Taiwan's No. 2 chipmaker UMC eyes entering cutting-edge race

[XYang2023]

Company evaluates 6-nm partnerships to offset China pressure in older chips

TAIPEI — Taiwan’s second-largest contract chipmaker, United Microelectronics Corp., (UMC) is assessing the feasibility of venturing into cutting-edge chip production, a segment dominated by TSMC, Samsung and Intel, Nikkei Asia has learned.

UMC is exploring future growth drivers, four people said, including potentially 6-nanometer chip production, which is suitable for making advanced connectivity chips for Wi-Fi, radio frequency and Bluetooth, AI accelerators for various applications and core processors for TVs and cars.

Taiwan's No. 2 chipmaker UMC eyes entering cutting-edge race

Company evaluates 6-nm partnerships to offset China pressure in older chips

asia.nikkei.com

 

[XYang2023]

https://twitter.com/x/status/1939608597769646186

 

[XYang2023]

In that case, Intel foundry nodes would be:
Intel 16, Intel 12, Intel 6, Intel 4/3, Intel 18A, Intel 14A

 

[Barnsley]

Cmon UMC!!

You can do it!!

 

[siliconbruh999]

In that case, Intel foundry nodes would be:
Intel 16, Intel 12, Intel 6, Intel 4/3, Intel 18A, Intel 14A

Intel doesn't have viable 7nm class process node to sell

 

[XYang2023]

Intel doesn't have viable 7nm class process node to sell

I was saying Intel 6 through collaboration with UMC.

 

[XYang2023]

https://twitter.com/x/status/1939620364797243816

 

[osnium]

Maybe UMC can emerge as a critical 49% owner for Intel Foundry with US based PE firms and Intel products as a consolidated 51% owner. TD can be directed through UMC’s experienced engineers while the fabs use Intel’s US and EU based capacity.

 

[Fred Chen]

UMC collaborated with IBM before.

 

[siliconbruh999]

Maybe UMC can emerge as a critical 49% owner for Intel Foundry with US based PE firms and Intel products as a consolidated 51% owner. TD can be directed through UMC’s experienced engineers while the fabs use Intel’s US and EU based capacity.

The only thing that UMC can provide Intel is PDK Experience and Customers not Tech Intel is way ahead of UMC in that regard.

 

[osnium]

Maybe UMC can emerge as a critical 49% owner for Intel Foundry with US based PE firms and Intel products as a consolidated 51% owner. TD can be directed through UMC’s experienced engineers while the fabs use Intel’s US and EU based capacity.

The only thing that UMC can provide Intel is PDK Experience and Customers not Tech Intel is way ahead of UMC in that regard.

Doing technology development with an eye towards ensuring robust PDKs and bringing customers is very valuable.

 

[Daniel Nenni]

The only thing that UMC can provide Intel is PDK Experience and Customers not Tech Intel is way ahead of UMC in that regard.

Very true. Having outside eyes looking in is valuable as well. UMC has VERY satisfied and loyal customers. Intel Foundry could learn from that. I'm surprised we have not heard more about Tower Semi doing the same thing and how about Globalfoundries stuck at 14nm they got from Samsung? Intel needs to fill fabs, absolutely.

 

[XYang2023]

Very true. Having outside eyes looking in is valuable as well. UMC has VERY satisfied and loyal customers. Intel Foundry could learn from that. I'm surprised we have not heard more about Tower Semi doing the same thing and how about Globalfoundries stuck at 14nm they got from Samsung? Intel needs to fill fabs, absolutely.

Intel did mention it recently:

"Most people kind of settled in that that must mean ’27, and that’s generally kind of what we’re thinking is we can be breakeven. It doesn’t require a ton of revenue for Foundry. It’s somewhere in the, you know, single digit, called low to mid single digit billions of revenue that foundry’s got to get from external sources. But I would just remind you that some of that’s going to be our partnership with UMC, some of that’s going be our partnership with Tower, some of that’s gonna be packaging, and some of that’s going to be 18 a, some of that actually is gonna be older generations, know, like, example, Intel sixteen. So it’s not a ton that has to come from 18 a for it to to work."

https://www.investing.com/news/transcripts/intel-at-jpmorgan-conference-strategic-focus-under-new-leadership-93CH-4042544

 

[Daniel Nenni]

Intel did mention it recently:

"Most people kind of settled in that that must mean ’27, and that’s generally kind of what we’re thinking is we can be breakeven. It doesn’t require a ton of revenue for Foundry. It’s somewhere in the, you know, single digit, called low to mid single digit billions of revenue that foundry’s got to get from external sources. But I would just remind you that some of that’s going to be our partnership with UMC, some of that’s going be our partnership with Tower, some of that’s gonna be packaging, and some of that’s going to be 18 a, some of that actually is gonna be older generations, know, like, example, Intel sixteen. So it’s not a ton that has to come from 18 a for it to to work."

https://www.investing.com/news/transcripts/intel-at-jpmorgan-conference-strategic-focus-under-new-leadership-93CH-4042544

Right, but from what I remember Intel was going to manufacture additional 65nm BCD/RF-SOI capacity for Tower versus Tower getting access to Intel's FinFET processes.

 

[BruceA]

Leading edge node development takes billions of dollars.

IBM is a pretender only look at Samsung success and in a few years see Rapidus for what IBM does.

Intel has t been reliable since 10nm. There are no takers for Intel 4/3 or 18A. There is no confidence in the ecosystem to their PDKs nor their ability to manufacture with the yield the real leadership foundry does.

Doubt Intel could offer UMC much

 

[siliconbruh999]

Leading edge node development takes billions of dollars.

IBM is a pretender only look at Samsung success and in a few years see Rapidus for what IBM does.

Intel has t been reliable since 10nm. There are no takers for Intel 4/3 or 18A. There is no confidence in the ecosystem to their PDKs nor their ability to manufacture with the yield the real leadership foundry does.

Maybe it is true of their ecosystem but not the yield they are yielding pretty big dies on Intel 3 nearly ~600mm2 for their Server CPU also they are shipping 100s of thousands of these.

 

[cheesehead]

Hmm, somewhat confused on this part of the news item:

6-nanometer chip production, which is suitable for making advanced connectivity chips for Wi-Fi, radio frequency and Bluetooth

Is this just another case of main stream journalism, or do they really intend to build Wi-Fi and Bluetooth chips on 6nm? I have barely seen such technology on nodes below 28nm. The feature set of a 6nm chip in this area would require very specific features to justify the cost disadvantage

 

[Artificer60]

Maybe UMC can emerge as a critical 49% owner for Intel Foundry with US based PE firms and Intel products as a consolidated 51% owner. TD can be directed through UMC’s experienced engineers while the fabs use Intel’s US and EU based capacity.

Except for the fact that UMC doesn't do leading edge processes. There is a big difference between development at the leading edge and development on trailing nodes. I don't think that UMC's engineers could help Intel much in this regard. As siliconbruh999 said they do bring foundry experience to the table and that is the thing that Intel desperately needs right now. I see the potential for a significant win for both companies from this deal.

 

[dkr1986]

Right, but from what I remember Intel was going to manufacture additional 65nm BCD/RF-SOI capacity for Tower versus Tower getting access to Intel's FinFET processes.

Yes, Tower is going to use Intel's 300mm capable fab space in New Mexico, tool it with equipment worth up to $300 mil and fab 65nm chips.

Except for the fact that UMC doesn't do leading edge processes. There is a big difference between development at the leading edge and development on trailing nodes. I don't think that UMC's engineers could help Intel much in this regard. As siliconbruh999 said they do bring foundry experience to the table and that is the thing that Intel desperately needs right now. I see the potential for a significant win for both companies from this deal.

Intel gets Intel 6 as foundry offering with foundry grade PDKs, IPs and libraries (& to use Intel 7 based fully depreciated equipments and older fab capacity) and UMC gets another finfet tech node, US footprint and readily available fab space. Looks like Win-Win for both.

Is this just another case of main stream journalism, or do they really intend to build Wi-Fi and Bluetooth chips on 6nm? I have barely seen such technology on nodes below 28nm. The feature set of a 6nm chip in this area would require very specific features to justify the cost disadvantage

Intel has been using TSMC's N6 node for IO & Platform control tiles on Client CPUs. Maybe they find use on those lines for other customers. UMC\Intel 12 was announced in 1Q'24 and will be HVM in 1Q'27. So this UMC\Intel 6 will likely be roughly ~2028.

 

[Xebec]

The only thing that UMC can provide Intel is PDK Experience and Customers not Tech Intel is way ahead of UMC in that regard.

UMC can also provide capital funds for equipment, and perhaps even existing buildings/fabs that can be upgraded. (Intel is definitely minding their cash now)