Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/government-should-invest-in-tsm-and-not-intel.14792/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Government should invest in TSM and not Intel

Arthur Hanson

Well-known member
When it comes to semis, TSM is without question the clear leader over Intel and intends to not only continue leading but constantly increase it. Intel has had its chance and failed. Any investment should be an ownership stake or other clearly defined benefit, that's a true investment. I see no way Intel can offer anything substantial in finance, technology, or advanced production over TSM, not even close. TSM offers the benefit of an entire ecosystem of companies, many US-based, that would benefit the US directly to a significant edge over Intel. Just building a high-end semi workforce would be a substantial long-term benefit. Apple going with TSM only confirms this thesis.
 
I guess DoD and DoE agree it with you. That's why they encourage and lure (with subsidies) TSMC to setup a fab in Arizona in the first place.

DoD and DoE understand who is manufacturing those leading edge chips for them today, who can actually deliver the products for national security related applications on time and in time, and the future landscape among various semiconductor manufacturers. The challenge for them is how to deal with those politicians from the states that Intel and Globalfundries operate.
 
FIRST LOOK AT OAK RIDGE’S “FRONTIER” EXASCALER, CONTRASTED TO ARGONNE’S “AURORA”

“Aurora” is burning twice as much electricity to deliver slightly less performance than “Frontier”. And at $1 per watt per year to keep a supercomputer running, it could cost close to $60 million a year power “Aurora”, which adds up to close to $240 million over four years. At only 29 megawatts, you are talking only $116 million for “Frontier”. “Aurora” better be damned efficient computationally to pay that heavy power bill, which is only driven in part by having 50 percent more GPUs to deliver slightly less raw performance a year later than “Frontier”.


Intel/Cray failed to deliver the old "Aurora" and the new Aurora exascale supercomputer on time at Argonne National Laboratory. But AMD/Cray is on track to deliver "Frontier" exascale supercomputer, the first in US, at Oak Ridge National Laboratory by the end of 2021. By partnering with AMD instead of Intel, Cray can deliver a exascale supercomputer on schedule. It surely means something to many people.
 
Last edited:
DoD thinking can sometimes be the opposite here...

Investing in TSMC itself isn't a bad idea, but if something happened to TSMC (hack, war, Chinese stuff (tm), etc), a non-American company - there's substantial security risk there.

They really should invest in both .. Intel because it's the safe government choice, and TSMC because it's the clear leader and will give more advanced technology for at least the next few years.
 
"They really should invest in both .. Intel because it's the safe government choice,"

Intel is not a safe choice because not only they don't deliver what they promised but also their product and technology offering are limited. Intel got tons of taxpayers' money and certainly they will get more. But at the end of day DoD and DoE need real semiconductor products to make and operate fighter jets, missiles, satellites, and nuclear weapons. To DoD and DoE or any those national security related agencies, the major question is not how much they should invest in Intel. A more important issue is how do they diversify their sources (may or may not include Intel) in order to leverage the whole semiconductor supply chain and ecosystem, with advance capability, cost, and reliability in mind.
 
Keynote from Air Force Research Laboratory at CadenceLIVE Americas 2021

"Procurement Objectives

With the average life of different products in the 10- to 30-year range and a viable supply chain that is outsourced and offshored, the following are some key goals:
  • 1. Capping the lifecycle cost
  • 2. Improving cost and schedule predictability
  • 3. Access to a reliable and secure ecosystem
  • 4. Better visibility into suppliers"
IMO, DoD and DoE don't have too much confidence on Intel for #1 and #2 compare to TSMC. And their approach is to bring TSMC's leading edge fab to US to achieve the goal #3 and #4.
 
"They really should invest in both .. Intel because it's the safe government choice,"

Intel is not a safe choice because not only they don't deliver what they promised but also their product and technology offering are limited. Intel got tons of taxpayers' money and certainly they will get more. But at the end of day DoD and DoE need real semiconductor products to make and operate fighter jets, missiles, satellites, and nuclear weapons. To DoD and DoE or any those national security related agencies, the major question is not how much they should invest in Intel. A more important issue is how do they diversify their sources (may or may not include Intel) in order to leverage the whole semiconductor supply chain and ecosystem, with advance capability, cost, and reliability in mind.
Boeing, Lockheed Martin, Raytheon, and others don't exactly deliver what they say on time at cost consistently.. but there's a reason they get the contracts..
 
When it comes to semis, TSM is without question the clear leader over Intel and intends to not only continue leading but constantly increase it. Intel has had its chance and failed. Any investment should be an ownership stake or other clearly defined benefit, that's a true investment. I see no way Intel can offer anything substantial in finance, technology, or advanced production over TSM, not even close. TSM offers the benefit of an entire ecosystem of companies, many US-based, that would benefit the US directly to a significant edge over Intel. Just building a high-end semi workforce would be a substantial long-term benefit. Apple going with TSM only confirms this thesis.
You already posted this idea not so long ago. Why did you want to repeat it? It's not like the governmeent is going to listen to you. Also, what's the "high-end semi workforce" you are talking about? Fab workers? I doubt TSMC is going to bring their process development to US. And that's where know how is. Don't you want to have it in US?
 
Boeing, Lockheed Martin, Raytheon, and others don't exactly deliver what they say on time at cost consistently.. but there's a reason they get the contracts..
Agree. But in that kind of situations most those companies are the final product assemblers or prime contractors. There are limited or no alternatives DoD or DoE can go to.

In the world of semiconductors, DoD and DoE do have choices and they are using various suppliers for a long time while Intel is just one of many. With or without Intel, DoD and DoE can and will move forward. The DoE's Exascale supercomputer project shows Intel is in serious trouble about their capability and credibility.
 
Last edited:
You already posted this idea not so long ago. Why did you want to repeat it? It's not like the governmeent is going to listen to you. Also, what's the "high-end semi workforce" you are talking about? Fab workers? I doubt TSMC is going to bring their process development to US. And that's where know how is. Don't you want to have it in US?
TSMC's 1nm process node development appears to be a collaborative process with US universities (MIT and the University of California at Berkeley) and is supportive by various branches of the US government. I'm sure TSMC has its secret sauce (or "know how") for creating their cutting edge chips but the US now has some of that research/knowledge as well.

Of course, TSMC isn't going to share their technical know-how unless they receive something of equal or better value. TSMC has already stated that they NEED access to US technology in order to do business and that there is no alternative. The US has some leverage. That's why Huawei (TSMC's second biggest costumer in 2019) has had to sell a portion of their smartphone business and why they have so much trouble securing chips now.
 
The world is more interconnected than ever and this makes the collaboration of all types key. TSM building fabs in Arizona will be a win for all parties, even Intel if they handle it right.
 
Back
Top