DeWine calls on Biden administration to release Intel cash, says company won't leave Ohio

[hist78]

For those who are interested in the origination of CHIPS Act, US government strategy, and the story behind the TSMC Arizona project, I posted an 2022 EE Times interview with Keith Krach, former Under Secretary of State for Economic Growth, Energy, and the Environment in the Trump administration.

Architect of CHIPS Act Speaks on Its Impact

"In an exclusive interview with EE Times, Keith Krach, former Under Secretary of State for Economic Growth, Energy, and the Environment in the Trump administration, speaks on the significance of the CHIPS Act, which has since been passed by the House in a 243–187 vote." "Along with former...

semiwiki.com

 

[lilo777]

The bottom line is that, so far, CHIPS act has been a total failure. If anything, the US are farther behind other countries in advanced chip manufacturing today than they were two years ago. If it was not for political pressure (bans, sanctions etc.) even China would have already been ahead of US by now and the way US government "invests" in advanced semiconductor manufacturing, China will get ahead eventually.

 

[hist78]

The bottom line is that, so far, CHIPS act has been a total failure. If anything, the US are farther behind other countries in advanced chip manufacturing today than they were two years ago. If it was not for political pressure (bans, sanctions etc.) even China would have already been ahead of US by now and the way US government "invests" in advanced semiconductor manufacturing, China will get ahead eventually.

Is it a CHIPS Act failure or an Intel's failure?

One of the CHIPS Act top priorities is TSMC Arizona fab project. It will begin high volume production for AMD, Apple, Nvidia, Broadcom, and Qualcomm starting in Q1 2025.

Starting from scratch and now with 2,000 employees onsite, it's a major achievement for CHIPS Act, US Government, Congress, and TSMC.

 

[XYang2023]

Is it a CHIPS Act failure or an Intel's failure?

One of the CHIPS Act top priorities is TSMC Arizona fab project. It will begin high volume production for AMD, Apple, Nvidia, Broadcom, and Qualcomm starting in Q1 2025.

Starting from scratch and now with 2,000 employees onsite, it's a major achievement for CHIPS Act, US Government, Congress, and TSMC.

I would call it a failure for the U.S. If there were a conflict in the Taiwan Strait now, the TSMC fabs in Arizona could do very little. Additionally, TSMC continues to build sub-3nm fabs in Taiwan, with no geographic diversification.

 

[lilo777]

Is it a CHIPS Act failure or an Intel's failure?

One of the CHIPS Act top priorities is TSMC Arizona fab project. It will begin high volume production for AMD, Apple, Nvidia, Broadcom, and Qualcomm starting in Q1 2025.

Starting from scratch and now with 2,000 employees onsite, it's a major achievement for CHIPS Act, US Government, Congress, and TSMC.

Definitely one can't blame Intel for the failure of the US government. The top priority of the CHIPS act was to help domestic advanced chip manufacturing not the TSMC Arizona fab projects. Those FABs, by design, are going to stay two generations behind the most advanced node. They will be good for chips for outdated iPhone models. "Two generations back" used to be the red line for US government requirements on construction of FABs in China. Besides, advanced semiconductor manufacturing is impossible without advanced R&D and TSMC FABs contribute nothing in this regard.

 

[hist78]

Definitely one can't blame Intel for the failure of the US government. The top priority of the CHIPS act was to help domestic advanced chip manufacturing not the TSMC Arizona fab projects. Those FABs, by design, are going to stay two generations behind the most advanced node. They will be good for chips for outdated iPhone models. "Two generations back" used to be the red line for US government requirements on construction of FABs in China. Besides, advanced semiconductor manufacturing is impossible without advanced R&D and TSMC FABs contribute nothing in this regard.

The 2022 interview below with Keith Krach, former Under Secretary of State for Economic Growth, Energy, and the Environment in the Trump administration, can help us understand more about the CHIPS Act and the US government's strategy.

Architect of CHIPS Act Speaks on Its Impact

"In an exclusive interview with EE Times, Keith Krach, former Under Secretary of State for Economic Growth, Energy, and the Environment in the Trump administration, speaks on the significance of the CHIPS Act, which has since been passed by the House in a 243–187 vote." "Along with former...

semiwiki.com

 

[tedc]

Definitely one can't blame Intel for the failure of the US government. The top priority of the CHIPS act was to help domestic advanced chip manufacturing not the TSMC Arizona fab projects. Those FABs, by design, are going to stay two generations behind the most advanced node. They will be good for chips for outdated iPhone models. "Two generations back" used to be the red line for US government requirements on construction of FABs in China. Besides, advanced semiconductor manufacturing is impossible without advanced R&D and TSMC FABs contribute nothing in this regard.

The CHIPS Act was not implemented to only help US headquartered companies. And it's broader than just Intel (which does a large part of its R&D in Israel, not just in the US). The goal is to foster investments in semiconductor manufacturing in the US, which includes the broader ecosystem (like wafers, packaging, materials, equipment, R&D, etc.).

Getting industry leaders to invest on US soil -- that's what the goal is. TSMC has built the most advanced logic foundry in the US that's ready for high volume production and brought with it new facilities from Amkor, Global Wafers, etc. That's a "win" and we should take it.

Intel is just one of the many players vying for this funding. Yes, understand that the ideal scenario is for Intel to be a viable alternative for leading edge chips, but we can't take an all or nothing approach. The CHIPS Act is hedging and it hasn't been all roses, but there have been some successes and we should acknowledge them.

 

[samwilde]

The CHIPS Act was not implemented to only help US headquartered companies. And it's broader than just Intel (which does a large part of its R&D in Israel, not just in the US). The goal is to foster investments in semiconductor manufacturing in the US, which includes the broader ecosystem (like wafers, packaging, materials, equipment, R&D, etc.).

From what I understand Intel does process development in the US. That then that gets replicated (copy exactly) elsewhere. Including Israel.

Intel does conduct R&D in Israel but I think that is for chip design, software, and things like that.

 

[Artificer60]

From what I understand Intel does process development in the US. That then that gets replicated (copy exactly) elsewhere. Including Israel.

Intel does conduct R&D in Israel but I think that is for chip design, software, and things like that.

You would be correct. All actual process development on leading edge technology takes place in the US. Intel Oregon does the process development while Intel Arizona does the packaging development. Once a leading edge process is developed additional process enhancement/improvement may take place at other fabs, but the original development is all done in the US.

 

[lilo777]

The CHIPS Act was not implemented to only help US headquartered companies. And it's broader than just Intel (which does a large part of its R&D in Israel, not just in the US). The goal is to foster investments in semiconductor manufacturing in the US, which includes the broader ecosystem (like wafers, packaging, materials, equipment, R&D, etc.).

Getting industry leaders to invest on US soil -- that's what the goal is. TSMC has built the most advanced logic foundry in the US that's ready for high volume production and brought with it new facilities from Amkor, Global Wafers, etc. That's a "win" and we should take it.

Intel is just one of the many players vying for this funding. Yes, understand that the ideal scenario is for Intel to be a viable alternative for leading edge chips, but we can't take an all or nothing approach. The CHIPS Act is hedging and it hasn't been all roses, but there have been some successes and we should acknowledge them.

Just imagine that Intel/IFS goes the way of Global Foundries (i.e. decides to save money and concentrate on mature processes). That would be a national disaster. It would mean that US were left without advanced semicondactor R&D and manufacturing completely. With just TSMC FABs (and not the best ones), US would be as advanced in this domain as Ireland (with its Intel FAB) is now. Obviously it's an exaggeration but you get the point. If that were to happen, if US still decided to restore their semiconductor prowess, the government would have to spend hundreds of billions of dollars to achieve that. And they would have to come up with a permanent framework of incentives to make private investors interested in this industry. Alternatively, they can come up with sanctions against Taiwan and South Korea to deprive them semiconductor manufacturing hardware supplies. And it's not as ridiculous as one might think. That's exactly how US are trying to compete with China.

 

[tedc]

From what I understand Intel does process development in the US. That then that gets replicated (copy exactly) elsewhere. Including Israel.

Intel does conduct R&D in Israel but I think that is for chip design, software, and things like that.

Thanks for the correction. I was definitely thinking about the design teams in Israel.

Just imagine that Intel/IFS goes the way of Global Foundries (i.e. decides to save money and concentrate on mature processes). That would be a national disaster. It would mean that US were left without advanced semicondactor R&D and manufacturing completely. With just TSMC FABs (and not the best ones), US would be as advanced in this domain as Ireland (with its Intel FAB) is now. Obviously it's an exaggeration but you get the point. If that were to happen, if US still decided to restore their semiconductor prowess, the government would have to spend hundreds of billions of dollars to achieve that. And they would have to come up with a permanent framework of incentives to make private investors interested in this industry. Alternatively, they can come up with sanctions against Taiwan and South Korea to deprive them semiconductor manufacturing hardware supplies. And it's not as ridiculous as one might think. That's exactly how US are trying to compete with China.

Yes, understand the concern that TSMC and Samsung keep their most advanced processes in Taiwan and South Korea. Until 2021, the US (including Intel) was perfectly fine with that approach because those are allied countries.

Let's get back to the Ohio cash issue though: I find it hard to believe that the the CHIPS Act money is what's holding up construction here. If Intel had the commitments (internal and customer) to support the capacity, they would build it out. What they seem to be looking for is a loan to do other stuff, while they delay the Ohio fabs until 2026. That's a perfectly fine approach, but if another company (like Samsung) did that, we'd be criticizing the CHIPS Act for not disbursing funds appropriately.

 

[hist78]

Thanks for the correction. I was definitely thinking about the design teams in Israel.


Yes, understand the concern that TSMC and Samsung keep their most advanced processes in Taiwan and South Korea. Until 2021, the US (including Intel) was perfectly fine with that approach because those are allied countries.

Let's get back to the Ohio cash issue though: I find it hard to believe that the the CHIPS Act money is what's holding up construction here. If Intel had the commitments (internal and customer) to support the capacity, they would build it out. What they seem to be looking for is a loan to do other stuff, while they delay the Ohio fabs until 2026. That's a perfectly fine approach, but if another company (like Samsung) did that, we'd be criticizing the CHIPS Act for not disbursing funds appropriately.

The non-binding $8.5 billion CHIPS Act grants announcement for Intel covers Intel sites in Arizona, Ohio, Oregon, and New Mexico. I believe the majority of the $8.5 billion will go to Arizona, Ohio, and Oregon, with each site likely receiving around $2 billion to $3 billion in grant money.

Among these, Intel’s Fab 52 and 62 in Arizona are projected to cost about $30 billion (49% owns by private equity firm Brookfield), while the Ohio facility is expected to cost approximately $28 billion. This means that CHIPS Act grant funding will likely cover around 10% or less of the total project costs at each Intel site. This distribution reflects the CHIPS Act’s legislative intent: the grant money is designed to serve as an incentive and matching fund for approved projects, not as a primary funding source or to cover the majority of any project’s cost.

It’s also important to remember that each Intel fab requires 4 to 5, or even 6 years, to complete, and Intel does not pay the entire cost on day one.

What’s holding Intel back? What are the serious issues that the CHIPS Program Office has identified?

 

[tedc]

What’s holding Intel back? What are the serious issues that the CHIPS Program Office has identified?

Some helpful context in a recent NYT article (The White House Bet Big on Intel. Will It Backfire?, 10/24/2024):

"... Intel’s business has recently worsened, causing concern on Capitol Hill and stoking the fears of some administration officials who have worried for years about the company’s ability to deliver on its promises."

"To prevent taxpayer money from being wasted, the administration created certain milestones for the companies it would subsidize. Companies must hit benchmarks — like building a plant, producing chips or signing up customers — before they can receive portions of their funds. Like most companies, Intel has not yet received its first financial infusion under the program. And in recent months, it has been sparring with the government over the milestones it must hit before it gets funding, five people familiar with the talks said."

"As they designed the chips program, Ms. Raimondo and Biden officials weighed how to support Intel while safeguarding taxpayer money. They were already questioning whether Intel could follow through on its promises, several officials said."

"Caitlin Legacki, a former Commerce Department official, said there also was a fear that “Intel is going to take chips money, build an empty shell of a factory and then never actually open it, because they don’t have customers.” This is why the program was designed to withhold payments until companies hit certain milestones, she said."

 

[lilo777]

Some helpful context in a recent NYT article (The White House Bet Big on Intel. Will It Backfire?, 10/24/2024):

"... Intel’s business has recently worsened, causing concern on Capitol Hill and stoking the fears of some administration officials who have worried for years about the company’s ability to deliver on its promises."

"To prevent taxpayer money from being wasted, the administration created certain milestones for the companies it would subsidize. Companies must hit benchmarks — like building a plant, producing chips or signing up customers — before they can receive portions of their funds. Like most companies, Intel has not yet received its first financial infusion under the program. And in recent months, it has been sparring with the government over the milestones it must hit before it gets funding, five people familiar with the talks said."

"As they designed the chips program, Ms. Raimondo and Biden officials weighed how to support Intel while safeguarding taxpayer money. They were already questioning whether Intel could follow through on its promises, several officials said."

"Caitlin Legacki, a former Commerce Department official, said there also was a fear that “Intel is going to take chips money, build an empty shell of a factory and then never actually open it, because they don’t have customers.” This is why the program was designed to withhold payments until companies hit certain milestones, she said."

Let's not forget that this is the government that has been watching for decades as the US leadership in semiconductor industry (manufacturing specifically) was being squandered. US advanced semiconductor manufacturers folded one by one. Intel is the last one standing. If Intel fails, the fact that the government saved $8B dollars won't help anyone. It will be a complete and total failure of the US as a country. These government officials are just the clerks fearing for their careers. Considering the scale of the industry and the problems, one time $50B payment was a laughable sum to begin with. If the US lose the ability to produce advanced silicon, government investment in the broader ecosystem ("wafers, packaging, materials, equipment") will be wasted as well. There won't be anything to package.

 

[MKWVentures]

Money needs to be a rebate. Intel buys, then government re-imburses. Intel is not spending in Ohio .... when they spend, they will get money

Money up front is a horrible idea for large corporations. I literally had this discussion with people in Washington DC this week

 

[tedc]

Let's not forget that this is the government that has been watching for decades as the US leadership in semiconductor industry (manufacturing specifically) was being squandered. US advanced semiconductor manufacturers folded one by one. Intel is the last one standing. If Intel fails, the fact that the government saved $8B dollars won't help anyone. It will be a complete and total failure of the US as a country. These government officials are just the clerks fearing for their careers. Considering the scale of the industry and the problems, one time $50B payment was a laughable sum to begin with. If the US lose the ability to produce advanced silicon, government investment in the broader ecosystem ("wafers, packaging, materials, equipment") will be wasted as well. There won't be anything to package.

The incentives and cost structure clearly aren't there yet. Even Intel chose to produce its most advanced silicon in Ireland instead of the US.

 

[tedc]

Money needs to be a rebate. Intel buys, then government re-imburses. Intel is not spending in Ohio .... when they spend, they will get money

Money up front is a horrible idea for large corporations. I literally had this discussion with people in Washington DC this week

I don't see the milestone approach as unreasonable at all. Short of giving out free cash, what are other options? Does the US government co-invest in new fabs like Brookfield and Apollo did?

 

[MKWVentures]

I don't see the milestone approach as unreasonable at all. Short of giving out free cash, what are other options? Does the US government co-invest in new fabs like Brookfield and Apollo did?

It is milestone approach and it should be IMO. Other people were indicating there should be some pre-funding (never do prefunding to a large company). US is not open to co-investment at this time.

The best idea is ALWAYS to give a rebate on purchases within 1 quarter subject to staffing requirements. If requirements are not met, Intel pays a fine .... this is the way many state subsidies work and the way many other countries use subsidies