly wouldn't have amd around especially availability playing keyrole for awhile

Competitive in performance, but not power consumption.

I'm rather amazed the latest i9 doesn't melt down into a puddle of slag.

[cj09beira]

2 hours ago, porina said:

How far did they get with equipment purchase? They might have got some equipment for 7nm development before it was killed off, but I doubt it would be enough to roll out for any significant production volume. Also assuming GoFlo's 7nm could have been comparable to TSMC 7nm if they continued, this is not really worth anything to Intel as they have that level of tech already with 10SF and incoming 10SFE.

 

If we make the big assumption this goes ahead, it would not impact Intel's plans to 7nm at all. It would be no small job to refit GoFlo's assets to align with Intel's internal standards. I'd think they'd largely keep existing capacity "as is" in the short term and refit it to a future Intel node gradually. Maybe some Intel 7nm volume is possible but it would take significant further investment to push it to 5nm and beyond. The main benefit being it is probably cheaper than building fabs from scratch even considering the up front costs.

 

Having more capacity might also play into the foundry business they kicked off again. Not everything needs to be made on leading edge process.

how far?, they had already bought several EUV machines for production so the was probably work done at the fabs to support it, intel doesn't want Glo Fo for their tech, they want it to expand capacity of their own nodes most likely.

How good was Glo Flo's 7nm?, it was being done with the help of ibm and samsung it was quite promising, them giving up probably also made samsung fall behind some what as they were all of a sudden left alone to finish it, i still think amd dropping glo fo's 7nm wasn't their best choice long term, as it affected not only glo fo but also samsung and ibm.

 

Now on topid: (viewer discretion is advised.)

Spoiler

Spoiler

Spoiler

 

 

[Vishera]

There are antitrust and competition concerns with this deal,

And the new executive order of the Biden administration can be a problem for the deal.

Quote

Executive Order on Promoting Competition in the American Economy

 

In the Order, the President:

• Announces an Administration policy of greater scrutiny of mergers, especially by dominant internet platforms, with particular attention to the acquisition of nascent competitors, serial mergers, the accumulation of data, competition by “free” products, and the effect on user privacy.

 

[Taf the Ghost]

3 hours ago, porina said:

How far did they get with equipment purchase? They might have got some equipment for 7nm development before it was killed off, but I doubt it would be enough to roll out for any significant production volume. Also assuming GoFlo's 7nm could have been comparable to TSMC 7nm if they continued, this is not really worth anything to Intel as they have that level of tech already with 10SF and incoming 10SFE.

 

If we make the big assumption this goes ahead, it would not impact Intel's plans to 7nm at all. It would be no small job to refit GoFlo's assets to align with Intel's internal standards. I'd think they'd largely keep existing capacity "as is" in the short term and refit it to a future Intel node gradually. Maybe some Intel 7nm volume is possible but it would take significant further investment to push it to 5nm and beyond. The main benefit being it is probably cheaper than building fabs from scratch even considering the up front costs.

 

Having more capacity might also play into the foundry business they kicked off again. Not everything needs to be made on leading edge process.

GloFo's would have been still DUV, like TSMC. EUV was 5nm for them, on their roadmap.

[cj09beira]

2 minutes ago, Taf the Ghost said:

GloFo's would have been still DUV, like TSMC. EUV was 5nm for them, on their roadmap.

They already had the euv machines though, at least 2 that i remember seeing in a article about how they were selling them after the cancelling of the 7nm node.

[Taf the Ghost]

2 minutes ago, cj09beira said:

They already had the euv machines though, at least 2 that i remember seeing in a article about how they were selling them after the cancelling of the 7nm node.

Right, because they still have to produce test nodes long before they roll out the full node into production. GloFo canceled and/or sold off their 5nm, 3nm and 1nm research as well. Nodes have a long, long development time.

[Taf the Ghost]

35 minutes ago, Vishera said:

There are antitrust and competition concerns with this deal,

And the new executive order of the Biden administration can be a problem for the deal.

 

American regulatory approval is unlikely to be that big of a hold up. German/EU or maybe even Saudi might be a different story. There'd also be some sort of side deal with AMD in the process.

 

On the deal itself, if Intel can acquire GloFo for all of 30B USD in stock (current market cap is over 225B), that's actually a good deal for them and opens up the foundry sales business instantly.  At the same time, GloFo is talking about an IPO next year, so it might be a simple as Intel was "kicking the tires" on the entire situation.

[Vishera]

4 minutes ago, Taf the Ghost said:

American regulatory approval is unlikely to be that big of a hold up.

 

Change of policy:

Quote

The Order also encourages the leading antitrust agencies to focus enforcement efforts on problems in key markets and coordinates other agencies’ ongoing response to corporate consolidation. The Order:

• Calls on the leading antitrust agencies, the Department of Justice (DOJ) and Federal Trade Commission (FTC), to enforce the antitrust laws vigorously and recognizes that the law allows them to challenge prior bad mergers that past Administrations did not previously challenge.
• Announces a policy that enforcement should focus in particular on labor markets, agricultural markets, healthcare markets (which includes prescription drugs, hospital consolidation, and insurance), and the tech sector.
• Establishes a White House Competition Council, led by the Director of the National Economic Council, to monitor progress on finalizing the initiatives in the Order and to coordinate the federal government’s response to the rising power of large corporations in the economy.

[Brooksie359]

2 hours ago, pas008 said:

Yeah funny thing is Intel 14+++++++++++++

Think I forgot some pluses

it's is still competitive with AMD's TSMC 7nm mainstream 

If Intel went chiplet we prolly wouldn't have amd around especially availability playing keyrole for awhile

14+++++++ is not competitive with TSMC 7nm. Is Intel cpus still holding in there? Yeah but that is a cpu not a fab process. The main advantage of better fab processes is power consumption and on that end tsmc wipes the floor with Intel 14nm++++

[pas008]

2 hours ago, StDragon said:

Competitive in performance, but not power consumption.

I'm rather amazed the latest i9 doesn't melt down into a puddle of slag.

 

21 minutes ago, Brooksie359 said:

14+++++++ is not competitive with TSMC 7nm. Is Intel cpus still holding in there? Yeah but that is a cpu not a fab process. The main advantage of better fab processes is power consumption and on that end tsmc wipes the floor with Intel 14nm++++

Wipes the floor? Looks pretty good for being plus the the millionth and they haven't even done chiplet lol

https://www.techpowerup.com/review/intel-core-i7-11700kf/21.html

 

 

[Brooksie359]

32 minutes ago, pas008 said:

 

Wipes the floor? Looks pretty good for being plus the the millionth and they haven't even done chiplet lol

https://www.techpowerup.com/review/intel-core-i7-11700kf/21.html

 

 

Yes wipes the floor. Did you even look at what you linked? At the part where they go over energy efficiency the 5800x wipes the floor with the 10700k. 

[pas008]

44 minutes ago, Brooksie359 said:

Yes wipes the floor. Did you even look at what you linked? At the part where they go over energy efficiency the 5800x wipes the floor with the 10700k. 

That's not Wiping the floor considering its 14nm+ to the billionth compared to 7nm

Seems exaggerated lol

[Brooksie359]

6 minutes ago, pas008 said:

That's not Wiping the floor considering its 14nm+ to the billionth compared to 7nm

Seems exaggerated lol

What do you mean considering? When comparing how competitive two fab techniques are we only consider how the two stack up. We don't give Intel a handicap simply because it's name is 14nm+++++ instead of say 7nm. 14nm+++++ is simply significant worse than 7nm and the benchmarks show that. The cold hard truth is that tsmc 7nm is much better than intels 14+++++. 

[pas008]

9 minutes ago, Brooksie359 said:

What do you mean considering? When comparing how competitive two fab techniques are we only consider how the two stack up. We don't give Intel a handicap simply because it's name is 14nm+++++ instead of say 7nm. 14nm+++++ is simply significant worse than 7nm and the benchmarks show that. The cold hard truth is that tsmc 7nm is much better than intels 14+++++. 

Lol read my first post Intel's 14nm is still competitive considering its 14nm ++++++

[derr12]

Didn't tsmc license 7nm tech from someone else before dropping it?