Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/8-months-in-what-is-happening-with-bidens-chips-act.17780/
        )

    [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
)

8 Months in, What is Happening with Biden's CHIPS Act?

benb

Well-known member
Excerpts:
Mike Schmidt, Director of the CHIPS program office
Todd Fisher, Chief Investment Officer of the CHIPS program office
$39 billion is for manufacturing incentives $11 billion is for R&D
Todd and I are on that incentive side, on that $39 billion side.
We view ourselves fundamentally as managing a $39 billion investment program on behalf of taxpayers.

"Number one, we need to make a set of investments that are advancing our economic and national security. So we have a set of strategic objectives, but then we have to do so in a way that is safeguarding taxpayer dollars and really getting good deals. So we built our office here around chief strategy officer managing an office of strategy, technology and policy. That's Morgan Dwyer. She comes from the senior leadership of the Defense Department and the Office of Science Technology Policy.

And then Todd, bringing his considerable financial commercial expertise as the investment office, you know, gives us a way of kind of like mediating a conversation and coming up with an overall approach that is going to going to advance both the taxpayer protection side of the equation, but also the kind of strategic and economic."

"And so we say in leading edge logic, we want to have at least two large scale leading edge logic manufacturing clusters. We talk about the importance of advanced packaging, which is absolutely critical, both from a supply chain resiliency standpoint, but also from a technology leadership standpoint. We talk about memory, we talk about current generation to mature."

"And so the success will be to see that number instead of continuing to go down, to start to go up. And then in the very specific areas that we've already laid out, meaning leading edge logic where we now today produce 0% of the world's chips in leading edge, we've said very clearly we want to see two self-sustaining ecosystems at a minimum, and we have no real advanced packaging in this country either."

"We want to see quite robust aspects there, and we want to start to see cost competitive memory plants being done in this country. And we want to see all the pieces of the ecosystem, including suppliers, come together around those goals. And so that's how we're thinking about measuring ourselves."

"Yeah, we'll do like a review of the application and we'll decide, you know, based on the criteria we've laid out based on our economic and national security objectives, this is a project we want to support that will lead us to extending what we call a preliminary memorandum of terms, but basically a term sheet, right? And then you're in a commercial relationship with the company where we're going to try to align on a set of economic terms that allows a project or a set of projects to move forward. If we can align, then we'll do our due diligence and we'll move towards making that award."

"The other thing I would just say is on the kind of engineering side of things, this is just a personal view, but I think we as a country would be really better off if more engineers were interested in working on semiconductors, as opposed to say software. And I think all of of us can, you know, play a role in trying to shape that narrative. I mean, semiconductors, it is really the peak of human innovative capacity. And I think, you know, we need to make semiconductors cool here."

"The large leading edge fabs that are written mostly about, are already focused on certain geographic locations. That is where many of them have even started to construct on those locations. And we hope and expect to see, you know, broad ecosystems of suppliers and customers and others, you know, build around those ecosystems. However, that is not all of our money.

We also are trying to support current and mature nodes of semiconductors. Some of the things that really cause some of the slowdowns in the pandemic – auto, healthcare, MRI machines, etc., that are not like the leading edge of logic, but are, you know, mature nodes that have been produced for a while but are still really necessary to produce lots and lots of different consumer products and auto products and communications products."

"And we are ultimately going to take a portfolio level approach, right? Where we're going to try to maximize what we can achieve on the leading edge side, but also we hope to see really strong applicants on the current generation and mature side that are meaningfully contributing to supply chain resiliency and meaningfully providing that kind of resilience to critical industries like critical infrastructure like auto, like the defense, industrial base obviously being a critically important factor there from a national security perspective."

"But private investment is going to be driving this, right? And so fundamentally, when it comes to the kind of supply demand dynamics of any particular fab or any particular award, we're going to do our due diligence, as Todd said, we think that's core to our job, but we're going to be investing in a sense alongside private capital who will also have a measure of confidence that yes, there will be ups and downs as there always are in this industry, but that the overall demand profile is going to lead to healthy and self-sustaining dynamics."
 
"...leading edge logic ..., we've said very clearly we want to see two self-sustaining ecosystems at a minimum"

What follows is pure speculation. The US needs to be prepared for war, by possessing two domestic sources for all the pieces of a manufacturing ecosystem, from R&D to test to packaging. I think it is less likely that Samsung or TSMC could be part of one of these ecosystems.

First, TSMC and Samsung manufacturing ecosystems get their technology from Korea and Taiwan, respectively. That is a key weakness in a war. A sustainable manufacturing ecosystem needs to have R&D in the US. Otherwise it's not sustainable for long.

Second, TSMC and Samsung are foreign. US taxpayer money benefitting foreign corporations could produce a backlash from US voters who are footing the bill. The US domestic politics of supporting TSMC and Samsung are not as good as supporting Intel, by far.

Third, Samsung and TSMC (and Intel as well, I think) do their test and packaging overseas. Intel must have at least a little test and packaging in the US, for R&D.

So, we can be pretty sure Intel is in, but who is #2? Could AMD be planning to get back into manufacturing? Or will Samsung, or TSMC, relocate R&D, test and packaging to the US, creating a complete ecosystem? Would Korea or Taiwan countenance this?
 
"...leading edge logic ..., we've said very clearly we want to see two self-sustaining ecosystems at a minimum"

What follows is pure speculation. The US needs to be prepared for war, by possessing two domestic sources for all the pieces of a manufacturing ecosystem, from R&D to test to packaging. I think it is less likely that Samsung or TSMC could be part of one of these ecosystems.

First, TSMC and Samsung manufacturing ecosystems get their technology from Korea and Taiwan, respectively. That is a key weakness in a war. A sustainable manufacturing ecosystem needs to have R&D in the US. Otherwise it's not sustainable for long.

Second, TSMC and Samsung are foreign. US taxpayer money benefitting foreign corporations could produce a backlash from US voters who are footing the bill. The US domestic politics of supporting TSMC and Samsung are not as good as supporting Intel, by far.

Third, Samsung and TSMC (and Intel as well, I think) do their test and packaging overseas. Intel must have at least a little test and packaging in the US, for R&D.

So, we can be pretty sure Intel is in, but who is #2? Could AMD be planning to get back into manufacturing? Or will Samsung, or TSMC, relocate R&D, test and packaging to the US, creating a complete ecosystem? Would Korea or Taiwan countenance this?
I agree about TSMC and Samsung R&D.

Agree that for defense purposes Intel needs to drag some ecosystem companies to the US.

Intel and GF are probably the two target ecosystems for leading edge logic. Though GF would need a jumpstart to get there, and I'm not sure how that happens.
 
Excerpts:
Mike Schmidt, Director of the CHIPS program office
Todd Fisher, Chief Investment Officer of the CHIPS program office
$39 billion is for manufacturing incentives $11 billion is for R&D
Todd and I are on that incentive side, on that $39 billion side.
We view ourselves fundamentally as managing a $39 billion investment program on behalf of taxpayers.

"Number one, we need to make a set of investments that are advancing our economic and national security. So we have a set of strategic objectives, but then we have to do so in a way that is safeguarding taxpayer dollars and really getting good deals. So we built our office here around chief strategy officer managing an office of strategy, technology and policy. That's Morgan Dwyer. She comes from the senior leadership of the Defense Department and the Office of Science Technology Policy.

And then Todd, bringing his considerable financial commercial expertise as the investment office, you know, gives us a way of kind of like mediating a conversation and coming up with an overall approach that is going to going to advance both the taxpayer protection side of the equation, but also the kind of strategic and economic."

"And so we say in leading edge logic, we want to have at least two large scale leading edge logic manufacturing clusters. We talk about the importance of advanced packaging, which is absolutely critical, both from a supply chain resiliency standpoint, but also from a technology leadership standpoint. We talk about memory, we talk about current generation to mature."

"And so the success will be to see that number instead of continuing to go down, to start to go up. And then in the very specific areas that we've already laid out, meaning leading edge logic where we now today produce 0% of the world's chips in leading edge, we've said very clearly we want to see two self-sustaining ecosystems at a minimum, and we have no real advanced packaging in this country either."

"We want to see quite robust aspects there, and we want to start to see cost competitive memory plants being done in this country. And we want to see all the pieces of the ecosystem, including suppliers, come together around those goals. And so that's how we're thinking about measuring ourselves."

"Yeah, we'll do like a review of the application and we'll decide, you know, based on the criteria we've laid out based on our economic and national security objectives, this is a project we want to support that will lead us to extending what we call a preliminary memorandum of terms, but basically a term sheet, right? And then you're in a commercial relationship with the company where we're going to try to align on a set of economic terms that allows a project or a set of projects to move forward. If we can align, then we'll do our due diligence and we'll move towards making that award."

"The other thing I would just say is on the kind of engineering side of things, this is just a personal view, but I think we as a country would be really better off if more engineers were interested in working on semiconductors, as opposed to say software. And I think all of of us can, you know, play a role in trying to shape that narrative. I mean, semiconductors, it is really the peak of human innovative capacity. And I think, you know, we need to make semiconductors cool here."

"The large leading edge fabs that are written mostly about, are already focused on certain geographic locations. That is where many of them have even started to construct on those locations. And we hope and expect to see, you know, broad ecosystems of suppliers and customers and others, you know, build around those ecosystems. However, that is not all of our money.

We also are trying to support current and mature nodes of semiconductors. Some of the things that really cause some of the slowdowns in the pandemic – auto, healthcare, MRI machines, etc., that are not like the leading edge of logic, but are, you know, mature nodes that have been produced for a while but are still really necessary to produce lots and lots of different consumer products and auto products and communications products."

"And we are ultimately going to take a portfolio level approach, right? Where we're going to try to maximize what we can achieve on the leading edge side, but also we hope to see really strong applicants on the current generation and mature side that are meaningfully contributing to supply chain resiliency and meaningfully providing that kind of resilience to critical industries like critical infrastructure like auto, like the defense, industrial base obviously being a critically important factor there from a national security perspective."

"But private investment is going to be driving this, right? And so fundamentally, when it comes to the kind of supply demand dynamics of any particular fab or any particular award, we're going to do our due diligence, as Todd said, we think that's core to our job, but we're going to be investing in a sense alongside private capital who will also have a measure of confidence that yes, there will be ups and downs as there always are in this industry, but that the overall demand profile is going to lead to healthy and self-sustaining dynamics."

"And so we say in leading edge logic, we want to have at least two large scale leading edge logic manufacturing clusters."

I guess there will be three clusters, Phoenix Arizona, Austin Texas, and Albany New York to begin with. Each of these three has strong industrial, research, academic, and labor pool for semiconductor ecosystem.

The federal government may also do something around intel's Ohio site in New Albany, Ohio. But it's a long way for Ohio to become a cluster that is as strong as other top three clusters mentioned above.
 
Last edited:
"...leading edge logic ..., we've said very clearly we want to see two self-sustaining ecosystems at a minimum"

What follows is pure speculation. The US needs to be prepared for war, by possessing two domestic sources for all the pieces of a manufacturing ecosystem, from R&D to test to packaging. I think it is less likely that Samsung or TSMC could be part of one of these ecosystems.

First, TSMC and Samsung manufacturing ecosystems get their technology from Korea and Taiwan, respectively. That is a key weakness in a war. A sustainable manufacturing ecosystem needs to have R&D in the US. Otherwise it's not sustainable for long.

Second, TSMC and Samsung are foreign. US taxpayer money benefitting foreign corporations could produce a backlash from US voters who are footing the bill. The US domestic politics of supporting TSMC and Samsung are not as good as supporting Intel, by far.

Third, Samsung and TSMC (and Intel as well, I think) do their test and packaging overseas. Intel must have at least a little test and packaging in the US, for R&D.

So, we can be pretty sure Intel is in, but who is #2? Could AMD be planning to get back into manufacturing? Or will Samsung, or TSMC, relocate R&D, test and packaging to the US, creating a complete ecosystem? Would Korea or Taiwan countenance this?


IMO, "two self-sustaining ecosystems at a minimum" means TSMC and Global Foundries plus a rising star/inspiring new comer - Intel.

DoD, DOE, NASA and US industries have too many existing and critical relationship with TSMC and GlobalFoundries. Strengthen TSMC's and GlobalFoundries' ecosystems in US is a very efficient, logical, and practical approach.

Also TSMC's ecosystem has many major American companies as its members such as the top three US semi equipment suppliers (Applied Materials, LAM Research, KLA-Tencor), top 4 EDA companies (Candence, Synopsys, Siemens/Mentor Graphics, Ansys), 5 US companies of the top 6 fabless companies in the world (Qualcomm, Nvidia, Broadcom, AMD, Marvell), and countless US IP companies.

There is no reason for US government to skip such powerful and efficient ecosystem led by TSMC.
 
Excerpts:
Mike Schmidt, Director of the CHIPS program office
Todd Fisher, Chief Investment Officer of the CHIPS program office
$39 billion is for manufacturing incentives $11 billion is for R&D
Todd and I are on that incentive side, on that $39 billion side.
We view ourselves fundamentally as managing a $39 billion investment program on behalf of taxpayers.

"Number one, we need to make a set of investments that are advancing our economic and national security. So we have a set of strategic objectives, but then we have to do so in a way that is safeguarding taxpayer dollars and really getting good deals. So we built our office here around chief strategy officer managing an office of strategy, technology and policy. That's Morgan Dwyer. She comes from the senior leadership of the Defense Department and the Office of Science Technology Policy.

And then Todd, bringing his considerable financial commercial expertise as the investment office, you know, gives us a way of kind of like mediating a conversation and coming up with an overall approach that is going to going to advance both the taxpayer protection side of the equation, but also the kind of strategic and economic."

"And so we say in leading edge logic, we want to have at least two large scale leading edge logic manufacturing clusters. We talk about the importance of advanced packaging, which is absolutely critical, both from a supply chain resiliency standpoint, but also from a technology leadership standpoint. We talk about memory, we talk about current generation to mature."

"And so the success will be to see that number instead of continuing to go down, to start to go up. And then in the very specific areas that we've already laid out, meaning leading edge logic where we now today produce 0% of the world's chips in leading edge, we've said very clearly we want to see two self-sustaining ecosystems at a minimum, and we have no real advanced packaging in this country either."

"We want to see quite robust aspects there, and we want to start to see cost competitive memory plants being done in this country. And we want to see all the pieces of the ecosystem, including suppliers, come together around those goals. And so that's how we're thinking about measuring ourselves."

"Yeah, we'll do like a review of the application and we'll decide, you know, based on the criteria we've laid out based on our economic and national security objectives, this is a project we want to support that will lead us to extending what we call a preliminary memorandum of terms, but basically a term sheet, right? And then you're in a commercial relationship with the company where we're going to try to align on a set of economic terms that allows a project or a set of projects to move forward. If we can align, then we'll do our due diligence and we'll move towards making that award."

"The other thing I would just say is on the kind of engineering side of things, this is just a personal view, but I think we as a country would be really better off if more engineers were interested in working on semiconductors, as opposed to say software. And I think all of of us can, you know, play a role in trying to shape that narrative. I mean, semiconductors, it is really the peak of human innovative capacity. And I think, you know, we need to make semiconductors cool here."

"The large leading edge fabs that are written mostly about, are already focused on certain geographic locations. That is where many of them have even started to construct on those locations. And we hope and expect to see, you know, broad ecosystems of suppliers and customers and others, you know, build around those ecosystems. However, that is not all of our money.

We also are trying to support current and mature nodes of semiconductors. Some of the things that really cause some of the slowdowns in the pandemic – auto, healthcare, MRI machines, etc., that are not like the leading edge of logic, but are, you know, mature nodes that have been produced for a while but are still really necessary to produce lots and lots of different consumer products and auto products and communications products."

"And we are ultimately going to take a portfolio level approach, right? Where we're going to try to maximize what we can achieve on the leading edge side, but also we hope to see really strong applicants on the current generation and mature side that are meaningfully contributing to supply chain resiliency and meaningfully providing that kind of resilience to critical industries like critical infrastructure like auto, like the defense, industrial base obviously being a critically important factor there from a national security perspective."

"But private investment is going to be driving this, right? And so fundamentally, when it comes to the kind of supply demand dynamics of any particular fab or any particular award, we're going to do our due diligence, as Todd said, we think that's core to our job, but we're going to be investing in a sense alongside private capital who will also have a measure of confidence that yes, there will be ups and downs as there always are in this industry, but that the overall demand profile is going to lead to healthy and self-sustaining dynamics."

The Bloomberg interview audio can be heard at the link below:

 
It's the R&D part of the ecosystem that is the difficult part. TSMC and Samsung would like to take US money without relocating research, because this is their lifeblood. That is likely to be the crux of negotiation. I don't think the US will accept being a leading-edge technology colony of Korea and Taiwan.
 
As predicted, TSMC would like $15B from the USG CHIPS Act, but finds the reporting, auditing, and profit-sharing requirements unacceptable.


“Some of the conditions are unacceptable and we aim to mitigate any negative impact from these and will continue discussions with the U.S. government,” Mr. Liu told attendees at an industry meeting on March 30 in Taiwan.

TSMC is concerned that the economics of the Arizona project may not work if its potential profit is capped by the government, and it also sees problems calculating the profit of one or two factories in a global manufacturing operation, according to people familiar with the company’s stance in talks with the U.S.

The government’s demands for extensive access to TSMC’s books and operations are another sticking point, especially in an industry where companies tend to keep secret basic facts like who their customers are.

With all of those tales TSMC is telling about costs in the US, I wonder how the profit-sharing objectives would be set? Either the Biden administration changes their tune (unlikely), or this doesn't end well for TSMC.
 
While these CHIPS bozos fixate on "safeguarding taxpayer dollars" and "getting good deals" and "making semiconductors cool", GF and Intel are prepping more mass layoffs as we speak because the money won't arrive in time for the next quarterly report.

Daddy Reagan would be rolling in his grave to see how the "small government" "free market" ideology that dismantled tax revenue from corporations and allowed them to outsource most of our productive capacity has led to the gov't directly meddling in their books and operations to "share" in their profits. Defense contractors and banks have pilfered trillions of taxpayer dollars over the last decades just to immiserate people with pointless wars and debt, but for a real productive industry that requires an empowered domestic labor force it's suddenly important to put up a bunch of red tape and perform audits. Oh and that massive growing market in Asia that everyone was counting on for long term viability is now totally off limits. This should go well.
 
I wish the stock markets were this predictable. Sorry for the WSJ reference (it requires a subscription).


I'm still sticking to my prediction that none of the big fab companies take the CHIPS money. The Biden administration, and Congress for that matter, is going to look awfully silly if the only thing that comes out of the Act is the semiconductor equivalent of the NIH, which I think will be useless.
 
Samsung and Hynix have a 1 year waiver on their Xi’an fabs, reported October 2022, which will expire or renew. If not renewed, they will be cut off from US tech like the rest of China. This may lead them to explore relocating somewhere else, possibly the US, possibly gaining assistance from the CHIPs act.

Edit: Reading deeper in the article, Hynix also has ex-Intel Dalian fab and older Wuxi fabs which may fall under the export ban.

 
Last edited:
Back
Top