Question [WSJ] Intel in talks to buy GloFo

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Failnaught

Member
Aug 4, 2008
26
25
91
This last year to year and a half has been very strange when it comes to Intel news. I have no idea what to make of it anymore.

Could it be that Intel doesn't even know what they want to do? For a while they couldn't find a CEO with any kind of technical background (Swan was the first CEO from the business side, I think, and publicly said he doesn't want the job). Then some senior technical people left and/or were let go. Pretty much immediately afterwards Pat Gelsinger was hired as the CEO. Rumor has it he's an engineer's engineer, and previously turned down the Intel CEO job a few years ago when they appointed Bob Swan instead. And then we hear a rumor that they booked a lot of volume at the leading edge TSMC process (probably they signed the contact a few years ago). I thought they were getting out of leading edge fab, and/or getting into a strategic relationship with TSMC?

Now they want to buy GloFo. So no way, they are doubling down on their fab and definitely pushing the newest nodes and concentrating the best talent in the USA in the same place to push it through. Plus it helps to pick up some customers and the managers who know how to run that business.

Sounds to me like they were of two minds on this, and eventually the fab side won the argument, expelled the people who wanted to go fabless. Could it be possible Pat G waited for the dust to settle on these internal political problems, before he signed up to be the CEO? If so, that's a shrewd move, though it probably cost Intel a couple of years.
 

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136
I thought they were getting out of leading edge fab, and/or getting into a strategic relationship with TSMC?

Buying GloFo doesn't mean that they will ever go back to the leading edge. They still need the leading edge for their CPU products.

As for Pat you have to factor in that VMWare is screwed because of The Cloud. He may not have felt that way when Swan was given the job.
 

eek2121

Platinum Member
Aug 2, 2005
2,904
3,906
136
Yes, came into this thread to make a similar point: this looks a like a play to be the sole (significant) recipient of that state aid. I'm sure for show some smaller players will get some pocket change though.

More likely, an easy avenue that allows Intel to both rebrand and split it's foundry from the CPU business.

Yes, but if someone has "Eyes wide shut perspective or understanding", that person he cannot see that red reality.

I hope you aren't referring to me. Let's see who has more team red in their household, me or you? Who owns more AMD stock? Me, or you? (hint: "Eek2121" is the answer to both questions). Shoot, you can't even refrain from using a bold RED FONT constantly in many of your posts...

I'm pro AMD currently, but I'm not blind to the competition, unlike a few others here. I know Intel has the ability to execute because they have done so repeatedly, and are about to do so again. You should stop attacking others and start asking AMD where their >8 core laptops are, among other things.
 

scannall

Golden Member
Jan 1, 2012
1,944
1,638
136
A few things. I am somewhat skeptical about this leak. Not say it could never happen but there are more reasons it's a bad idea than a good one.

We have all seen how poorly acquisitions go with Intel. And there is no reason to expect that this one would somehow be the one that worked. Meshing any other corporate culture with Intel goes poorly. With their 'Not Invented Here' mindset it makes actual progress difficult. One of the upsides is buildings that are already fabs, which is a good thing. Those cost a lot. A potential good thing that would likely be wasted is personnel that actually know the contract fab business.

Andy Groves 'Paranoia' comments have shifted from competitors to different departments inside Intel. Can Pat change the direction of that ship? Maybe, I hope so anyway but it won't be easy.

They have spent the last 15 years or more making bad decisions and riding the wave of their predecessors. And now finally they want a bailout and relief for their MBa's? Really? Having squandered much of their upper level talent, and many dollars on pumping stock prices instead of their tech... No way.
 

A///

Diamond Member
Feb 24, 2017
4,352
3,154
136
If so, it's not known publicly to anyone. A recent analysis showed current and future Intel EUV purchases, can't recall the name of the firm, but it was pretty grim. 20K wpm o_O in 2023 I think.
I can't tell if this is a serious post or you being coy. You'll have to jog my memory of recent confirmation PR statements.
 

A///

Diamond Member
Feb 24, 2017
4,352
3,154
136
Could it be that Intel doesn't even know what they want to do? For a while they couldn't find a CEO with any kind of technical background (Swan was the first CEO from the business side, I think, and publicly said he doesn't want the job). Then some senior technical people left and/or were let go. Pretty much immediately afterwards Pat Gelsinger was hired as the CEO. Rumor has it he's an engineer's engineer, and previously turned down the Intel CEO job a few years ago when they appointed Bob Swan instead. And then we hear a rumor that they booked a lot of volume at the leading edge TSMC process (probably they signed the contact a few years ago). I thought they were getting out of leading edge fab, and/or getting into a strategic relationship with TSMC?
Don't quote me on this, but those prior signed orders were likely for non performant products they didn't want to waste their own machine time over. Non computer performant products aren't lithography sensitive. I'd assume it's more to do with their leading contracts with universities and labs requiring their products for super computers and other CPUs that would benefit from the node shrink, but would take up valuable wafer creation time at Intel's own fabs because the margins on those would be slim pickings. There's been rumors for a while that Intel's previously ousted employees, big and small, are being invited back because their technical background be it engineering or running fabs is critical going forward.

Sounds to me like they were of two minds on this, and eventually the fab side won the argument, expelled the people who wanted to go fabless. Could it be possible Pat G waited for the dust to settle on these internal political problems, before he signed up to be the CEO? If so, that's a shrewd move, though it probably cost Intel a couple of years.
Nah, it doesn't make sense. The people who wanted to go fabless were younger and got fired/made redundant the past few years. It's hard to blame Intel's past CEOs for their missteps when their own group managers messed things up long term, too. VR's actions while at Intel damaged the company for years to come. No 2-4 year turnaround will help the company as it will be a long term rebuilding process.

If Alderlake is successful then it's successful, but Intel must prove through multiple generations it has what it takes to succeed. They will be playing the AMD game just as AMD did. There's still a lot of questions that have gone unanswered about Alderlake, and I suspect when push comes to shove, the mainstream consumer will repeat the same line of thinking during the Pentium IV => C2D era when it comes to core counts as speed was the then concern and oft referred to argument. Though not at all surprised as the inverse has been through the last two generations with the 10th gen being better value and being better performing with an overclock.
 

Ajay

Lifer
Jan 8, 2001
15,332
7,792
136
Not that I know of. Samsung and TSMC have the bulk of outstanding orders with some smaller but known foundries ordering up equipment, too. To my knowledge, Intel still has some outstanding orders from ASML.

If so, it's not known publicly to anyone. A recent analysis showed current and future Intel EUV purchases, can't recall the name of the firm, but it was pretty grim. 20K wpm

E09U4jrVoAMa5X-

E09U5KvVoAAoYZC
Thanks. Mizuho Securities - I thought it was Muhuzo - lol.
 

Failnaught

Member
Aug 4, 2008
26
25
91
Buying GloFo doesn't mean that they will ever go back to the leading edge. They still need the leading edge for their CPU products.

Will they ever get back to the leading edge? That will take some skill and luck. Could they? Why not? AMD won back the performance crown after years of being nearly insolvent. Sure, they made smart decisions and executed like the grim reaper, but that's what Intel needs to do now. Will they try? Why buy out the other fab if they didn't want to play at the leading edge? And how do you grow as a business if you're stuck on an old process indefinitely? As the competition with AMD recently shows, having your own fab is a huge advantage. Even though Intel was a node or two behind, AMD still couldn't get more market share than Intel. It makes sense for Intel to keep trying, even if they won't have success inside of 5 years.

Another thing is that progress should be faster if you're a few years behind the leader. You can just reverse engineer/copy what the other guys are doing. So for Intel to be perpetually 1 node behind TSMC and maybe match Samsung, is not a bad outcome.
 

moinmoin

Diamond Member
Jun 1, 2017
4,933
7,619
136
Intel historically didn't keep older process nodes but converted its fabs to newer nodes to reach the quantity its known for. Due to that Intel is limited to newer nodes whereas TSMC has plenty of older nodes available still. So buying GloFo (if that comes to pass) might not be some for anything leading edge but actually for all the older and specialized nodes GloFo offers but Intel doesn't.
 

Arkaign

Lifer
Oct 27, 2006
20,736
1,377
126
Intel historically didn't keep older process nodes but converted its fabs to newer nodes to reach the quantity its known for. Due to that Intel is limited to newer nodes whereas TSMC has plenty of older nodes available still. So buying GloFo (if that comes to pass) might not be some for anything leading edge but actually for all the older and specialized nodes GloFo offers but Intel doesn't.

This is a very good way to offer a more diverse portfolio of products and fab options for things that don't need top end nodes. There are so many IOT, industrial, auto, aerospace etc ICs that need sourcing, and as we've seen from the shortages and increasing stress on TSMC's output, it's better for the industry if more investment in concurrent processes is made.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
it's better for the industry if more investment in concurrent processes is made.

Only if said investments lead to an increase in capacity. I'm not seeing any signs (yet) that Intel's money play will lead to GF or Intel building out more fab capacity state-side or to dealing with supply chain concerns.
 

JasonLD

Senior member
Aug 22, 2017
485
445
136
Only if said investments lead to an increase in capacity. I'm not seeing any signs (yet) that Intel's money play will lead to GF or Intel building out more fab capacity state-side or to dealing with supply chain concerns.

I don't see why they wouldn't increase the capacity when they are capacity constrained.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
I don't see why they wouldn't increase the capacity when they are capacity constrained.

All Intel is doing is trading $30 billion in assets to gain control of an existing foundry. That includes 0 announced foundry starts tied to the deal, assuming it goes through.
 

Vattila

Senior member
Oct 22, 2004
799
1,351
136
To make sense of this, we need to start to think of Intel as two separate entities: Intel Design and Intel Manufacture.

The IDM 2.0 strategy, recently announced by the new CEO Pat Gelsinger, sets out the future for Intel, of which Intel Foundry Services (IFS) is a big part. I think the path is clear. IFS will increasingly take over Intel Manufacture, while Intel Design migrates to a more independent and flexible industry-standard fabless operating model (Jim Keller worked on this transformation, in part, as I understand it, establishing the proper methodologies and abstraction levels between design and manufacture, as well as between different design groups — see discussion thread "Keller's role at Intel").

The IDM 2.0 strategy stipulates the use of external foundries as they see fit for future products, e.g. to ensure product performance targets are met and are competitive and can be produced in sufficient volume. Former CEO Bob Swan boldly announced this change of direction after the announcement last summer of the unfortunate 7nm process delay. He unequivocally said they would consider using external manufacture for main products in the 2023 time frame, if 7nm wasn't ready by then, while also adopting a disintegration design model (a.k.a. chiplet design), in which chiplets are made on the most suitable process, internally or externally. This change of direction requires an enormous operational change for Intel Design to a much more flexible design methodology, including changes to design tool use, standards and procedures.

Meanwhile, IFS is clearly set out to become an independent business, with its own accounting and management. I think it will gradually take over all of Intel's manufacturing operations, until formally spun off, either as a wholly owned subsidiary (like Samsung Foundry) or as a joint venture with other investors. Pat Gelsinger is a pretty smart guy. He probably understands that independence is important, and that the lack thereof stunted previous efforts to provide foundry services. He also probably understands that he needs to go big or go home. Scale is important if they aim to have any chance of establishing IFS as a credible competitor to the leading-edge foundries in the industry, i.e. TSMC and Samsung.

"Samsung too was a vertically integrated Integrated Device Manufacturing unit before it created a subsidiary by the name of Samsung Foundries. This subsidiary belongs to Samsung semiconductors. Samsung like Intel was a vertically integrated company."

Intel Vs Samsung: The Foundry Scenario | by Vijayalakshmi Swaminathan | The Research Nest | Medium

So the rumoured takeover of GlobalFoundries should be viewed as a merger between IFS and GlobalFoundries into an independent foundry business, operating wholly independent from Intel Design, with the complete dedication to customer needs. Anything less will not work, and probably would not pass industry scrutiny and regulation, I think.

PS. For more discussion (and a poll) about the separation of design and manufacture at Intel, see:

 
Last edited:

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
So the rumoured takeover of GlobalFoundries should be viewed as a merger between IFS and GlobalFoundries into an independent foundry business, operating wholly independent from Intel Design, with the complete dedication to customer needs. Anything less will not work, and probably would not pass industry scrutiny and regulation, I think.
Right now intel sells everything their current FABs can produce so if they wanted to implement IFS they would have to reduce the production of their own products, that what makes them the most amounts of money, to start producing stuff for others where they would only get a percentage.
Or they could buy up additional FABs, like glofo, so they don't need to let other makers into their own FABs.
Added bonus, they can use glofos old/different stuff for the methodologies and abstraction levels between design and manufacture you mentioned, without having to rely on supply from 3rd party FABs where they would have to fight against others (apple) for wafers.
 
  • Like
Reactions: Tlh97 and Vattila

Mopetar

Diamond Member
Jan 31, 2011
7,797
5,899
136
If Intel wanted to go that route I think they'd be better served by building more fabs specifically designed for third parties. Honestly the regulatory hurdles they would need to crawl over in order to acquire Global Foundries make the entire idea a fool's errand so they'd be far better served building more of their own. The U.S. government has been going on about needing more production in the country which puts Intel in a pretty good position as far as getting a lot of tax breaks for building new fabs.
 

A///

Diamond Member
Feb 24, 2017
4,352
3,154
136
If Intel wanted to go that route I think they'd be better served by building more fabs specifically designed for third parties. Honestly the regulatory hurdles they would need to crawl over in order to acquire Global Foundries make the entire idea a fool's errand so they'd be far better served building more of their own. The U.S. government has been going on about needing more production in the country which puts Intel in a pretty good position as far as getting a lot of tax breaks for building new fabs.
This still assumes the story is legitimate. Intel would be better off using their old fabs and offering foundry time for rock bottom prices with the US taxpayer picking up most of the tab of subsidies to offer X deals through X years. Buying GloFlo doesn't necessarily mean production in country would improve, not anytime soon. It's replacing one owner with another. You need to build out additional fabs to increase supply. None of this makes any sense if you consider Intel is trying to parlay the EU to build a fab in Dresden. This is doubly weird because there is a GloFlo fab in Dresden. I sometimes suspect this buying GloFlo rumor was stretched from that "rumor" about building out EU fabs to tune of $20, excluding this supposed and paraded around rumor of wanting to buy all of GloFlo for $30B. Intel with their broken record for the past six years, many lies to their investors and board, political ambitions and internal strife wants to spend at least $50B and in excess to fulfill their dreams?
 
  • Like
Reactions: Tlh97 and Saylick

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
If Intel wanted to go that route I think they'd be better served by building more fabs specifically designed for third parties.

Agreed. Especially since there's a push to get more semiconductor production "moved' to the United States. They easily could have worked out some kind of a collaboration deal with GF to share IP/expertise if Intel really needed it that didn't cost them $30 billion.
 

Vattila

Senior member
Oct 22, 2004
799
1,351
136
The Next Platform's take on the rumoured takeover bid:

"Sometimes you build a business from scratch and sometimes you buy one to build a foundation upon. With Intel wanting to be regarded as the second foundry option after Taiwan Semiconductor Manufacturing Corp, an acquisition of Globalfoundries – even for $30 billion, a 5X multiple over earnings – gives Intel a chance to really build this Intel Foundry Services business. According to data from IC Knowledge, which is cited by our friend Aaron Rakers at Wells Fargo in a recent report, TSMC has around 36 percent of wafer starts per month (WSPM) capacity across the foundries of the world, but Intel has around 18 percent and Globalfoundries has around 11 percent. The combined companies are within spitting distance with TSMC, especially if Intel starts counting its own chip manufacturing costs as revenue for Intel Foundry Services, which we think it will do. While people make a lot of noise about the advanced nodes because they are necessary for datacenter devices like CPUs, GPUs, DPUs, and other kinds of compute engines and network devices, a little less than half of the company’s revenues come from chips with 16 nanometer and larger transistor geometries. (To be specific, 5 nanometer is 18 percent of revenues and 7 nanometer is 31 percent.) That old stuff persists for a long, long time and helps pay the bills on the new processes. This is what Intel has been missing, and we said as much when the whole Intel Foundry Services idea was floated in earnest."

 

Vattila

Senior member
Oct 22, 2004
799
1,351
136
Semiconductor Engineering's take on the rumoured takeover bid focusses on the potential for government contracts:

"The company appears to be warming to the idea of becoming a trusted foundry for the U.S. Department of Defense. That could add billions of dollars in government contracts, and it could make Intel’s foundry efforts highly profitable, especially with the company’s chiplet approach. Acquiring GF would provide specialty processes, such as SiGe and FD-SOI, along with some general 300mm leading-edge capacity that already is up and running, fuelled by a variety of foundry customers. As of today, there are no leading-edge foundries owned by U.S. companies. Intel did not accept military contracts in the past because there wasn’t enough volume in custom chip designs. But the focus on chiplets and advanced packaging is a common cause for both Intel and the DoD, and a potentially lucrative one for Intel. And with the U.S. government’s newfound willingness to invest in domestic semiconductor technology in order to remain competitive with China, Intel could well become the beneficiary of steady government business and possibly a direct cash infusion."

 
  • Like
Reactions: lightmanek