Intel in talks to buy GlobalFoundries for about $30 billion

sidthesexist · 16 Jul 2021 at 10:58

Not seen this posted yet?

July 15 (Reuters) - Intel Corp (INTC.O) is in talks to buy semiconductor manufacturer GlobalFoundries Inc for about $30 billion, the Wall Street Journal reported on Thursday, citing people familiar with the matter.

https://www.reuters.com/business/intel-talks-buy-globalfoundries-about-30-billion-wsj-2021-07-15/

Doesn't Intel already have a better node than what Global Foundries were using? Or would they be going for the 7nm that they were developing in the past?

KompuKare · 16 Jul 2021 at 11:19

Have to say I am more surprised by the price, if accurate. (Of course, GF 14nm was actually licensed from Samsung.)

With the rumoured $billions the US government is considering pouring into semiconductor fabricarion, this might be Intel trying to get all that money by saying they are the only American fab.

Grim5 · 16 Jul 2021 at 11:53

GloFo whole behind Intel and TSMC still makes a lot of small chips for non performing critical nodes such as audio chips and also their process is better than what the Russians and Chinese have

Zarax · 16 Jul 2021 at 12:09

GloFo would be a perfect starting point for Intel's renewed foundry business, easy capacity with decent performance for many processes (their 12nm is not that far behind Intel 14nm) and a snub at AMD, no matter how insignificant it is.

smilingcrow · 16 Jul 2021 at 13:58

It would also give them the buildings which could be re-tooled for newer processes.
That would save them time to ramp up rather than building from scratch.

feck199 · 17 Jul 2021 at 09:48

Need to expand!

dave28 · 17 Jul 2021 at 10:00

Intel making a big move, that is some serious money to spend. It completely dwarfs what other companies can spend

stooeh · 17 Jul 2021 at 10:41

Dont they make the IO dies for Ryzen chips?

Cyber-Mav · 17 Jul 2021 at 15:55

Intel buying amd's fabs thats interesting. Guess amd are going all in with TSMC and no turning back.
Seems way overpriced to me though.

LePhuronn · 18 Jul 2021 at 00:03

stooeh said:
Dont they make the IO dies for Ryzen chips?

Not with Zen 4, the IO die is apparently moving to 6nm. But, AMD have committed to a multi-year deal for 12nm and 14nm wafers. No idea what they're going to produce on it though. Chipsets maybe?

Haz123 · 18 Jul 2021 at 00:47

Intel has severely under invested in RnD for the past decade and it has bit them in the backside How buying one of the other Western based fabs will help them I do not know but I would prefer to see them using that huge sum of money creating new fabs with better more efficient processes. Intel being stuck on 14nm for so long has hurt us all.

Rroff · 18 Jul 2021 at 01:26

May not be Intel who is actually buying it - there is a lot of effort in the US at the moment to secure semi-conductor fabrication in the US which doesn't rely on Taiwan.

clv101 · 18 Jul 2021 at 09:37

smilingcrow said:
It would also give them the buildings which could be re-tooled for newer processes.

Buildings are worth millions, not billions. The real estate won't be a significant consideration in the deal.

Donnie Fisher · 18 Jul 2021 at 09:42

I agree, I dont think it would be for buildings alone. I also think a large part of it is to secure capacity thats outwith the reliance purely of the east.

LePhuronn · 18 Jul 2021 at 11:48

Haz123 said:
...but I would prefer to see them using that huge sum of money creating new fabs with better more efficient processes.

What do you think they've been trying to do for the past 5+ years? Do you think Intel have just been sat there milking 14nm through choice? Intel haven't made new fabs with more efficient processes because they don't have any efficient processes that work. 10nm SuperFin coming with Alder Lake is arguably the first new process that has any meaningful volume to it.

Rroff · 18 Jul 2021 at 11:56

clv101 said:
Buildings are worth millions, not billions. The real estate won't be a significant consideration in the deal.

A fully setup fab goes for billions - one of Intel's recent builds IIRC from breaking ground to production was like 8bn USD. The 2 new ones in Arizona is like a 20bn investment.

jigger · 18 Jul 2021 at 12:12

smilingcrow said:
It would also give them the buildings which could be re-tooled for newer processes.
That would save them time to ramp up rather than building from scratch.

Intel has its own buildings. Hectares of them sat idol. Property is literally the last thing Intel need.

jigger · 18 Jul 2021 at 12:22

Wasn’t Intel lobbying against the likes of TSMC and GloFlo? Offering to invest huge amounts of money in its own European semiconductor businesses.

Haz123 · 18 Jul 2021 at 13:00

LePhuronn said:
What do you think they've been trying to do for the past 5+ years? Do you think Intel have just been sat there milking 14nm through choice? Intel haven't made new fabs with more efficient processes because they don't have any efficient processes that work. 10nm SuperFin coming with Alder Lake is arguably the first new process that has any meaningful volume to it.

Yes. They made a choice at board level to put less money into RnD and pay higher dividends to increase their stock value. They made that choice around 10 years ago and stagnated ever since. They lack efficent processes that work because they did not put enough money into RnD and made bad choices on what to research. In the past intel would hedge their bets and invest into all potential processes and it paid off several times in the past.

In a decade they went from the worlds leading chip fab to meh, also rans , average. TSMC did not get the lead they have by luck , they put huge sums of money into RnD and infrastructure whilst Intel did not. Apple use the best fabs, in the past that was Intel but not anymore and the people to blame for that are the previous CEO and Board.

jigger · 18 Jul 2021 at 13:15

If Intel was any other company the only fix would be a top down reorganisation. But it’s Intel and they win by default.