I agree Intel should have been chasing smaller and boutique customers a lot sooner. There are a few small good news bits... I mean they landed Cobalt2 for 18A from Microsoft earlier this year. They claim to have been retooling for better arm fabrication... I don't know we'll see. Intels fab business is a long way away from getting orders from anyone doing anything consumer facing I would think. I just don't think anyone is gong to put a lot of faith in them to pull off a big mission critical geneation of much of anything. Tooling up for third party arm is probably a good way forward. If they can please Microsoft... they could maybe rely on them to run many generations of Cobalt and perhaps even their AI accelerators. (I don't know if they have a second gen AI chip in the works... but if they do I would assume they are going to use TMSC again until Intel can prove themselves with C2) Other players like Amazon might be in the mix as well if they can show some results in that regard. Those types of products are a little less sensitive to defect rates, as they are internal use and not for sale chips... MS is still saving lots of money as they direct contract and only end user of the part. (I just mean they can afford to throw away a few dies.... they don't have tight margin targets like Broadcom would) There is a very big buisness in doing all the custom US data center CPU / AI stuff over the next handful of years... and ya most of it will be ARM, and perfect customers for Intel. (If they don't F the first big ones up... if MS walks away next year and contracts TMSC Intel is screwed)
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
-
A Great friend to the HardForum with a great kid that he is trying to get a scholorship to continue his schooling. Please give hime a vote! Only 24 hours left! Thanks.
If you have an VOTE FOR KEENAN!
Intel to split off foundary business
- Thread starter HockeyJon
- Start date
Intel not using their own fabs for Arrow Lake is a pretty big tell tell sign of their fab capability. Intel is basically saying their fabs cannot make it as good as TSMC, plus Intel can make more profits/money outsourcing then using their own manufacturing facilities. Ouch! Big fail.
Then expect others would then use their fabs after that? I can’t believe anything Intel puts out. Intel over and over again said Arrow Lake would use 20a, when was that decided that 20a was crap but then not tell their investors anything until much later that TSMC will be used? Yep, just before launch. Does Intel have any processors scheduled to use 18a? What are these fabs going to make? Shutdown I presumed.
It is like cutting off a limb no longer functional and just killing the rest of the body. The jest is it sounds more like Intel has really given up the fab business and trying to separate it now.
Then expect others would then use their fabs after that? I can’t believe anything Intel puts out. Intel over and over again said Arrow Lake would use 20a, when was that decided that 20a was crap but then not tell their investors anything until much later that TSMC will be used? Yep, just before launch. Does Intel have any processors scheduled to use 18a? What are these fabs going to make? Shutdown I presumed.
It is like cutting off a limb no longer functional and just killing the rest of the body. The jest is it sounds more like Intel has really given up the fab business and trying to separate it now.
I don't know if its as much given up... as the non fab side can't afford to do another 14+++++ type gen, they need something solid to stem the bleeding.
Having said that there stock has been a mess cause their plans seem top be pretty Fd up. Panther lake is schedule for next year on 18a. It sounds like its going to be mobile only.... and then the next big Desktop arch Nova lake is apparently TMSC 2nm for '26 if that goes well. (or if they aren't a division of Qualcomm by then lol)
If Intel makes it to next year the big test for 18a will be Clearwater Forest their chiplet based server chips.... and Microsofts Cobalt 2. If they ship both of those generally on time that would be a big win. F those 2 up and if it hasn't already happened Intel gets eaten by someone. I mean best case and I mean best case they look like geniuses if clearwater does what they hope and smash AMD who keep eating at their server market. Then it looks like ok they contracted their consumer stuff to the fab everyone else uses for consumer stuff... and reserved their own 18a to fab the super high margin chiplet server chips and boutique datacetner stuff like Microsofts Cobalt CPUs. Investors would be very happy if that plays out... it means Intel offloads all its I9 type production to the fab competition, and starts turning out the $$$ parts at home.
A lot of what ifs, the big telling maybe how well Arrow Lake design competes with AMD somewhat lack luster Zen 5, at least in the consumer space. Design team vs design team. Who will win?
Been hearing next year from Intel way too much, which all their plans end up teetering out, about the only thing one can bank on.
Personally I doubt they have to worry about next year. Intel is probably looking at a hostile acquisition before the end of the year.
Nvidia wanted to own arm? I mean why not pick Intel apart at this point. Considering their valuation at the moment, I'm sure more then one company is considering what Intel would actually cost them. Intel actually made it easier the last year and then today by making it as easy as possible to cast aside the part no one really wants.
Microsoft contract + Amazon + Military all on 18a and who knows maybe Intel too will produce future chips using Intel -> sounds like something is at least cooking. If good or not, who knows?
Companies would be evaluating whether they would like the attention from authorities and if the low chances of such a takeover getting approved would be worth that attention.
Nvidia couldn't get arm. I don't think that, with current regulatory and national security sentiments, anyone will able to grab intel.
The issue with ARM was how Nvidia would have ended up being the company in control of a architecture used by everyone else. Intel doesn't have that problem. Buying Intel gives Nvidia x86... but Intel doesn't license x86 anymore. Regulators and ARM license holders were concerned Nvidia would be a bad landlord. No one else is effected with x86 in anyway. AMD owns their own x86 patents and their cross patents for life are written to survive Intel. Its not like Nvidia could buy Intel and charge AMD more... AMD doesn't pay for x86 they own it. Any extensions NV could do with x86 in the future would still be cross patented to AMD under their x86_64 deal. The only real issue could be the US wanting to protect Intels fab business. However there is no issue there the US gov was recently willing to withhold Chips funding cause they don't trust Intel management.... they might prefer Nvidia or Qualcomm being the Fab company, realistically anyone that buys Intel probably only wants the design side. So a split of the two halves actually makes it even more likely Intel gets absorbed by someone else imo. Intel would continue just not as a design company. I think there is also a very real possibility that more then one company becomes invovled and to skirt regulators Intels GPU business goes to a company like Microsoft or even Qualcom.
Twisted Kidney
Supreme [H]ardness
- Joined
- Mar 18, 2013
- Messages
- 4,741
It just keeps looking more and more like a death spiral.
- Joined
- Jan 14, 2012
- Messages
- 2,373
The tales of Intel's density advantage in their silicon should largely be debunked already. Because the claims of intel 10nm were that it had 2x the density of TSMC 7nm, yet for all their amazing density they had to cut cores out of 14nm and 10nm designs... Not to mention yields at said density weren't even viable. So, I think we all know just how well that all turned out for Intel. Clearly 10nm is superior running at temperatures akin to the surface of the sun, pulling wattage that makes old AMD FX series (32nm) chips look efficient by comparison.
This is exactly why I thought Pat's idea of trying to make Intel a competitor for cutting edge nodes with TSMC a farce. What company would run the risk of losing control of your trade secrets to have your chips manufactured for 10 or 20% cheaper, or even 5 or 10 percent faster performance? Not gonna happen by a sane company. The fact TSMC doesn't compete with its customers is a big deal.
I cannot find Intel ever claiming something like that are you sure ?
What I can find are very similar number while a bit higher for Intel, 106.10 vs 96.49 (some TSMC 7 were higher at 115)
https://semiwiki.com/semiconductor-...02-semicon-west-intel-10nm-and-gf-7nm-update/
Could it not be the terrible yield and heat that lead to cut core and not too low density ?
- Joined
- Jan 14, 2012
- Messages
- 2,373
Intel's slides all showed how dense their chips were. Other than that, it was always the intel fanboys on here that extolled the density virtues of Intel and how much better it was. But we never saw it, in practice.
- For Context, I recall many the conversation on the [H] where i as told I didn't understand the density advantage that Intel had over everyone else and how damn good their 10nm was going to be vs the competition. It never manifested, intel kept falling behind. The initial 10nm batch only found it's way into laptop CPUs in a single Lenovo laptop that couldn't match clocks and was not efficient enough to warrant mass production. 11th Gen Intel was garbage on 10 nm. 12th gen is pretty good, we all know what happened in 13th and 14th Gen chips. So, technically, Intel has had 5 generations to refine the tech and get better yields and they still have a 30 % failure rate...
Intel 10nm never yielded well. I think this was covered because Intel can absorb the losses as they were the Fab.
Last edited:
- Joined
- Jan 14, 2012
- Messages
- 2,373
Every generation of Intel from 22nm up has been it with massive delays. I think 10 nm was the worst. It was what? 6 years late?
- Joined
- Aug 31, 2009
- Messages
- 4,576
I can't see this being a positive thing. Intel not being a fab leader is bad juju. But I think ever since the 14nm fiasco the writing has been on the wall.
So maybe the density number are true ? (and intel never claimed double the TSMC 7 density ?)
For example:
Intel reports a density of 100.76MTr/mm2 (mega-transistor per squared millimetre) for its 10nm process, while TSMC's 7nm process is said to land a little behind at 91.2MTr/mm2 (via
That 10% more not 2x or I am missing some claim they made in some way ?
Has for never seeing it, considering how much the die shrinked by cores from 11th gen to 12th and how much those e-core can do by mm of die I am not so sure if that the case, giant density gain seem to have been true.
Gideon
Supreme [H]ardness
- Joined
- Apr 13, 2006
- Messages
- 4,258
10nm was a continuing disaster for Intel, they made it work but it never lived up to the hype about how it was superior to TSMC.
- Joined
- Jan 14, 2012
- Messages
- 2,373
Lets just say I'm wrong and you're right.
What does that do? What is your point?
Intel has been on their back foot since 3000 series AMD nearly caught up with them and 5000 series met and exceeded their performance at lower wattage and lower clocks.
They have been churning out the same thing until 12th gen and that was a last ditch, theoretical concept for a next gen processor that was designed for something like 7nm or beyond. But they couldn't get their Fabs to that point.
By the time it launched, TSMC 5nm had been in the wild for a good time (September 2020 launch), was there ever an hype in 2021 that Intel 10 was superior to TSMC 5 ? I was maybe out of the loop, but that would have been wild, we were already in the Samsung has an hard time keeping up with TSMC and TSMC N4P was upcoming soon.
I am more trying to understand, you say Intel claimed double density of TSMC, source (is this something fanboys said or Intel said) ? You say we never saw the density in action doing anything, while Intel 10 small e-core seem to compete well with Intel 14 core:
https://www.techspot.com/review/2374-intel-alder-lake-architecture/
Isn't this not a sign of a massive density upgrade being possibly true ?
As for what does that do, that made 24 core chips with a significant iGPU possible, but regardless the end result is what matter, not need to talk about density and other factor (there will be many, the end product CPU tell us more what we need to know and we do not know enough about that stuff to try or need to understand it from the transistor level)
Last edited:
- Joined
- Jan 14, 2012
- Messages
- 2,373
We can argue, infinitely, about the fine details of the die density and such, but it's pointless. Intel made a lot of damn claims that were proven false. I could go dig up all the articles and perhaps I am even exaggerating the double density (that people on here shilled for Intel for YEARS) claims. But non of it matters... because at the end of the day, Intel failed to deliver and they have been barely hanging on since AMD started executing well.
The writing has been on the wall since the Ryzen 1700 took the consumer marketplace by storm. Intel LAUGHED at that processor and said AMD was "Gluing" CPUs together. They kept on laughing until they realized they had been lapped, completely, by a smaller company they tried to kill many times over but failed to destroy.
- Joined
- Jan 14, 2012
- Messages
- 2,373
There has been Hype on the [H] for years about how superior Intel was. It's funny, now that they have been proven false, that people conveniently forget how hard they shilled for Intel.
I'm not attacking you. I'm just recalling getting my ass handed to me about this so called density superiority Intel had. No one thought AMD was going to Lap Intel at their own game.
Gideon
Supreme [H]ardness
- Joined
- Apr 13, 2006
- Messages
- 4,258
Plenty on here argued that once Intel had it running like it was supposed to, they would overtake AMD again. So yeah there was hype despite TSMC 5nm, and TSMC was often knocked by those same people for making up nm ratings. These days they are pretty quiet about the supposed superiority of Intel Fabs.
Yes I would imagine during the years before of future announcement, but I am talking by the time the 12900k in 2021, was it still going on ? Apple M1 Silicon was out by that point, TSMC incredible nodes was quite proven by then.
I feel people are bringing random fan wars to this subject a bit.... not actual Intel or Anandtech type of claims.
Last edited:
- Joined
- Jan 14, 2012
- Messages
- 2,373
It's possible on the fan wars. However, I wasn't referring back and forth to Anandtech at the time those where going on. There was a time that we all could just sit out here in this ecosystem and debate endlessly without digging too deeply outisde HardOCP's walls. It's still a possibility here but the forums don't seem as lively as they once were. This place was my one stop shop for information and computer advice. I miss it.
Intel made enough false claims to bury themselves before we even get the fan wars involved.
I think there's a bit of fanboyism going on, but more than that I think this board is understandably too focused on desktop and gaming. Desktop is a distant 3rd to server and laptop, and desktop gaming is just a portion of that.
I searched through this thread and the only mentions of "EPYC" and "Xeon" are in a few people's sigs. Also not a lot of talk about laptops. We desktop users have been getting the short end of the stick lately. Another overclocked and overvolted Intel 7 offering from Intel until Arrow Lake launches, and over on the AMD side we get a CCD or two an EPYC didn't want and a big 'ol discussion about whether or not the latest iteration is "meh". Arrow Lake sounds like it's aimed more at laptops. Better power efficiency is going to be the biggest improvement aside from hopefully not overvolting and frying themselves. I guess they figure it's good enough if it's faster than what you can get in a laptop.
Intel launched the Xeon 6700e series on Intel 3 in June, up to 144 e-cores. 6900p, up to 128 p-cores, is supposed to be coming soon, and they say they'll fill in the 6900e (up to 288 e-cores) and 6700p early next year. Those are the chips to watch if you want to see how Intel is really doing on the fab side until they get 18A going.
Lunar Lake and Arrow Lake are also interesting, but only on the chip design side since TSMC is making both. Still, they'll be indicative of what Intel can do if they either get their fabs up to speed or go fabless and use TSMC. Personally I'm pulling for Intel to get those fabs working and profitable. Why? Made in USA. Wouldn't mind if Intel's chip design business went poof. x86 might finally die if AMD had to stand alone. Problem is if they went bankrupt someone would buy it and keep making x86 chips.
I searched through this thread and the only mentions of "EPYC" and "Xeon" are in a few people's sigs. Also not a lot of talk about laptops. We desktop users have been getting the short end of the stick lately. Another overclocked and overvolted Intel 7 offering from Intel until Arrow Lake launches, and over on the AMD side we get a CCD or two an EPYC didn't want and a big 'ol discussion about whether or not the latest iteration is "meh". Arrow Lake sounds like it's aimed more at laptops. Better power efficiency is going to be the biggest improvement aside from hopefully not overvolting and frying themselves. I guess they figure it's good enough if it's faster than what you can get in a laptop.
Intel launched the Xeon 6700e series on Intel 3 in June, up to 144 e-cores. 6900p, up to 128 p-cores, is supposed to be coming soon, and they say they'll fill in the 6900e (up to 288 e-cores) and 6700p early next year. Those are the chips to watch if you want to see how Intel is really doing on the fab side until they get 18A going.
Lunar Lake and Arrow Lake are also interesting, but only on the chip design side since TSMC is making both. Still, they'll be indicative of what Intel can do if they either get their fabs up to speed or go fabless and use TSMC. Personally I'm pulling for Intel to get those fabs working and profitable. Why? Made in USA. Wouldn't mind if Intel's chip design business went poof. x86 might finally die if AMD had to stand alone. Problem is if they went bankrupt someone would buy it and keep making x86 chips.
DukenukemX
[H]F Junkie
- Joined
- Jan 30, 2005
- Messages
- 11,635
This is why Pat Gelsinger was "retired". Whatever Pat was cooking was taking far too long to finish, plus Arrow Lake was a disaster. Shareholders demanded growth and now they'll get it.
hititnquitit
2[H]4U
- Joined
- Apr 16, 2018
- Messages
- 3,129
Why would TSMC want anything to do with an Intel foundry?
The only way I can see this being remotely true is if they are given full control of the entire foundry and process. Sending a group of engineers... That doesn't pass the ultra control freak TSMC smell test.
The only way I can see this being remotely true is if they are given full control of the entire foundry and process. Sending a group of engineers... That doesn't pass the ultra control freak TSMC smell test.
Volume.
Intel isn't as good as TSMC's latest and greatest, but Intel is using all the same hardware as TSMC in virtually identical configurations.
TSMC's edge comes from its material sciences and its design software.
But even without those It's not like Intel's fab processes are bad, they just aren't as good.
TSMC could repackage them and sell them as is for many of their customers who are still out on the 22+ nm nodes.
The overwhelming bulk of silicon printed every year is still upwards of 90nm.
TSMC could do this and put so much of its competition out of business because it's biggest issue has been keeping up with volume, Intel operates more fabs than TSMC does, so yeah...
TSMC could do a lot with Intel's fabs because their issue is caused by a decade of bad management, TSMC would rope that in hard.
Over the past few months, Frank Yeary, the interim executive chairman of Intel, has spoken with administration officials and leaders of Taiwan Semiconductor Manufacturing Company about a deal that would separate Intel’s ailing manufacturing business from its semiconductor design and product business, according to four people with knowledge of the plan, who spoke on the condition of anonymity.
TSMC, which produces an estimated 90 percent of the world’s most advanced semiconductors, would assume control of Intel’s manufacturing business and take a majority stake in the business alongside a consortium of investors that could include private equity firms and other tech companies, the four people said.
The Trump administration has encouraged TSMC to do the deal. Howard Lutnick, President Trump’s nominee for commerce secretary, has been involved in the conversations and considers them one of the most consequential challenges of his new job, two of the people familiar with the discussions said.
Late last year, Intel’s board approached TSMC to gauge its interest in some sort of partnership, two of the people familiar with the talks said. In January, TSMC’s chief executive, C.C. Wei, met separately with Mr. Lutnick and Mr. Yeary to discuss how a tie-up might work.
Mr. Yeary has been speaking to Mr. Lutnick regularly about the idea since then, three of these people said. The Intel chairman’s interest in cleaving the company has also opened the door for suitors interested in acquiring Intel’s product business, including Qualcomm.
Some details of the discussions were previously reported by Digitimes, a Taiwanese news outlet, and Bloomberg.
Intel and TSMC declined to comment. Mr. Lutnick did not respond to a request for comment.
A Qualcomm spokeswoman declined to comment.
TSMC could address Mr. Trump’s demands by simply building more manufacturing capacity in the United States, said Stacy Rasgon, a semiconductor analyst at Bernstein Research. TSMC, which received up to $6.6 billion in grants from the CHIPS Act, is building three factories in Arizona and has the ability to expand there.
https://archive.is/cQM6a
TSMC, which produces an estimated 90 percent of the world’s most advanced semiconductors, would assume control of Intel’s manufacturing business and take a majority stake in the business alongside a consortium of investors that could include private equity firms and other tech companies, the four people said.
The Trump administration has encouraged TSMC to do the deal. Howard Lutnick, President Trump’s nominee for commerce secretary, has been involved in the conversations and considers them one of the most consequential challenges of his new job, two of the people familiar with the discussions said.
Late last year, Intel’s board approached TSMC to gauge its interest in some sort of partnership, two of the people familiar with the talks said. In January, TSMC’s chief executive, C.C. Wei, met separately with Mr. Lutnick and Mr. Yeary to discuss how a tie-up might work.
Mr. Yeary has been speaking to Mr. Lutnick regularly about the idea since then, three of these people said. The Intel chairman’s interest in cleaving the company has also opened the door for suitors interested in acquiring Intel’s product business, including Qualcomm.
Some details of the discussions were previously reported by Digitimes, a Taiwanese news outlet, and Bloomberg.
Intel and TSMC declined to comment. Mr. Lutnick did not respond to a request for comment.
A Qualcomm spokeswoman declined to comment.
TSMC could address Mr. Trump’s demands by simply building more manufacturing capacity in the United States, said Stacy Rasgon, a semiconductor analyst at Bernstein Research. TSMC, which received up to $6.6 billion in grants from the CHIPS Act, is building three factories in Arizona and has the ability to expand there.
https://archive.is/cQM6a
The White House official said the Trump administration supports foreign companies investing and building in the U.S. but is "unlikely" to support a foreign firm operating Intel's factories.
Earlier, Bloomberg had reported that Trump's team raised the idea of a deal between the two firms in recent meetings with officials from TSMC, and that they were receptive, citing a person familiar with the matter.
The White House official did not immediately comment on whether Trump's team met with the companies and raised the idea of such a deal. The official did not immediately respond to follow-up questions about whether TSMC could be involved with Intel's factories through a joint entity headquartered in the United States.
https://www.reuters.com/technology/...fter-trump-team-request-bloomberg-2025-02-14/
Earlier, Bloomberg had reported that Trump's team raised the idea of a deal between the two firms in recent meetings with officials from TSMC, and that they were receptive, citing a person familiar with the matter.
The White House official did not immediately comment on whether Trump's team met with the companies and raised the idea of such a deal. The official did not immediately respond to follow-up questions about whether TSMC could be involved with Intel's factories through a joint entity headquartered in the United States.
https://www.reuters.com/technology/...fter-trump-team-request-bloomberg-2025-02-14/
Outside of volume and control over the semiconductor industry, I'm not sure. They've already got an advanced fab in Arizona.
Advanced ish. It’s a 4nm fab which is basically just 5+++.
So it’s the most mature of the advanced nodes.
Which is fine. Things like cars and other integrated electronics don’t need the most advanced node. The only things that truly need the most advanced mode are those where power usage is a factor.
Most vehicle electronics are still on nodes measuring in the high double digits, it will be a slog of a fight to get them to move ahead with smaller nodes. Due to the nature of the vehicle industries supply chain and extreme cost cutting from just in time manufacturing they have been famously reluctant or unable to adopt newer components.
There are walls upon walls of text out there detailing and complaining about the issue and how it’s a problem.
In a new report from UDN, we're hearing that the supply chain and foreign media are reporting that the main goal of the Trump administration is to increase Intel's wafer manufacturing capacity through TSMC in order to strengthen "Made in America" from President Trump.
The US hopes to see TSMC holding a 20% stake in IFS (Intel Foundry Services) through a technology valuation or an actual cash investment.
details regarding the form of investment and related amounts have not been finalized.
Read more: https://www.tweaktown.com/news/1033...ith-qualcomm-and-broadcom-chipping/index.html
Sources indicate that Broadcom has been closely assessing Intel’s chip design and marketing businesses.
https://twitter.com/RockBolig/status/1891258388702740689
The US hopes to see TSMC holding a 20% stake in IFS (Intel Foundry Services) through a technology valuation or an actual cash investment.
details regarding the form of investment and related amounts have not been finalized.
Read more: https://www.tweaktown.com/news/1033...ith-qualcomm-and-broadcom-chipping/index.html
Sources indicate that Broadcom has been closely assessing Intel’s chip design and marketing businesses.
https://twitter.com/RockBolig/status/1891258388702740689
crazycrave
2[H]4U
- Joined
- Mar 31, 2016
- Messages
- 2,247
We know that Tom already spoke about Celestial in Dec 2024, and that part is not a rumor as he said the hardware was already done and the rumor now is that Intel wants to Fab it's own chips.
https://www.msn.com/en-us/news/tech...S&cvid=8b9e335bf9154caaf2e06918d29b62e8&ei=41
- We could be seeing a new Xe3P architecture for Intel's rumored next-gen 'Celestial' GPUs
- Rumors hint at Intel potentially using an in-house INTC process instead of the usual TSMC
- This could lead to stronger competition against AMD and Nvidia as Intel may be making a bigger effort on its GPUs
https://www.msn.com/en-us/news/tech...S&cvid=8b9e335bf9154caaf2e06918d29b62e8&ei=41
westrock2000
[H]F Junkie
- Joined
- Jun 3, 2005
- Messages
- 9,672
I’ll keep the names out of it, but my employer bought a 300mm fab from Intel and we also kept their employees. It’s been a couple years now and the fab still is not at “production” level expectations. This isn’t like a startup fab, it was an existing fab, with existing employees. Our leadership has expressed that the work culture at the fab just is not productive. I have never heard a ex-Intel employee or a tool vendor say they loved Intel or enjoy going there. Usually it’s a long the lines of “people are too afraid to speak up or rock the boat, or try something new”. I’ve interacted with teams over there and individually they seem fine and willing to try things. Maybe Intel leadership just beat the vigor out of the employees. So hopefully we can reinvigorate them. I used to think that my company was dysfunctional and that surely other companies had their act together. But then we got that fab and it’s like yikes, maybe other companies are just as bad, or worse. It’s a great industry, but you have to be open to new ideas & solutions from any level of the workforce.
TSMC Will Not Take Over Intel Operations, Observers Say
TSMC is considering taking a controlling stake in Intel Foundry as urged by U.S. President Donald Trump’s administration, according to a Bloomberg report last week, citing one person close to the matter.A few days later, the Wall Street Journal reported that Broadcom is interested in buying Intel Products if a potential investor takes over Intel Foundry.
Intel and TSMC declined to comment on the matter.
The reported plan to pair TSMC with Intel will fizzle, according to Handel Jones, CEO of International Business Strategies (IBS), which advises electronics companies and investors.
Taking on Intel Foundry would be a “battleship anchor” on TSMC’s bottom line, TechInsights vice chair Dan Hutcheson told EE Times.
There is no chance of an agreement, Hutcheson said.
Paul Triolo, who advises global tech clients at Washington, D.C.-based Albright Stonebridge Group, said: TSMC had to acquiesce to Washington before when it stopped selling chips to China’s Huawei, agreed to build fabs in Arizona and now considers helping Intel.
The Trump administration’s overarching aim is for more U.S. investment from TSMC, according to C.Y. Huang, president of FCC Partners, an investment bank based in Taiwan.
https://www.eetimes.com/tsmc-will-not-take-over-intel-operations-observers-say/