Intel CEO optimistic about CHIPS Act’s future after trading texts with JD Vance

[Daniel Nenni]

Mark Bohr was an Intel Senior Fellow and he helped make decisions that led to high-k MG and finfets. Fellows have also helped with key technological development and thus truly are the best engineers at Intel. It’s a body blow to Intel to lose them.

Mar Bohr proves my point. Mark predicted the collapse of the fabless model which was ridiculous and the result of wearing IDM blinders his entire career. As a result the IDM business model has collapsed:

Intel says fabless model collapsing… really? - Semiwiki

There is an interesting discussion in the SemiWiki forum in response to the EETimes article: Intel exec says fabless model 'collapsing'. Definitely an interesting debate, one worth our time since the advertising click hungry industry pundits will certainly jump all over it. Clearly I’m biased...

semiwiki.com

 

[Daniel Nenni]

Which country has a larger consumer market? When you compare by Household final consumption expenditure, the US is roughly as large as the other top 5 combined.

The problem with Taiwan wafer tariffs is that it will hurt America, especially the semiconductor industry which is dominated by US companies who do not make wafers (Apple, Qualcomm, Broadcom, Nvidia, AMD, and even Intel). 60%+ of TSMC's revenue comes from the US. There is no way for 60% of TSMC's wafers to be made in the US. anytime soon. China tariffs I'm fine with. We have had a trade imbalance with China for many years, other countries as well. Level the playing field. You buy from us and we buy from you. The US consumer will not make that decision so the Federal Government must level things up, my opinion.

 

[Daniel Nenni]

Intel’s Chips Act funding may be safe under the upcoming change in the administration in the US reports a contributor to SemiWiki

Intel’s Chips Act money looking safe

Intel’s Chips Act funding may be safe under the upcoming change in the administration in the US reports a contributor to SemiWiki “Intel CEO Pat

www.electronicsweekly.com

 

[blueone]

Mar Bohr proves my point. Mark predicted the collapse of the fabless model which was ridiculous and the result of wearing IDM blinders his entire career. As a result the IDM business model has collapsed:

Intel says fabless model collapsing… really? - Semiwiki

There is an interesting discussion in the SemiWiki forum in response to the EETimes article: Intel exec says fabless model 'collapsing'. Definitely an interesting debate, one worth our time since the advertising click hungry industry pundits will certainly jump all over it. Clearly I’m biased...

semiwiki.com

I don't think it proves your point. Bohr is an engineer, down in the guts of process technology. That's where and how he got appointed being a Senior Fellow based on technical innovation. The fact that he was wrong about the efficacy of the foundry business model does not detract from his contributions in engineering and technologies. It would be nice if all of these technical leaders were equally skilled in business trends, but they often aren't. Nor are the people most aware of market and business trends usually technical leaders. IMO, Intel would arguably be technically less competent than it is today if Bohr was not an engineering leader.

 

[XYang2023]

The problem with Taiwan wafer tariffs is that it will hurt America, especially the semiconductor industry which is dominated by US companies who do not make wafers (Apple, Qualcomm, Broadcom, Nvidia, AMD, and even Intel). 60%+ of TSMC's revenue comes from the US. There is no way for 60% of TSMC's wafers to be made in the US. anytime soon. China tariffs I'm fine with. We have had a trade imbalance with China for many years, other countries as well. Level the playing field. You buy from us and we buy from you. The US consumer will not make that decision so the Federal Government must level things up, my opinion.

I think it is the other way around. Tariffs on wafers place financial penalties on U.S. companies, whose profit margins (except for Intel) are exceptionally high compared to other industries. The policy should encourage them to adopt U.S.-based fabs. Tariffs on China should be moderated, as their impacts can be felt by individual households, who are price-sensitive.

Another policy could be that if the wafers are produced in the U.S., they can be sold to China with fewer restrictions.

https://www.reuters.com/technology/us-revoked-some-export-licenses-chinas-huawei-2024-05-07/

 

[hist78]

The problem with Taiwan wafer tariffs is that it will hurt America, especially the semiconductor industry which is dominated by US companies who do not make wafers (Apple, Qualcomm, Broadcom, Nvidia, AMD, and even Intel). 60%+ of TSMC's revenue comes from the US. There is no way for 60% of TSMC's wafers to be made in the US. anytime soon. China tariffs I'm fine with. We have had a trade imbalance with China for many years, other countries as well. Level the playing field. You buy from us and we buy from you. The US consumer will not make that decision so the Federal Government must level things up, my opinion.

For those who proposed high tariff on semiconductor products imported into US, they often don't check some basic data before making such suggestion. Using Taiwan as an example, its semiconductor or electronic products (like PCs) import tax is very low, many of them are in zero. At the same time, Taiwan is a major market for US semiconductor companies. Taiwan, with the population of 23.5 million, has an out of proportion role in buying US semiconductor products.

For example, in 2023 Intel top four market by revenue are China ($14.85 billion, population 1.4 billion), US ($13.96B, 335 million), Singapore ($8.6B, 5.9 million) and Taiwan ($6.87B, 23.4 million). Source: statista.com.

There are very few unsettled trade issues between Taiwan and US. There is no case or there is no easy way for United States to impose high tariffs on semiconductor import without seriously damaging US itself.

That left the weapon system sales as one possible routes to increase US export to Taiwan. But the problem is that both countries tried very hard to do that already, only met with US defense manufacturers' slow and inefficient production performance. The military sales to Taiwan currently has reached a $20 ~ $24 billion backlog that both US Congress and executive branch are alarmed and frustrated.

TSMC cannot make 2nm chips abroad now: MOEA

Taiwan’s technology protection rules prohibits Taiwan Semiconductor Manufacturing Co (TSMC, 台積電) from producing 2-nanometer chips abroad, so the company must keep its most cutting-edge technology at home, Minister of Economic Affairs J.W. Kuo (郭智輝) said yesterday. Kuo made the remarks in...

semiwiki.com

Source: statista.com.

 

[samwilde]

Mar Bohr proves my point. Mark predicted the collapse of the fabless model which was ridiculous and the result of wearing IDM blinders his entire career. As a result the IDM business model has collapsed:

I agree that his point of view is ridiculous. As a result of Rock's Law the cost to build a fab will continue to increase which will make the IDM model increasingly untenable vs the Foundry model. In the 1980s a lot of computer manufacturers had their own manufacturing facilities operating as IDMs including Commodore, HP, TI, IBM, DEC, and others. And today none of the existing computer manufacturers have their own chip making facilities. The closest would be someone like Samsung but even they are failing.

The problem with Taiwan wafer tariffs is that it will hurt America, especially the semiconductor industry which is dominated by US companies who do not make wafers (Apple, Qualcomm, Broadcom, Nvidia, AMD, and even Intel). 60%+ of TSMC's revenue comes from the US. There is no way for 60% of TSMC's wafers to be made in the US. anytime soon. China tariffs I'm fine with. We have had a trade imbalance with China for many years, other countries as well. Level the playing field. You buy from us and we buy from you. The US consumer will not make that decision so the Federal Government must level things up, my opinion.

I agree with this as well. It is the same thing as the tariffs on Japanese DRAM in the 1980s. It hurts most of the computer industry in the US to benefit only a couple of players. If the US slaps tariffs on chip imports from Taiwan it will only hurt itself.

I think the best option is to do what is already happening. Level the playing field for US fabs with state aid, and possibly demand US fabbed components for computer purchases by the US government like China and Russia are doing. That will attract way less WTO scrutiny and will serve much the same purpose.

 

[XYang2023]

For those who proposed high tariff on semiconductor products imported into US, they often don't check some basic data before making such suggestion. Using Taiwan as an example, its semiconductor or electronic products (like PCs) import tax is very low, many of them are in zero. At the same time, Taiwan is a major market for US semiconductor companies. Taiwan, with the population of 23.5 million, has an out of proportion role in buying US semiconductor products.

For example, in 2023 Intel top four market by revenue are China ($14.85 billion, population 1.4 billion), US ($13.96B, 335 million), Singapore ($8.6B, 5.9 million) and Taiwan ($6.87B, 23.4 million). Source: statista.com.

There are very few unsettled trade issues between Taiwan and US. There is no case or there is no easy way for United States to impose high tariffs on semiconductor import without seriously damaging US itself.

That left the weapon system sales as one possible routes to increase US export to Taiwan. But the problem is that both countries tried very hard to do that already, only met with US defense manufacturers' slow and inefficient production performance. The military sales to Taiwan currently has reached a $20 ~ $24 billion backlog that both US Congress and executive branch are alarmed and frustrated.

TSMC cannot make 2nm chips abroad now: MOEA

Taiwan’s technology protection rules prohibits Taiwan Semiconductor Manufacturing Co (TSMC, 台積電) from producing 2-nanometer chips abroad, so the company must keep its most cutting-edge technology at home, Minister of Economic Affairs J.W. Kuo (郭智輝) said yesterday. Kuo made the remarks in...

semiwiki.com

View attachment 2451
Source: statista.com.

Tanks have proven to be ineffective in the recent war in Ukraine. Additionally, Taiwan is a mountainous island. Generally, people in Taiwan dislike such spending.

Taiwan to finally get first US Abrams tanks, but can they stand up to the PLA?

First 38 battle tanks arriving in December seen to boost cross-strait deterrence, but some analysts cite modern threats like suicide drones.

www.scmp.com

 

[blueone]

I agree that his point of view is ridiculous. As a result of Rock's Law the cost to build a fab will continue to increase which will make the IDM model increasingly untenable vs the Foundry model. In the 1980s a lot of computer manufacturers had their own manufacturing facilities operating as IDMs including Commodore, HP, TI, IBM, DEC, and others. And today none of the existing computer manufacturers have their own chip making facilities. The closest would be someone like Samsung but even they are failing.

Note that the video is from 2012. Situations change, perhaps Bohr's opinion has too.

 

[samwilde]

Note that the video is from 2012. Situations change, perhaps Bohr's opinion has too.

It was self-evident even then. AMD had spun out Globalfoundries already at that point.

 

[hist78]

I think it is the other way around. Tariffs on wafers place financial penalties on U.S. companies, whose profit margins (except for Intel) are exceptionally high compared to other industries. The policy should encourage them to adopt U.S.-based fabs. Tariffs on China should be moderated, as their impacts can be felt by individual households, who are price-sensitive.

Another policy could be that if the wafers are produced in the U.S., they can be sold to China with fewer restrictions.

https://www.reuters.com/technology/us-revoked-some-export-licenses-chinas-huawei-2024-05-07/

I think you may have overlooked the fact that the Republic of China (Taiwan) has been a U.S. ally for many years. The strong military, political, and economic relationship between them is well established and benefits both countries. No U.S. administration or Congress would attempt to damage this relationship; rather, they aim to strengthen it.

On the other hand, the Chinese Communist Party (CCP) and its forever leader, Chairman Xi, have become increasingly violent, oppressive, and aggressive both domestically and internationally. The bipartisan long-term US policy and actions toward CCP/PRC clearly aim at deterrence and containment, including those imposed on semiconductors.

 

[Artificer60]

That left the weapon system sales as one possible routes to increase US export to Taiwan. But the problem is that both countries tried very hard to do that already, only met with US defense manufacturers' slow and inefficient production performance. The military sales to Taiwan currently has reached a $20 ~ $24 billion backlog that both US Congress and executive branch are alarmed and frustrated.

My father served in the Navy. At one point in time rather than working on the aircraft in his squadron he was working on the production line at LTV, because they were having issues delivering their contracted volumes.

A slow, inefficient defense industry is nothing new. Outside of WWI,I when the whole nation was focused on national defense, I can't think of any time that hasn't been true.

 

[XYang2023]

I think you may have overlooked the fact that the Republic of China (Taiwan) has been a U.S. ally for many years. The strong military, political, and economic relationship between them is well established and benefits both countries. No U.S. administration or Congress would attempt to damage this relationship; rather, they aim to strengthen it.

On the other hand, the Chinese Communist Party (CCP) and its forever leader, Chairman Xi, have become increasingly violent, oppressive, and aggressive both domestically and internationally. The bipartisan long-term US policy and actions toward CCP/PRC clearly aim at deterrence and containment, including those imposed on semiconductors.

What are the driving factors behind the election outcome? The economy—specifically, how people feel about inflation—not ideology or foreign policies.

 

[blueone]

It was self-evident even then. AMD had spun out Globalfoundries already at that point.

I'm not sure I'm understanding your intent by calling out GF, but they aren't a good example of anything I can think of. In 2012 Intel had a lead in leading edge fab process, and I'm sure that colored the view of many engineers in Intel about whether a foundry model was better for Intel.

 

[samwilde]

I'm not sure I'm understanding your intent by calling out GF, but they aren't a good example of anything I can think. In 2012 Intel had a lead in leading edge fab process, and I'm sure that colored the view of many engineers in Intel about whether a foundry model was better for Intel.

AMD was the only competitor to Intel which still had its own fabs. When AMD sold their fabs that should have been a wake up call for anyone in the industry to figure out the future was the fabless model. While I am of the opinion that Intel should presently continue its IDM 2.0 strategy, which is similar to the "virtual gorilla" strategy AMD had when Jerry Sanders III was still heading AMD, in the long run even Intel won't be able to feed a cost effective fab with just its products.

 

[Daniel Nenni]

I don't think it proves your point. Bohr is an engineer, down in the guts of process technology. That's where and how he got appointed being a Senior Fellow based on technical innovation. The fact that he was wrong about the efficacy of the foundry business model does not detract from his contributions in engineering and technologies. It would be nice if all of these technical leaders were equally skilled in business trends, but they often aren't. Nor are the people most aware of market and business trends usually technical leaders. IMO, Intel would arguably be technically less competent than it is today if Bohr was not an engineering leader.

I met Mark Bohr at a conference, he sat next to me and we chatted. My take-away was that he was past his prime. This was during the whole 10nm debacle which you failed to mention. Intel 14nm was also a debacle but it was the first FinFET process so the delay and bad yield was much less of an issue than at 10nm.

Intel at 10nm ... a conversation with Mark Bohr of Intel
Intel's Process Architecture and Integration team have done what many have said couldn t be done: the first true 10nm in production. They have scaled another node with greater density and performance-per-watt. Most importantly, they did all this with lower cost-per-transistor. In other words, Moore’s Law is still alive and well. In this video ...

My point is that sometimes a fellow position is a window seat versus being in the trenches so I would not give more value to a fellow than an up and comer.

 

[blueone]

I met Mark Bohr at a conference, he sat next to me and we chatted. My take-away was that he was past his prime. This was during the whole 10nm debacle which you failed to mention. Intel 14nm was also a debacle but it was the first FinFET process so the delay and bad yield was much less of an issue than at 10nm.

I didn't mention the 10nm debacle because the video was from 2012. Also, I don't know anything about the sequence of decisions that led to the failure of 10nm.

My point is that sometimes a fellow position is a window seat versus being in the trenches so I would not give more value to a fellow than an up and comer.

And I haven't disagreed with you that some Fellows and Senior Fellows are not impressive, to be kind. Some starting right from their appointment to the position. I can also tell you, from first hand experience, you can get in a lot of trouble by disagreeing with some of them. I think you're being much more realistic though with this comment by adding the words "sometimes" as compared with your earlier post (#26).

 

[blueone]

AMD was the only competitor to Intel which still had its own fabs. When AMD sold their fabs that should have been a wake up call for anyone in the industry to figure out the future was the fabless model.

Not only isn't this a reasonable conclusion, Intel executives at the time seemed convinced that AMD divested from GF because of Intel's superior technology, superior management, and far superior level of potential investment.

While I am of the opinion that Intel should presently continue its IDM 2.0 strategy, which is similar to the "virtual gorilla" strategy AMD had when Jerry Sanders III was still heading AMD, in the long run even Intel won't be able to feed a cost effective fab with just its products.

I agree on these points. Personally, I think the merchant chip market is under attack by in-house designs and the dramatic continued growth of cloud computing. So I really think Intel needs to figure out the foundry business ASAP. IDM2.0 is only necessary to help fund the transition.

 

[hist78]

Tanks have proven to be ineffective in the recent war in Ukraine. Additionally, Taiwan is a mountainous island. Generally, people in Taiwan dislike such spending.

Taiwan to finally get first US Abrams tanks, but can they stand up to the PLA?

First 38 battle tanks arriving in December seen to boost cross-strait deterrence, but some analysts cite modern threats like suicide drones.

www.scmp.com

What are the driving factors behind the election outcome? The economy—specifically, how people feel about inflation—not ideology or foreign policies.

My father served in the Navy. At one point in time rather than working on the aircraft in his squadron he was working on the production line at LTV, because they were having issues delivering their contracted volumes.

A slow, inefficient defense industry is nothing new. Outside of WWI,I when the whole nation was focused on national defense, I can't think of any time that hasn't been true.

That's why US government and Congress are very worried about the product delivery timeline and quality. Some domestic suppliers are just too slow or often cancelled their products left and right before the scheduled arrival. Some of them even moved production to foreign countries.

One of the several strategies US adopted is to bring foreign suppliers to US soil as long as they are from those US allies. More competitions and more suppliers, no matter it's from UK, Norway, Italy, Taiwan, Japan, South Korea, or US, are very important in solving the problems.

 

[hist78]

I met Mark Bohr at a conference, he sat next to me and we chatted. My take-away was that he was past his prime. This was during the whole 10nm debacle which you failed to mention. Intel 14nm was also a debacle but it was the first FinFET process so the delay and bad yield was much less of an issue than at 10nm.

Intel at 10nm ... a conversation with Mark Bohr of Intel
Intel's Process Architecture and Integration team have done what many have said couldn t be done: the first true 10nm in production. They have scaled another node with greater density and performance-per-watt. Most importantly, they did all this with lower cost-per-transistor. In other words, Moore’s Law is still alive and well. In this video ...

My point is that sometimes a fellow position is a window seat versus being in the trenches so I would not give more value to a fellow than an up and comer.

@Daniel Nenni had a 5 years old post and discussion about Mark Bohr, link listed below. It's very interesting to read it five years later.

Intel's Mark Bohr to Retire

This is an interesting article. Kind of a fluff piece but an easy read. Remember, Mark was the one who said the fabless ecosystem was dead. I met Mark after that at a symposium. He sat right next to me and we chatted on and off for three days. He is a low key and humble guy despite what the PR...

semiwiki.com