top | item 30357370

Intel to Acquire Tower Semiconductor for $5.4B

150 points| frxx | 4 years ago |intc.com | reply

80 comments

order
[+] lnsru|4 years ago|reply
When reading headline I was thinking about Intel removing its competitor. Then I searched a bit and found this: “ Tower Semiconductor Ltd. is an Israeli company that manufactures integrated circuits using specialty process technologies, including SiGe, BiCMOS, SOI, mixed-signal and RFCMOS, CMOS image sensors, non-imaging sensors, power management, and non-volatile memory as well as MEMS capabilities.”

Looks like that Intel wants to go deeper in foundry business offering more services. I can speculate, that Intel has very good silicon process for logic circuits and will improve their offerings with technology from this acquisition. Sounds like a reasonable strategy to me.

[+] throwaway4good|4 years ago|reply
So it is some kind of complement to Intel's business? By the price Intel is paying compared to Tower's revenue, it really sounds like Intel is paying a lot without being able to add anything?!
[+] bayindirh|4 years ago|reply
Tower is one of the most capable CMOS image sensor fabs. They're generally used by MFT manufacturers for their image sensors.

The other, and the biggest player in this image sensor business is Sony.

This is a completely different lithography and technology when compared to CPUs and GPUs (larger feature sizes, different requirements, etc.).

Intel's transforming itself to be a fab operator it seems.

[+] spupe|4 years ago|reply
Intel has a clear strategy of becoming a big player in the semiconductor industry in the next decade. I honestly think it will work. Many people have observed that it is a cash-intensive industry with tiny margins, but once these huge investments have been made and there are only two or three companies that can dominate the market for five to ten years, you can extract huge value before any competitor catches up.
[+] kken|4 years ago|reply
>f becoming a big player in the semiconductor industry in the next decade

Did you mean

"a big player in the semiconductor _foundry_ industry in the next decade" ?

Intel has been the biggest semi player for most years during the last three decades.

[+] the-dude|4 years ago|reply
This comment could have been written 50 years ago.
[+] Shadonototra|4 years ago|reply
They don't have a strategy, they react to their downfall and are rushing to find a solution

They are lucky the pandemic happened, it gave them the opportunity to take a step back and realize how screwed they are

[+] mmmBacon|4 years ago|reply
Where do you get that the profit margins are tiny? TSMC has a net profit margin of 38%; that’s higher than Google (27%).
[+] mhh__|4 years ago|reply
> becoming a big player

Are you saying they aren't a big player now?

[+] chrisjc|4 years ago|reply
> manufactures integrated circuits using specialty process technologies, including SiGe, BiCMOS, SOI, mixed-signal and RFCMOS, CMOS image sensors, non-imaging sensors, power management, and non-volatile memory

Aren't these the kinds of technologies used to produce the kinds of chips that are in great shortage atm?

Or perhaps a better way of stating what I'm trying to say is... aren't these the kinds of technologies that could be used to produce chips that can replace those that are in great shortage atm?

Not as dependent on highly advanced lithography machinery from the likes of ASML, but not so ancient as those that have recently stopped churning out the same basic chips after 20-30 years.

[+] mastax|4 years ago|reply
Yes. On the investor call one of the points was that Intel's cash would allow tower to expand and modernize its old 150mm and 200mm fabs, and to expand PMIC production enough to be competitive in that market.
[+] nickpeterson|4 years ago|reply
Intel will succeed as long as it isn’t grossly mismanaged simply because the US government needs them to succeed. I’m also surprised they don’t pressure Apple to start building fabs in the US as well considering they own the entire stack now.
[+] Hamuko|4 years ago|reply
I'd foresee Apple ordering chips from Intel Foundry Services' new US fabs before Apple actually building their own fabs. Although AFAIK, IFS fabs do not yet compete in the same nodes as TSMC.
[+] jcranmer|4 years ago|reply
Apple doesn't fab chips itself; those come from TSMC… who are building fabs in the US now.
[+] gigatexal|4 years ago|reply
Is this Intel’s PASemi moment or something entirely different?

Remember PASemi were the folks Apple bought that gave us the A series chips and made Apple the chip powerhouse it is today.

[+] sanxiyn|4 years ago|reply
I think it's entirely different. Tower is a specialty foundry.
[+] graderjs|4 years ago|reply
Scooped you by 21 hours: https://news.ycombinator.com/item?id=30344944

I think this comment will be downvoted, but just want to run the experiment to be sure. I'll delete if so .. No worries.

[+] MichaelApproved|4 years ago|reply
Since you’re conducting an experiment, I’ll provide my feedback. Usually I’d downvote a simple comment talking about being first to post but you also provided a link to the previous discussion which I thought was worth upvoting.

Regarding the talk of downvotes, I’d usually downvote for all the classic reasons but you turned it into an experiment which will spark discussion, so that’s an extra reason to upvote.

I take my upvote/downvote powers very seriously. Or maybe a bit too seriously?

Edit: after making this comment I clicked and saw your post got no comments, so I must now consider it to be a “simple” comment with little value.

Hmmmm. I guess I’ll just take back my upvote but refrain from downvoting.

[+] ant6n|4 years ago|reply
You'll be downvoted for talking about downvoting (just as I will be). Just avoid.