Concerns Grow in Washington over Intel

[Daniel Nenni]

Any Fab investment in Israel is Risky Biz those fabs are Intel 7 so useless for anyone outside Intel

Intel reportedly scaling back R&D teams in Israel — several hundred talented staff will be laid off​

Credit: Intel
Intel plans to reduce its workforce by several hundred employees at its Israeli research and development centers, following similar job cuts at its facilities elsewhere, reports DigiTimes. This move is part of the company's general cost-cutting measures amid financial challenges, and it opens opportunities for competitors to attract Intel's talent. There is a catch, though, as dozens of Intel employees already left the company for Nvidia earlier this year, according to Globes.

Intel's Israeli operations employ around 11,700 people, including 7,800 in R&D roles and 3,900 in manufacturing, with the latest layoffs expected to focus on R&D while leaving manufacturing staff largely intact as Intel readies for the completion of Fab 38 in Kiryat Gat, the report says. The exact number of people that will either be proposed to leave voluntarily, or retire, or just laid off is not known, but Globes reports that Intel is set to cut its Israeli workforce by several hundred positions.

Intel has three R&D sites in Israel. Each of these Israeli centers has a specialized focus: Haifa works on CPUs, AI hardware, and software; Petah Tikva develops communications and AI solutions; and Jerusalem is involved in communications, software, and cybersecurity. The report does not specify which R&D sites will see more cuts, other than saying that layoffs are expected in these R&D centers. But let us try to detail at least one thing.

Intel's Israeli R&D team is responsible for multiple breakthrough microarchitectures, such as Banias, Yonah/Merom, and Nehalem, just to name a few. Each of these was strategically important for Intel at some point. In fact, the Banias processor was the heart of Intel's Centrino platform, the company's first platform developed specifically for laptops, which pretty much revolutionized the market. Yonah/Merom put Intel back on the map of high-performance PCs and replaced the power-hungry Netburst microarchitecture that powered Pentium 4 processors.

But while Intel itself is slashing its R&D talent in Israel, it should be noted that people are leaving the company, too. Nvidia has been hiring former Intel employees: in 2024 alone, at least 30 Intel employees moved to Nvidia's offices in Yokneam and Tel Aviv, with total hires from Intel estimated to range from 60 to 90, according to the Globes report which analyzed LinkedIn profiles. Nvidia is expected to hire more Intel employees as the layoffs continue, possibly bringing the number to about 100. As of June 30, Nvidia reported a workforce of 4,000 employees in Israel, reflecting the company’s growing presence in the region.

Many Intel employees who moved to Nvidia are benefiting from more attractive compensation packages. For instance, a junior hardware engineer at Nvidia can expect a starting annual salary of NIS 566,000 (around $151,500), about 33% higher than the equivalent role at Intel. Nvidia also offers significantly more valuable share options compared to Intel, with packages starting at NIS 56,200 annually ($15,045).

In addition to Nvidia, other major tech companies are also drawing former Intel employees. Apple, Amazon, and Intel's subsidiary Mobileye have recently hired from Intel's workforce in Israel. Some groups have joined Microsoft, Google, and Huawei, which operates a development center in Haifa.

https://www.yahoo.com/finance/news/intel-reportedly-scaling-back-r-154500411.html

 

[hist78]

TSMC receives substantial support from Taiwan's government.

Latest interview with the vice president of the Chung-Hua Institution for Economic Research:

Can you tell me the amount of subsides TSMC received from the Taiwanese government?

 

[XYang2023]

Can you tell me the amount of subsides TSMC received from the Taiwanese government?

It’s not just subsidies; TSMC receives comprehensive support, including diplomatic assistance. He also suggested substantial subsidies in areas such as green technology, AI, EVs, and more. The fact that TSMC consumes around 10% of Taiwan’s electricity is remarkable. TSMC and Taiwan's government are closely interconnected.

 

[XYang2023]

Does TSMC, one of the world's top ten companies, still need subsidies from Taiwan? Debate sparks between National Taiwan University and National Taipei University of Business.

Li-wen Chen
April 20, 2023

"Does TSMC, one of the world's top ten companies by market capitalization and the leader in semiconductor foundry, still need subsidies from Taiwan?" During the pandemic and trade war, TSMC successfully established a "silicon shield" with its advanced processes, earning it the nickname "Taiwan’s national treasure." TSMC enjoys government tax incentives and affordable utilities, yet over 70% of its shares are held by foreign investors, meaning that 70% of its dividends go abroad. The Global Asset Management Research Center at the National Taipei University of Business conducted a public opinion survey and is organizing an intercollegiate debate between National Taiwan University and National Taipei University of Business on May 21, aiming to present diverse perspectives on economic balance.

Among the arguments supporting continued subsidies for TSMC is the idea that "Taiwan cannot live without TSMC." From an economic standpoint, Taiwan’s semiconductor industry accounts for approximately 15% of the country’s GDP, with TSMC contributing 7.3% of GDP. In terms of employment, TSMC employs 58,543 people globally, with 46,130 in Taiwan, representing 38% of the nation’s semiconductor workforce. Moreover, TSMC is a global leader in advanced manufacturing processes, particularly in 3nm and 2nm technologies.

However, opponents argue that TSMC's shareholder structure shows that 72% of its shareholders are foreign investors. The company has long enjoyed tax benefits and consumes 6% of Taiwan's electricity. With the upcoming production ramp-up at new facilities in southern Taiwan, TSMC’s power consumption is expected to rise to between 12.5% and 16%.

From a tax incentive perspective, Taiwan's income tax rate was only 11% for the past two years, though the corporate tax rate is now 20%. TSMC receives approximately NT$60 billion in tax breaks annually, which could instead support NT$6,000 per household for Taiwan’s nine million families or fund more startups. Additionally, TSMC dominates the Taiwan Stock Exchange, accounting for 30% of its total market capitalization. This attracts passive investment funds and crowds out other Taiwanese companies, as no other single stock holds a 25% weight in the national index.

Dr. Sheng-yuan Chen, Director of the Global Asset Management Research Center and Distinguished Professor of Finance at National Taipei University of Business, stated that although TSMC is the world’s largest semiconductor company by market value, its large-scale expansions have caused labor and material shortages in the construction industry. He suggests considering the potential benefits of redirecting subsidies from TSMC to other Taiwanese industries.

Chairman of Yuanta Funds, Tsung-sheng Liu, noted that debates on public policy issues among young students encourage the younger generation to offer diverse perspectives on Taiwan's economic future. Both active and passive funds focus on TSMC as a target stock, and institutional investors pay close attention to policies impacting TSMC.

Nevertheless, Hao-chung Kuo, Director of Semiconductor Research at Foxconn, pointed out that Taiwan’s semiconductor industry is the result of decades of experience, primarily in wafer foundry and IC design, with the latter still dominated by the United States. Although TSMC’s advanced technology makes it desirable for U.S. manufacturing, Taiwan's semiconductor industry must continue to invest in talent to maintain its lead in the ICT sector. Major industries should position themselves well and ensure proper talent allocation for diversification.

全球十強台積電還需台灣來補助？ 台大台北商大辯論鳴槍

全球市值前十大、晶圓代工龍頭廠台積電，以先進製程成功在疫情貿易戰期間，建立的「矽盾」獲得護國神山美名。台積電享有政府租稅優惠、便宜水電，但外資持股達7成，等於股利7成都被外國人領走，臺北商業大學黃石全球資產管理研究中心透過公開民調，以及臺灣大學和臺北商業大學5月21日校際辯論賽，為台積電用電、人才的磁吸效應外，在經濟平衡上呈現多元看法。

tw.news.yahoo.com

 

[Daniel Nenni]

Can you tell me the amount of subsides TSMC received from the Taiwanese government?

I do not think that number is calculable. It has been 30+ years. I am sure it is significant. Maybe another reason for TSMC to build fabs outside of Taiwan is to get more incentives from the Taiwan Government?

 

[delong.height]

I do not think that number is calculable. It has been 30+ years. I am sure it is significant. Maybe another reason for TSMC to build fabs outside of Taiwan is to get more incentives from the Taiwan Government?

The 2nm fabs I heard is what TSMC get in subsidy as well, subsidies in Asia is quite unique, all government officials in local or state are super cooperated and want to make sure that businesses receive good support. Labor, talents, housing, transportation, utility, and many many other thing. I believed that in western world, we like to talk about how much money the companies are getting from the policymakers. But in China, same as in Taiwan, it covers every aspect of society, money is one but not that important. The entire society is engaging in making company happier in order to get much private investment. This is especially the case with China, but Taiwan isn't any way far from it. And I believed that there are many examples. Foxconn with zhengzhou, Tesla with Shanghai, Nio with hefei, CATL with tingde, and many many more. TSMC with Taiwan, and there're many other example, but TSMC gets a lot more spotlight.

 

[hist78]

I do not think that number is calculable. It has been 30+ years. I am sure it is significant. Maybe another reason for TSMC to build fabs outside of Taiwan is to get more incentives from the Taiwan Government?

It’s not just subsidies; TSMC receives comprehensive support, including diplomatic assistance. He also suggested substantial subsidies in areas such as green technology, AI, EVs, and more. The fact that TSMC consumes around 10% of Taiwan’s electricity is remarkable. TSMC and Taiwan's government are closely interconnected.

My suggestion is that if we don't know something for sure, we probably shouldn’t just assume it exists.

About three years ago, before the CHIPS Act passed, I listened to a podcast episode produced by a newspaper in a city near a major Intel campus. The reporter stated that semiconductor companies in China, South Korea, Taiwan, and Japan all receive large government subsidies, and that the U.S. government should do the same.

I called the reporter to ask about the amount of subsidy, if any, that TSMC received from the Taiwanese government in recent years, and whether it was significant. Several days later, the reporter called me back and told me he couldn’t find any evidence of a significant subsidy. I thanked him for his effort to answer my question. He told me that, as a reporter, it’s his responsibility to report the facts and that he appreciated my suggestion to correct his report.

I also remember that Pat Gelsinger likes to point out that China, South Korea, Taiwan, and Japan have lower costs for building fabs and that China, South Korea, and Japan provide large subsidies to their fabs. Pat skillfully grouped TSMC/Taiwan among the low-cost and high-subsidy countries. Obviously, TSMC is the major competitor Intel wants to catch up with, but Pat Gelsinger deliberately avoids discussing how TSMC and Taiwan’s semiconductor industry achieved their current success. Instead, he often talks about what the governments of China, South Korea, and Japan do to support their semiconductor industries. I can't recall him ever discussing what Taiwan or TSMC has done that Intel or the U.S. could learn from.

He also doesn’t mention the amount of subsidy TSMC received from the Taiwanese government. I suspect he doesn’t know it either—not even a rough estimate.

Unfortunately, I don't have Pat Gelsinger's phone number to call and ask him, like I did with that reporter.

About two years ago, the Taiwanese government introduced a tax credit incentive program for the tech industry based on R&D spending and advanced equipment procurement. Now, people might mistakenly believe that government subsidy has been with TSMC for the past 30+ years, but this is far from the truth.

A lack of real understanding about how Taiwan developed TSMC to the level it is today could lead a company or a country to terrible results.

 

[XYang2023]

My suggestion is that if we don't know something for sure, we probably shouldn’t just assume it exists.

About three years ago, before the CHIPS Act passed, I listened to a podcast episode produced by a newspaper in a city near a major Intel campus. The reporter stated that semiconductor companies in China, South Korea, Taiwan, and Japan all receive large government subsidies, and that the U.S. government should do the same.

I called the reporter to ask about the amount of subsidy, if any, that TSMC received from the Taiwanese government in recent years, and whether it was significant. Several days later, the reporter called me back and told me he couldn’t find any evidence of a significant subsidy. I thanked him for his effort to answer my question. He told me that, as a reporter, it’s his responsibility to report the facts and that he appreciated my suggestion to correct his report.

I also remember that Pat Gelsinger likes to point out that China, South Korea, Taiwan, and Japan have lower costs for building fabs and that China, South Korea, and Japan provide large subsidies to their fabs. Pat skillfully grouped TSMC/Taiwan among the low-cost and high-subsidy countries. Obviously, TSMC is the major competitor Intel wants to catch up with, but Pat Gelsinger deliberately avoids discussing how TSMC and Taiwan’s semiconductor industry achieved their current success. Instead, he often talks about what the governments of China, South Korea, and Japan do to support their semiconductor industries. I can't recall him ever discussing what Taiwan or TSMC has done that Intel or the U.S. could learn from.

He also doesn’t mention the amount of subsidy TSMC received from the Taiwanese government. I suspect he doesn’t know it either—not even a rough estimate.

Unfortunately, I don't have Pat Gelsinger's phone number to call and ask him, like I did with that reporter.

About two years ago, the Taiwanese government introduced a tax credit incentive program for the tech industry based on R&D spending and advanced equipment procurement. Now, people might mistakenly believe that government subsidy has been with TSMC for the past 30+ years, but this is far from the truth.

A lack of real understanding about how Taiwan developed TSMC to the level it is today could lead a company or a country to terrible results.

This analyst also suggested subsidizing the parts of the semiconductor supply chain that Taiwan does not cover to reduce costs and increase overall competitiveness.

If the U.S. wants to revive its manufacturing base, it needs to adopt strategies similar to those used by Asian countries. Relying solely on the free market won't work. For instance, BYD recently introduced a PHEV pickup in Australia, priced at the entry level of the Ford Ranger while matching the top trim in every aspect. Clearly, BYD is selling these at a loss to reshape the local market.

In semiconductor manufacturing, the U.S. should be competitive due to high levels of automation, especially with advancements in robotics and AI. The U.S. needs to remain committed and stop debating endlessly. While these discussions continue, countries like China and Taiwan are already nearing completion of multiple fabs.

It's no wonder that, as Trump pointed out, other countries sometimes view the U.S. as a joke.

 

[XYang2023]

Intel Ohio Update

Fall colors are at their peak and Ohio is ablaze with golds, reds, purple and orange leaves. Harvest season is my favorite time of year with bonfires, corn mazes and the sounds of leaves crunching underfoot.

We are committed to our Ohio expansion and have been busy the last few months making progress on our new manufacturing facility.

Most visibly, the Liebherr 12500 has joined our two ‘big birds’ on-site, and we were incredibly honored to have the fourth graders from Johnstown Elementary School help us name the crane. The crane will be used to lift the heavy steel roof trusses of the fab into place. It’s a big job for a big crane!

Recently, we reached an important ‘go vertical’ construction milestone, which refers to the building of structures that rise above ground level. A fab has important levels below ground dedicated to utilities and support equipment, though we mostly tend to think of the cleanroom level in semiconductor manufacturing. You can learn more about a fab’s structure here.

We also accepted the delivery of our fourth and final cold box “superload.” All the superloads, some of the largest pieces of equipment for our new facility, made the journey from Manchester in Adams County to the Ohio One campus in New Albany, Licking County. Together with our partners and the community, we are building a brighter future for what is becoming the Silicon Heartland.

The latest from Intel Ohio

Email from Intel Corp. Building the Silicon Heartland View as Webpage Sign up for this newsletter Share This Email Share This Email Share This Email Fall colors are at their peak and Ohio is ablaze wi

myemail.constantcontact.com

 

[delong.height]

My suggestion is that if we don't know something for sure, we probably shouldn’t just assume it exists.

About three years ago, before the CHIPS Act passed, I listened to a podcast episode produced by a newspaper in a city near a major Intel campus. The reporter stated that semiconductor companies in China, South Korea, Taiwan, and Japan all receive large government subsidies, and that the U.S. government should do the same.

I called the reporter to ask about the amount of subsidy, if any, that TSMC received from the Taiwanese government in recent years, and whether it was significant. Several days later, the reporter called me back and told me he couldn’t find any evidence of a significant subsidy. I thanked him for his effort to answer my question. He told me that, as a reporter, it’s his responsibility to report the facts and that he appreciated my suggestion to correct his report.

I also remember that Pat Gelsinger likes to point out that China, South Korea, Taiwan, and Japan have lower costs for building fabs and that China, South Korea, and Japan provide large subsidies to their fabs. Pat skillfully grouped TSMC/Taiwan among the low-cost and high-subsidy countries. Obviously, TSMC is the major competitor Intel wants to catch up with, but Pat Gelsinger deliberately avoids discussing how TSMC and Taiwan’s semiconductor industry achieved their current success. Instead, he often talks about what the governments of China, South Korea, and Japan do to support their semiconductor industries. I can't recall him ever discussing what Taiwan or TSMC has done that Intel or the U.S. could learn from.

He also doesn’t mention the amount of subsidy TSMC received from the Taiwanese government. I suspect he doesn’t know it either—not even a rough estimate.

Unfortunately, I don't have Pat Gelsinger's phone number to call and ask him, like I did with that reporter.

About two years ago, the Taiwanese government introduced a tax credit incentive program for the tech industry based on R&D spending and advanced equipment procurement. Now, people might mistakenly believe that government subsidy has been with TSMC for the past 30+ years, but this is far from the truth.

A lack of real understanding about how Taiwan developed TSMC to the level it is today could lead a company or a country to terrible results.

From the very beginning, it was government fund along with Philips, along with tax incentives that no where can be founded today. Do you think that TSMC can jump from nowhere and suddely be the biggest foundry in the world?

I mean it's still quite expensive even during the early days.

Also like what I said, tsmc consumes 7% of Taiwan's electricity, 7% percent of a developed region isn't a small number. Including power grid that are needed to transmit electricity. The electricity system in Taiwan is government owned vs private entity in the western states. And there's a quite big difference between government owned and private investment. For government owned, it's much cheaper. It's half vs in US. That is what I called incentives. It's free subsidies from the government

Water also another resource, Land also one, but also talent. There are many and many more incentives that is equivalent or maybe greater than giving cash. And it is simply hard to quantify the amount since all of these are not dollar terms. And over 30 years, it's simply a juggernaut amount to nurture a company like TSMC.


Now if you compare that to Intel, does Intel receive any? I know from talent perspective, the best are studying computer science, there is no government involvement but plenty in Taiwan. That is a big incentive.

 

[tooLongInEDA]

From the very beginning, it was government fund along with Philips, along with tax incentives that no where can be founded today. Do you think that TSMC can jump from nowhere and suddely be the biggest foundry in the world?

I mean it's still quite expensive even during the early days.

Also like what I said, tsmc consumes 7% of Taiwan's electricity, 7% percent of a developed region isn't a small number. Including power grid that are needed to transmit electricity. The electricity system in Taiwan is government owned vs private entity in the western states. And there's a quite big difference between government owned and private investment. For government owned, it's much cheaper. It's half vs in US. That is what I called incentives. It's free subsidies from the government

View attachment 2419

Water also another resource, Land also one, but also talent. There are many and many more incentives that is equivalent or maybe greater than giving cash. And it is simply hard to quantify the amount since all of these are not dollar terms. And over 30 years, it's simply a juggernaut amount to nurture a company like TSMC.


Now if you compare that to Intel, does Intel receive any? I know from talent perspective, the best are studying computer science, there is no government involvement but plenty in Taiwan. That is a big incentive.

Slightly off the main topic ... Energy prices aren't substantially higher in Europe or the US due to lack of subsidies or who owns the power generation and distribution. The main factor is regulatory costs, cross-subsidies for other power sources and a crazy pricing regime in many European countries where all energy suppliers get paid the rate of the least highest cost supplier. The chart is useful, but 2 years old - UK energy prices aren't going up 3x this year (home heating oil is 1/3 cheaper than 2 years ago). That's "in the rear view mirror" as we like to say in the chip business.

I note that Iran seems (with Sudan) to have the lowest cost - 1c/kWh ! I recall gas prices there in 2001 - you could fill up a car for under $1.

 

[tedc]

From the very beginning, it was government fund along with Philips, along with tax incentives that no where can be founded today. Do you think that TSMC can jump from nowhere and suddely be the biggest foundry in the world?

I mean it's still quite expensive even during the early days.

Also like what I said, tsmc consumes 7% of Taiwan's electricity, 7% percent of a developed region isn't a small number. Including power grid that are needed to transmit electricity. The electricity system in Taiwan is government owned vs private entity in the western states. And there's a quite big difference between government owned and private investment. For government owned, it's much cheaper. It's half vs in US. That is what I called incentives. It's free subsidies from the government

View attachment 2419

Water also another resource, Land also one, but also talent. There are many and many more incentives that is equivalent or maybe greater than giving cash. And it is simply hard to quantify the amount since all of these are not dollar terms. And over 30 years, it's simply a juggernaut amount to nurture a company like TSMC.


Now if you compare that to Intel, does Intel receive any? I know from talent perspective, the best are studying computer science, there is no government involvement but plenty in Taiwan. That is a big incentive.

You don't think Intel also gets large incentives every time it builds in Oregon, Arizona, Ohio, etc.? You don't think those states are doing everything they can to support Intel's investment (tax breaks, funding, land, electricity, water, people, etc.)? Intel didn't choose to build in Hillsboro just because it's pretty.

Now the US federal government could and should be doing more, but until 2020 it was never a national priority. (Although the government stepped in plenty during the memory wars.)

What the Taiwanese government has done is facilitate a tech ecosystem: industrial parks, research centers, tax breaks, energy, and land/water. None of this was directly aimed at TSMC, though it certainly benefited immensely. Morris Chang has even pointed how little the government supported TSMC financially in the early days (aside from seed money to get it off the ground). They were rightly convinced that TSMC needed to be competitive on its own merits.

 

[hist78]

From the very beginning, it was government fund along with Philips, along with tax incentives that no where can be founded today. Do you think that TSMC can jump from nowhere and suddely be the biggest foundry in the world?

I mean it's still quite expensive even during the early days.

Also like what I said, tsmc consumes 7% of Taiwan's electricity, 7% percent of a developed region isn't a small number. Including power grid that are needed to transmit electricity. The electricity system in Taiwan is government owned vs private entity in the western states. And there's a quite big difference between government owned and private investment. For government owned, it's much cheaper. It's half vs in US. That is what I called incentives. It's free subsidies from the government

View attachment 2419

Water also another resource, Land also one, but also talent. There are many and many more incentives that is equivalent or maybe greater than giving cash. And it is simply hard to quantify the amount since all of these are not dollar terms. And over 30 years, it's simply a juggernaut amount to nurture a company like TSMC.


Now if you compare that to Intel, does Intel receive any? I know from talent perspective, the best are studying computer science, there is no government involvement but plenty in Taiwan. That is a big incentive.

If we examine the graph you posted, we’ll find that it has limited relevance to the topic we’re discussing. In many countries, including the U.S. and Taiwan/ROC, there are at least three different electricity rate groups: residential, commercial (shopping malls, supermarkets, hospitals, etc.), and industrial (factories, food processing plants, fabs, etc.). Depending on factors such as volume, location, time of use, season, provider, type of power generation (solar, wind, coal, nuclear, etc.), business type, regulations, and many other variables, a semiconductor fab may pay a significantly different rate than a residential customer shown in this graph.

We can use the Salt River Project (SRP), which supplies power to Intel fabs in the Chandler, Arizona area, as an example. In 2022 (the most recent available data), SRP’s industrial rate was around $0.0733 per kilowatt-hour (kWh) – less than half of the $0.16 per kWh U.S. electricity rate listed in the graph. Did Intel pay $0.0733 per kWh for its fabs in Chandler? More than likely, it did not. Intel may have received a higher or lower rate than the published SRP industrial rate. In fact, we don’t even know if SRP is Intel’s electricity provider. Due to deregulation, in many states, a public utility company may only provide the distribution network rather than comprehensive electricity or natural gas services.

There are additional factors we need to consider:

In most countries, electricity generation and distribution are heavily regulated and subsidized by governments, including the US, Taiwan/ROC, and other major nations. In the US, for example, the Department of the Interior's Bureau of Reclamation (BOR) owns and operates major dams and power plants on the Colorado River, including the Hoover Dam and Glen Canyon Dam. The Tennessee Valley Authority (TVA), which generates a substantial amount of electricity annually, is also fully owned by the U.S. government.

Federal and local governments frequently fund billions of dollars in power generation, transmission, and distribution projects, benefiting both privately and publicly owned utilities. Tax credits and rebates to utility companies are common across the US.

Utility costs are a critical factor in semiconductor manufacturing. However, we can’t simply assume that TSMC enjoys a substantial, unfair advantage over Intel, Samsung, or GlobalFoundries due to government subsidy, just because TSMC’s business is performing well at the moment.

Have we considered that Taiwan, with a national GDP of $755.1 billion in 2023 (compared to State of Arizona's $522.8 billion in 2023) and only 1/8th the size of Arizona, really doesn't have the resources to outbid many larger countries in government subsidies for semiconductor development?

If government subsidies were so overwhelmingly powerful, then many larger and wealthier countries would have surpassed TSMC and Taiwan in semiconductor industry 10 or 20 years ago.

To reiterate my earlier point:

"A lack of real understanding about how Taiwan developed TSMC to the level it is today could lead a company or a country to terrible results."

This includes Intel as well.

 

[lilo777]

If we examine the graph you posted, we’ll find that it has limited relevance to the topic we’re discussing. In many countries, including the U.S. and Taiwan/ROC, there are at least three different electricity rate groups: residential, commercial (shopping malls, supermarkets, hospitals, etc.), and industrial (factories, food processing plants, fabs, etc.). Depending on factors such as volume, location, time of use, season, provider, type of power generation (solar, wind, coal, nuclear, etc.), business type, regulations, and many other variables, a semiconductor fab may pay a significantly different rate than a residential customer shown in this graph.

We can use the Salt River Project (SRP), which supplies power to Intel fabs in the Chandler, Arizona area, as an example. In 2022 (the most recent available data), SRP’s industrial rate was around $0.0733 per kilowatt-hour (kWh) – less than half of the $0.16 per kWh U.S. electricity rate listed in the graph. Did Intel pay $0.0733 per kWh for its fabs in Chandler? More than likely, it did not. Intel may have received a higher or lower rate than the published SRP industrial rate. In fact, we don’t even know if SRP is Intel’s electricity provider. Due to deregulation, in many states, a public utility company may only provide the distribution network rather than comprehensive electricity or natural gas services.

There are additional factors we need to consider:

In most countries, electricity generation and distribution are heavily regulated and subsidized by governments, including the US, Taiwan/ROC, and other major nations. In the US, for example, the Department of the Interior's Bureau of Reclamation (BOR) owns and operates major dams and power plants on the Colorado River, including the Hoover Dam and Glen Canyon Dam. The Tennessee Valley Authority (TVA), which generates a substantial amount of electricity annually, is also fully owned by the U.S. government.

Federal and local governments frequently fund billions of dollars in power generation, transmission, and distribution projects, benefiting both privately and publicly owned utilities. Tax credits and rebates to utility companies are common across the US.

Utility costs are a critical factor in semiconductor manufacturing. However, we can’t simply assume that TSMC enjoys a substantial, unfair advantage over Intel, Samsung, or GlobalFoundries due to government subsidy, just because TSMC’s business is performing well at the moment.

Have we considered that Taiwan, with a national GDP of $755.1 billion in 2023 (compared to State of Arizona's $522.8 billion in 2023) and only 1/8th the size of Arizona, really doesn't have the resources to outbid many larger countries in government subsidies for semiconductor development?

If government subsidies were so overwhelmingly powerful, then many larger and wealthier countries would have surpassed TSMC and Taiwan in semiconductor industry 10 or 20 years ago.

To reiterate my earlier point:

"A lack of real understanding about how Taiwan developed TSMC to the level it is today could lead a company or a country to terrible results."

This includes Intel as well.

The Federal Government owns 9 power agencies (including 4 Power Marketing Administrations and TVA) with 7% of net generation and 8% of transmission. They are a non-factor. Also, even if Intel gets lower rates for themselves, it does not cover all their needs. They have tons of suppliers and their costs get affected by the country's electricity costs.

Have we considered that Taiwan, with a national GDP of $755.1 billion in 2023 (compared to State of Arizona's $522.8 billion in 2023) and only 1/8th the size of Arizona, really doesn't have the resources to outbid many larger countries in government subsidies for semiconductor development?

Having resources and using them smartly are two very different things. In US, the overarching attitude towards the government is to make it as small as possible and not to let it "interfere" into private business activities. It has good and bad consequences. Also, it's not just about the subsidies. Regulations, prevailing salaries etc. matter a lot. For example, at this point, semiconductor industry in US simply can't compete with other industries (like AI, software design etc.) for talent because it can't offer competitive salaries. If US government is OK with having, say, Facebook instead of semiconductor manufacturing, they don't need to do anything. US will have the best social media in the world. Taiwan can't compete there.

 

[delong.height]

You don't think Intel also gets large incentives every time it builds in Oregon, Arizona, Ohio, etc.? You don't think those states are doing everything they can to support Intel's investment (tax breaks, funding, land, electricity, water, people, etc.)? Intel didn't choose to build in Hillsboro just because it's pretty.

Now the US federal government could and should be doing more, but until 2020 it was never a national priority. (Although the government stepped in plenty during the memory wars.)

What the Taiwanese government has done is facilitate a tech ecosystem: industrial parks, research centers, tax breaks, energy, and land/water. None of this was directly aimed at TSMC, though it certainly benefited immensely. Morris Chang has even pointed how little the government supported TSMC financially in the early days (aside from seed money to get it off the ground). They were rightly convinced that TSMC needed to be competitive on its own merits.

It's still not a national priority today. If it is, the chips act money would've been distributed. The state governments are doing a better job than the federal, I agree. But there's much more needed.

And to be honest,
Let's use Arizona as example, the GDP for Arizona in 2024 is 548B vs 755B of Taiwan. Arizona is 0.7 of Taiwan's GDP. That's certainly not a small number, so it should have the ability to support a TSMC, right?

Arizona is ranked 16 in the GDP list by state. So there are 15 state governments that can support a tech ecosystem for a new TSMC to exist, right? Literally there are 16 states capable of doing what Taiwan has been doing. But why don't we see that before the chips act? because the subsidies were not great, America never enable these deals to exist. And believe it or not, CHIPS act is once in a decade. Will there be a CHIPS act #2. I doubt it unless a COVID-like situation coming up tomorrow and wrecking supply chain once again or any geopolitical instability. There won't be a consensus between government and private sectors. And that's the thing Pat Gelsinger was complaining about. All of sudden, no one was worrying about supply chain stuff. Do you think that it will continue for the state to come up with large incentives it build? I doubt it because what Intel need now is every few year, they need a new fab buildout because how they changed their business model. They are going to milk the deprecated assets forever.

 

[delong.height]

If government subsidies were so overwhelmingly powerful, then many larger and wealthier countries would have surpassed TSMC and Taiwan in semiconductor industry 10 or 20 years ago.

Indeed it is powerful, and why it is needed. Every foundry received preferential treatment from their governments. You can argue that subsidies aren't that important, but it's because you are using TSMC as your example.

It has more than 90% of every leading edge foundry businesses, and a good amount of mature node that is making a pure profits. Not every foundry start out to be the same. They don't have the same runway. TSMC has been on the runway for many decades.

The points I've been trying to make, is that you need a attractive industrial policy in place in order to attract private investment. Giving out subsidy certainly isn't a bad thing. The problem, as it is still happening today, is that the government, or many government don't want to, it's a dirty business. And not something that can reflect on their book (GDP) immediately. Just like how building fab/spending capex is for many companies, it is a long term gain but short term damage.

 

[tedc]

It's still not a national priority today. If it is, the chips act money would've been distributed. The state governments are doing a better job than the federal, I agree. But there's much more needed.

And to be honest,
Let's use Arizona as example, the GDP for Arizona in 2024 is 548B vs 755B of Taiwan. Arizona is 0.7 of Taiwan's GDP. That's certainly not a small number, so it should have the ability to support a TSMC, right?

Arizona is ranked 16 in the GDP list by state. So there are 15 state governments that can support a tech ecosystem for a new TSMC to exist, right? Literally there are 16 states capable of doing what Taiwan has been doing. But why don't we see that before the chips act? because the subsidies were not great, America never enable these deals to exist. And believe it or not, CHIPS act is once in a decade. Will there be a CHIPS act #2. I doubt it unless a COVID-like situation coming up tomorrow and wrecking supply chain once again or any geopolitical instability. There won't be a consensus between government and private sectors. And that's the thing Pat Gelsinger was complaining about. All of sudden, no one was worrying about supply chain stuff. Do you think that it will continue for the state to come up with large incentives it build? I doubt it because what Intel need now is every few year, they need a new fab buildout because how they changed their business model. They are going to milk the deprecated assets forever.

Yes, but also look at it this way: Taiwan's tech ecosystem is primarily supported by the federal government (and somewhat at the city/county level). The US has way more resources than Taiwan has: city (Phoenix), state (Arizona), and federal (US) support.

CHIPS Act distribution is based on milestones. Intel hasn't met them (yet).

I think we can all agree that money alone doesn't solve Intel's problems. Right now, Gelsinger only seems to want money for fab shells. What else would additional funding be used for?

 

[delong.height]

Yes, but also look at it this way: Taiwan's tech ecosystem is primarily supported by the federal government (and somewhat at the city/county level). The US has way more resources than Taiwan has: city (Phoenix), state (Arizona), and federal (US) support.

I do agree that US has way more resource than Taiwan, and why US should be doing well and better than what Taiwan has done, in providing a good environment for business to thrive.

CHIPS Act distribution is based on milestones. Intel hasn't met them (yet).

I think we can all agree that money alone doesn't solve Intel's problems. Right now, Gelsinger only seems to want money for fab shells. What else would additional funding be used for?

If it has been sitting in the house for a year, waiting for Pelosi to add more requirements to that long list. Yes, Intel hasn't met them, none of them ever met.

And why CHIPS Act has been a failure, and not something that Biden administration can speak proudly of.

Believe it or not, since the fab shell has a super long lead time, of course PG want the money for fab shells. You can only manufacture if you have the space, that is the most problematic issue, as I believe.

I do think that US government should pay for the fab shell, lease them to companies like TSMC and Intel. The government should take on the responsibility to unlock that problem.

I certainly don't understand why it is based on milestones, and certainly I think people who want America to reach 20/80 are moron, this includes PG as well.

I think we can all agree that money alone doesn't solve Intel's problems.

And money will solve Intel's problem. The problem that Intel is so problematic is that revolution starts at a bad time. They want the money, but where can they get it. Client, nah, data center, people are buying Nvidia, what else?

 

[hist78]

Indeed it is powerful, and why it is needed. Every foundry received preferential treatment from their governments. You can argue that subsidies aren't that important, but it's because you are using TSMC as your example.

It has more than 90% of every leading edge foundry businesses, and a good amount of mature node that is making a pure profits. Not every foundry start out to be the same. They don't have the same runway. TSMC has been on the runway for many decades.

The points I've been trying to make, is that you need a attractive industrial policy in place in order to attract private investment. Giving out subsidy certainly isn't a bad thing. The problem, as it is still happening today, is that the government, or many government don't want to, it's a dirty business. And not something that can reflect on their book (GDP) immediately. Just like how building fab/spending capex is for many companies, it is a long term gain but short term damage.

To help everyone on this Semiwiki forum—and possibly to help Intel as well—can anyone estimate the subsidies that TSMC and Intel have each received over the past 20 years? I can only guess why Intel, treated TSMC as the largest competitor in the foundry business from Pat Gelsinger's perspective, keeps emphasizing the subsidies received by semiconductor companies in other countries without disclosing the amount TSMC has received. Isn’t that strange?

Without a clear understanding of the true issues and causes, simply injecting more money into Intel will not resolve the crisis it is facing.

 

[siliconbruh999]

To help everyone on this Semiwiki forum—and possibly to help Intel as well—can anyone estimate the subsidies that TSMC and Intel have each received over the past 20 years? I can only guess why Intel, treated TSMC as the largest competitor in the foundry business from Pat Gelsinger's perspective, keeps emphasizing the subsidies received by semiconductor companies in other countries without disclosing the amount TSMC has received. Isn’t that strange?

Without a clear understanding of the true issues and causes, simply injecting more money into Intel will not resolve the crisis it is facing.

Money will help but not completely solve it
it will be upon Intel to solve it