Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/intel-wins-us-government-project-to-develop-leading-edge-foundry-ecosystem.14572/
        )

    [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
)

Intel Wins US Government Project to Develop Leading-Edge Foundry Ecosystem

soAsian

Active member

"The U.S. Department of Defense (DOD) has recently sought to diversify its approach to securing advanced microprocessors by leveraging commercially available technologies developed by U.S. companies. Other than Intel, the majority of U.S.-based chip designers are fabless, which means they design and sell integrated circuits that are fabricated by contract manufacturers called foundries. Today, more than 80 percent of leading-edge manufacturing capacity is concentrated in Asia1, leaving the DOD with limited onshore access to foundry technology capable of meeting the country’s long-term needs for secure microelectronics."

I'm guessing this is Pat's IDM 2.0 core selling point by piggy back on Uncle Sam and get that govrt's money to beat out Samsung and TSMC (maybe?). If TSMC and Samsung double down on building fabs in the US. that might thwart Intel's plan(?)
 
Here's my two cents...

This is an Intel press release that essentially says we have government funds to develop a foundry ecosystem for the DoD for Intel 18A. That's great for Intel and good for the US government to have that option.

That said the US government made a concerted effort to bring TSMC to Phoenix, AZ. I am 100% sure that was to secure the latest "leading-edge process technologies" specifically FOR the DoD...otherwise what would be the point of the US subsidizing TSMC? TSMC's chairman, Mark Liu, explicitly stated that the Phoenix fab would be a profitable one for TSMC. I have no doubt that the US government will be leveraging TSMC in the very near future.
 
Here's my two cents...

This is an Intel press release that essentially says we have government funds to develop a foundry ecosystem for the DoD for Intel 18A. That's great for Intel and good for the US government to have that option.

That said the US government made a concerted effort to bring TSMC to Phoenix, AZ. I am 100% sure that was to secure the latest "leading-edge process technologies" specifically FOR the DoD...otherwise what would be the point of the US subsidizing TSMC? TSMC's chairman, Mark Liu, explicitly stated that the Phoenix fab would be a profitable one for TSMC. I have no doubt that the US government will be leveraging TSMC in the very near future.
Isn't TSMC's AZ fab is 5nm? isn't that going to be an old tech by the time the fab is done?
 
Here's my two cents...

This is an Intel press release that essentially says we have government funds to develop a foundry ecosystem for the DoD for Intel 18A. That's great for Intel and good for the US government to have that option.

That said the US government made a concerted effort to bring TSMC to Phoenix, AZ. I am 100% sure that was to secure the latest "leading-edge process technologies" specifically FOR the DoD...otherwise what would be the point of the US subsidizing TSMC? TSMC's chairman, Mark Liu, explicitly stated that the Phoenix fab would be a profitable one for TSMC. I have no doubt that the US government will be leveraging TSMC in the very near future.
DoD, DARPA, Navy, Air Force, Space Force, NSA, and DoE are facing several challenges in acquiring advanced, high performance, and specialized semiconductor products:

1. They are looking for stable and predictable suppliers/supplies from both short term and long term point of views.

One of their top principle is to aquire real and usable products on time. Any suppliers' roadmaps or theories can only entertain them to certain degree. An amazing chip on paper today won't help a F35 or a Patriot missle to fly this year or even next three years.

In this regard, Intel does have some credibility issues right now.



2. They have to secure and acquire chips supplies well ahead of the mass production of any weapon systems. They need the real product and plenty of time for hardware and software development and testing. It means a leading edge 3nm chip coming to market next year may not be meaningful for most weapon systems rollout from production line next three to four years.

3. Most defense and national security related semiconductor demand are in low volume. For example, Lockheed Martin only produces up to 180 F35 a year. How many chips does Lockheed Martin need every year for 180 F35 production? Compare it to any commercial semiconductor products, it's really a tiny volume. It can partially explain why TSMC Arizona fab initial monthly production volume is only 20,000 12-inch wafers. For DoD or DoE, this number is more than enough already.

4. DoD and DoE recognized a while back that commercial semiconductor development are getting faster and faster and they have to take advantage on it. Their collaboration with foundries are wide and deep. Some public available information can be seen here:


 
Yeah that was the problem with Intel's previous foundry effort, they needed more money and guidance from that bastion of lean, innovative, free market efficiency: the MIC.
 
DoD, DARPA, Navy, Air Force, Space Force, NSA, and DoE are facing several challenges in acquiring advanced, high performance, and specialized semiconductor products:

1. They are looking for stable and predictable suppliers/supplies from both short term and long term point of views.

One of their top principle is to aquire real and usable products on time. Any suppliers' roadmaps or theories can only entertain them to certain degree. An amazing chip on paper today won't help a F35 or a Patriot missle to fly this year or even next three years.

In this regard, Intel does have some credibility issues right now.



2. They have to secure and acquire chips supplies well ahead of the mass production of any weapon systems. They need the real product and plenty of time for hardware and software development and testing. It means a leading edge 3nm chip coming to market next year may not be meaningful for most weapon systems rollout from production line next three to four years.

3. Most defense and national security related semiconductor demand are in low volume. For example, Lockheed Martin only produces up to 180 F35 a year. How many chips does Lockheed Martin need every year for 180 F35 production? Compare it to any commercial semiconductor products, it's really a tiny volume. It can partially explain why TSMC Arizona fab initial monthly production volume is only 20,000 12-inch wafers. For DoD or DoE, this number is more than enough already.

4. DoD and DoE recognized a while back that commercial semiconductor development are getting faster and faster and they have to take advantage on it. Their collaboration with foundries are wide and deep. Some public available information can be seen here:



A new development revealed today about the Polaris supercomputer project at Argonne National Laboratory. I mentioned it yesterday that the Aurora supercomputer project at Argonne is seriously behind the schedule due to Intel unable to deliver needed CPUs. In the news the Polaris seems to become a bridge system to help user community to prepare their software for the future arrival (don't know when) of the Aurora. But I'm wondering if there's any reason not to scale up the Polaris and use it to replace the role that Aurora was planned to perform?

Polaris is based on AMD's Rome and Milan CPUs and Nvidia's A100 GPUs. I think DoE is trying to downplay the significance of Polaris. But will DoE eventually kill the Aurora project or just let it fade away?

Intel must find a way to solve this credibility problem. It hurts Intel and hurts those people inside and outside of government who work hard to promote Intel's products and technologies.


 
Back
Top