Semiconductors are essential to economic competitiveness and national security. The innovation of semiconductor technology is the foundation of promoting the digitalization of the global economy, artificial intelligence (AI) and 5G communications. For example, in terms of augmented reality or virtual reality experience, the Internet of Things, Industry 4.0 systems and autonomous vehicles, revolutionary technological applications are gradually becoming a commercial reality.
In addition, national defense also relies on mature electronic systems powered by advanced semiconductor components. The key areas of national defense modernization listed in the “U.S. National Defense Strategy 2018” include microelectronics, 5G, and quantum science. These are strategic areas that require US investment. Other key areas (such as cyber security, artificial intelligence, autonomous systems and advanced imaging equipment) are also extremely dependent on the development of the semiconductor industry. As digitally connected electronic systems become more and more important for managing advanced weapon systems and critical infrastructure, semiconductor suppliers that can provide both economy, reliability and component security will play a more critical role in national security The role of.
Due to the strategic significance of semiconductors to technological leadership and national security, many countries are now paying more attention to their position in the semiconductor value chain. The United States has always been a global leader in the semiconductor field. In the past 30 years, the United States has accounted for 45% to 50% of global semiconductor industry revenue. However, the current US share in the semiconductor manufacturing field is declining, accounting for only 12% of the global installed capacity.
The ongoing geopolitical friction between China and the United States, as well as the destruction caused by the COVID-19 pandemic, have also raised questions about the potential loopholes of US semiconductor companies in the global supply chain, mainly because manufacturing activities are concentrated in East Asia.
In recent years, the United States has initiated many programs related to the US Department of Defense, such as the “Trusted and Guaranteed Microelectronics” initiative, to ensure the safety of the domestic supply value chain at the manufacturing level. Taiwan Semiconductor Manufacturing Company (TSMC), the world’s largest semiconductor foundry company, announced in May 2020 that the company plans to build an advanced wafer fab in Arizona.
This plan is the first step to enhance the advanced semiconductor manufacturing capabilities of the United States. In order to meet the expected growth in semiconductor demand, from 2020 to 2030, global manufacturing capacity will increase significantly, which provides a market for the United States to attract more new fabs. In this report, we will analyze the case of the expansion of the semiconductor manufacturing industry in the United States.
We first discuss the current status and trends of the US semiconductor manufacturing industry to determine how the US share of global manufacturing capacity will change if the status quo continues, and the potential impact of this status and trend on the US semiconductor industry.
In order to understand the root cause of the continuous decline in the share of the United States in global manufacturing over the years, we analyzed the total cost difference between the construction and operation of three types of fabs in the United States and other countries and regions. We also studied the level of government incentives provided by various countries and regions. The analysis results show that compared with China and Taiwan, Singapore and other countries and regions that have important influence in semiconductor manufacturing, the cost of fabs in the United States is 40% to 70% higher than that in the United States. The incentives are directly related.
We have established an analytical model to assess the future trend of the United States’ share of global manufacturing capacity. Although this report does not provide policy recommendations, we propose that if the United States is to set a goal to occupy an important share of the future semiconductor manufacturing capacity and reverse the continued decline in US manufacturing capacity over the past 30 years, what additional additional measures should the US government introduce? Incentive plan.
In the long run, we think these incentive plans are particularly important, because closer R&D cooperation between chip design and manufacturing is needed to develop innovations in chip architecture and materials.
In this way, the performance of next-generation semiconductor chips in the United States can achieve a qualitative leap, and its cost can be greatly reduced. This is of strategic importance, because the technology industry and advanced defense systems depend on the development of the semiconductor industry.
The current position of the United States in the field of semiconductor manufacturing
The United States invented the integrated circuit and has long been a global leader in semiconductors. In the semiconductor industry, US companies always account for 45% to 50% of total global sales. The strong position of the United States in the entire value chain contributed to this position. American companies have more than 50% of the integrated market share in electronic design automation tools (EDA), core intellectual property (core IP), integrated circuit design and manufacturing equipment.
In contrast, the United States’ share of semiconductor manufacturing capacity (37% in 1990) is now only 12%. (See Exhibit 1) The share of manufacturing capacity in discrete, analog and optoelectronic products in the United States is still high (30%).
In fact, the United States is still a global manufacturing leader in specific areas of semiconductors, such as compound semiconductors, radio frequency and bulk acoustic wave (BAW) filters (although this status is now also challenged by new Asian investments). However, the US’s share of memory (4%) and logic systems (12%) is much lower, and the fastest-growing market segment is expected to drive 90% capacity growth in the next decade.
Chart 1-Compared with other semiconductor value chains and other strategic industries, the US manufacturing industry has a lower share
Source: The semiconductor value chain is derived from Gartner, Semiconductor Industry Association (SIA), VLSI Research, SEMI, and corporate finance market data based on BCG analysis; the United States’ share of manufacturing value added is derived from Oxford Economic Research Institute’s macroeconomic data BCG analysis.
Note: DAO refers to discrete, analog and optoelectronics; EDA refers to electronic design automation tools; ICT refers to information and communication technology; OSAT refers to outsourced semiconductor assembly and testing.
The decline in the share of US semiconductor manufacturing is in line with the overall trend of manufacturing in other US industries. The United States’ share of global manufacturing value added fell from 25% in the 1990s to 17% in 2018. 2 However, the US currently accounts for only 12% of the semiconductor manufacturing industry, which is far lower than the US share in other strategic industries such as aerospace (the US’s global manufacturing industry accounts for 49%), medical equipment and pharmaceuticals (approximately 25%). ) And petrochemical products (about 20%). Among industries that rely on advanced manufacturing, the United States only has a lower share of labor-intensive industries (consumer electronics accounted for 3%, computer and network hardware accounted for 8%) than its semiconductor manufacturing industry. Unlike several other industries, the US semiconductor industry has not experienced a wave of large-scale restructuring, which is related to the closure of US manufacturing plants and their relocation abroad. In contrast, in the past 30 years, US manufacturing capacity has grown at a cumulative annual growth rate of 7%. However, during the same period, global production capacity increased at an annual rate of 11%. The increase in installed capacity in the United States has been surpassed by Asian countries and regions such as Taiwan, South Korea, and China, because they have been investing heavily to become a manufacturing leader. (See Figure 2.)
Chart 2-Asian countries and regions exceed the growth of installed capacity in the United States. Source: VLSI research forecast; SEMI 2020 second quarter update; BCG analysis.
Note: All values are displayed as 8-inch equivalent; excluding those with a capacity of less than 5 kwpm or less than 8 inches.
Government policy has been a major factor in the strong growth of Asian semiconductors. These countries and regions put their strategic focus on semiconductors and support the development of their domestic manufacturing industries with preferential government funding, tax breaks and other government incentives, thereby making their economies more attractive.
At the same time, the semiconductor industry has witnessed the rise of the “fabless” model. Many US companies have adopted this business model, allowing them to focus on semiconductor design and commercialization, while at the same time relying on foreign manufacturing partners (also known as dedicated or pure “foundries”). These partners can obtain lower costs and more attractive government incentives in other countries. They can also reduce the risk of large-scale capital investment for a wide range of global clients. Dedicated fabs account for 38% of global manufacturing capacity, of which only 7% are located in the United States.
In contrast, the United States has a much higher share (14%) of the capabilities owned by the global integrated equipment manufacturer (IDM), which designs and manufactures its products in its own factories, so it can Both processes are placed in the same place to achieve synergy. The United States’ share of global manufacturing capacity is expected to decline further.
The latest data on the planned fab construction shows that the new capacity that the United States has developed only accounts for 6% of the total plan, and the new capacity that will start operations in the next five years will be located in the United States. This is significantly lower than the current share of the United States in global installed capacity (12%) and the United States’ share of global new capacity from 2010 to 2020 (10%).
We estimate that if no action is taken, by 2030, the United States’ share of manufacturing will be reduced to 10%. In contrast, mainland China plans to increase 40% of the world’s new capacity and may become a global leader in installed semiconductors.
By 2030, its manufacturing capacity will reach 24% of the world’s total production capacity, which is roughly equivalent to the share of China’s equipment manufacturers’ global demand for semiconductors.
Although China may still be one or two generations behind in manufacturing technology before 2030, its huge manufacturing base may accelerate its learning curve and narrow the gap.
In addition, synergy is particularly important in developing new semiconductor manufacturing capabilities in existing industrial clusters. In fact, semiconductor companies see it as one of the most important factors to consider when choosing a new fab location.
3 This synergy produces a self-reinforcing drive. Over time, the share of the US manufacturing industry will further decline, and the share of mainland China and other Asian countries that have established fabs will further expand. Ultimately, if there are no major changes under current conditions, it will become increasingly difficult for the United States to retain any strong manufacturing capabilities in the semiconductor manufacturing field.
Contrary to the United States, semiconductor manufacturing has been China’s top priority for a long time, and China has also been accelerating the realization of its goals in recent years. As shown in the figure below, China’s share of global semiconductor manufacturing capacity is growing rapidly.
Source: China market data from SEMI, IC Insights and VLSI Research; OECD; BCG analysis.
Why is semiconductor manufacturing so important
The manufacturing industry accounts for 45% of the global added value, and about 20% to 25% of the total R&D investment in the global semiconductor industry. Manufacturing is the center of the continuous development of the semiconductor industry. (See Figure 3). In the past fifty years, semiconductor manufacturing through the scaling of process nodes (commonly known as Moore’s Law) has continued to advance, and semiconductor performance and cost reduction have made amazing progress: the number of transistors per wafer has increased by nearly 10 million times. This makes the processor speed up by 100,000 times, and in the case of comparable performance, the annual cost is reduced by more than 45%.
The rapid development of this technological advancement prompted the transition from mainframes in the 1980s to smart phones in the 2010s, which is the driving force for productivity and economic growth.
Chart 3-The progress of semiconductor manufacturing has promoted the remarkable improvement of chip performance and cost. Source: Intel; Singularity.com; Wikipedia; BCG analysis.
Existing studies have shown that in industries where product design and manufacturing processes are closely related, the isolation of manufacturing and R&D will have a negative impact on industry development. In the semiconductor field, the success of the fabless business model requires cooperation between the design company and its foundry partners, but geographical proximity is not a necessary condition.
However, the semiconductor industry is seeking new breakthroughs in chip architecture and materials to maintain the pace of performance improvement and the cost of applying new key technologies such as AI or quantum computing. Progress in these new areas depends on the ever-increasing R&D cooperation between design and manufacturing.
In view of the leading position of the US semiconductor industry in basic science, integrated circuit design and production equipment, enhancing its manufacturing capabilities may lead it to lead the development of these innovative fields and create a new technological paradigm for the future. Maintaining the leading position in the United States provides important strategic advantages for defining the timing, standards, and business models of semiconductor manufacturing, thereby driving the pace of innovation across the entire value chain from manufacturing equipment and tools to design. Maintaining strong domestic manufacturing capabilities is also critical to ensuring a highly resilient supply chain for the US semiconductor industry.
Approximately 75% of the world’s semiconductor production capacity is concentrated in East Asia. Driven by the strong cluster effect, this number is expected to continue to grow. It is estimated that by 2030, mainland China alone will account for 25% of global manufacturing capacity. The Taiwan region of China currently accounts for 47% of the global capacity of the leading and advanced nodes (10 nanometers or less) used in advanced logic system devices, such as high-performance processors that power smart phones or data centers.
In terms of memory, South Korea accounts for about 30% of total semiconductor demand, while South Korea accounts for more than 40% of global capacity. As the COVID-19 crisis has shown, the high concentration of a country or region makes the global supply chain vulnerable to damage from natural disasters, epidemics or geopolitical conflicts.
Considering the strategic importance of the semiconductor industry to the US economy and national security, it is imperative to enhance the resilience of the supply chain through geographic diversification.
Having a larger semiconductor manufacturing space in the US can also bring more benefits to the US economy:
1. Develop local high-tech clusters to create high-quality employment opportunities and economic prosperity. A new standard-scale fab requires 3,000 to 6,000 employees, and the number of personnel depends on the specific product and technology.
This kind of direct job creation usually has a multiplier effect for the local economy. Over time, it can also help attract other companies in the value chain that hope to benefit from the cluster effect, such as closer integration in the semiconductor ecosystem. Collaboration, access to local talent pools, established support infrastructure, etc. The United States already has a number of vibrant semiconductor manufacturing clusters, such as Dallas and Austin (both in Texas), Portland (Oregon) and cities around Phoenix (Arizona).
2. Improve the balance of US merchandise trade. The US has a very large trade surplus in semiconductors, which will exceed US$8 billion in 2019. Having more fabs in the United States can expand this surplus by increasing the export of semiconductor products designed and manufactured in the United States, whether it is for end customers or overseas facilities to export outsourced semiconductor assembly and testing (OSAT) suppliers, thereby passing Packaging and testing finalize the production process.
Achieving the cluster effect in the value chain and enhancing the flexibility of its global supply chain are crucial to the continued competitiveness of the US semiconductor industry. This is not to say that the United States should blindly restore semiconductor manufacturing capabilities to achieve the goal of “self-sufficiency” in a broad sense. The semiconductor industry is essentially global, because countries and regions have obvious comparative advantages for different activities in the entire value chain. This feature enables American and foreign companies to obtain the best capabilities at the lowest economic cost, which can promote the “virtuous circle” of innovation behind technological breakthroughs in the industry. 6 In addition, just as the high concentration of East Asian manufacturing has led to loopholes in the supply chain of US semiconductors, the production capacity in the US to meet domestic market demand will also cause similar situations.
Understand U.S. competitiveness and alternative locations in other countries
The decline in the United States’ share of global semiconductor manufacturing capacity is not a question of lack of technical capabilities. In fact, the United States accounts for 28% of the global capacity of leading nodes and advanced nodes (10 nanometers or below), which is much higher than its 12% share of all process nodes. The American company is a global leader in the research and development of manufacturing process technology in all fields (logic systems, memory and simulation) and fab software, equipment and process control tools. Eight of the world’s top 20 companies involved in semiconductor manufacturing (including IDM and foundries) have been manufacturing in the United States, and together they account for more than 80% of current global production capacity.
Semiconductor manufacturers employ approximately 180,000 workers in the United States and operate wafer fabs in 18 US states.
So why did the company choose to build a fab outside the United States?
Based on relevant experience in the semiconductor industry, discussions with industry leaders, and surveys of US semiconductor companies involved in manufacturing, we determined: The company decides where to establish key standards for new fabs, and how to compare the United States with other alternative locations Compare the relative position obtained. (See Figure 4) Among the five most important factors, the US ranks very favorably: synergy with the existing scale, access to talent, and protection of intellectual property and assets. However, the United States is considered to be far behind other regions in terms of the two other key factors identified (labor costs and government incentives).
Chart 4-Although it scores high among the three fab site selection criteria, the United States has no economic competitiveness in the establishment of fabs. Source: BCG survey of SIA members, question C2: Choose the most fab location What are the important decisive criteria? Note: The chart does not show other factors that respondents did not identify as “important.”
Ranking comparison of the five criteria for fab site selection data source: SIA member survey data, question G1: Please rate the following countries/regions for each key decision factor in your fab (N = 6). The attractiveness index is quantified as 1-5: 5 = highly attractive, 1 = completely unattractive.
Indeed, the United States is currently not a cost-competitive region for semiconductor manufacturing. Because semiconductor fabs require a lot of investment.
In fact, the ratio of capital expenditure to revenue for the entire industry in 2019 exceeded 20%, and the semiconductor industry and power and utilities are the most capital-intensive sectors in the entire economy.
In order to quantify the cost difference between the United States and other regions, we benchmarked the total cost of ownership (TCO)8 of three representative types of fabs over a ten-year period to clarify what will be established from 2020 to 2030 New capacity 9 (See Figure 5.)
Figure 5-Three representative data sources for fabs used to measure the total cost of ownership in different locations: SIA; BCG analysis.
As shown in Figure 6, including land, buildings, and equipment, a standard-capacity advanced semiconductor factory costs approximately US$5 billion (for advanced analog factories) and US$20 billion (for advanced logic systems and memory factories) Capital expenditure.
This is much higher than the estimated cost of the next-generation aircraft carrier ($13 billion) or new nuclear power plants ($4 to 8 billion). In addition to the initial capital expenditure, we also calculated that the annual cash operating expenditure (labor, utilities, etc.) is approximately US$600 million to US$2 billion.
Therefore, without considering government incentives, the total TCO of a new fab will reach US$11 to 15 billion (advanced analog) and US$30 to 40 billion (for advanced logic systems or memory) within ten years. ).
Considering these costs, government-provided incentives are critical to investors who need support, and such incentives have become a regular part of the business case for new fab investments. Government incentives usually reduce upfront capital expenditures on land, buildings, and equipment, but they can also be extended to recurrent operating expenditures, such as labor costs. Overall, depending on the country/region, we estimate that government incentives can offset 15% to 40% of the total TCO of new factories (before incentives are deducted).
Figure 6-Government incentives have a significant impact on fab costs. Source: BCG analysis.
TCO includes capital expenditure (early land, construction and equipment) plus ten years of operating expenditure (labor, utilities, materials, taxes). The average value is the estimated value of the analyzed country or region (the United States, Japan, South Korea, Taiwan, China, Singapore and Germany).
For each type of fab under consideration, we have analyzed the upfront capital expenditures, annual operating costs and government incentives in different countries.
According to our analysis, among all three types of wafer fabs, the total cost of ownership of wafer fabs in the United States is about 25% to 30% higher than equivalent fabs in Taiwan or Singapore. (See Exhibit 7) In addition to structurally lower wages, China also provides very high government incentives, which seems to be even more cost-competitive.
In the United States, the total cost of ownership of fabs is about 50% higher than that in China, not even counting the additional advantages of financing costs provided by China through access to credit and equity below the cost of capital, a recent study by the Organization for Economic Cooperation (OECD) Operations and Development (OECD) shows that this is very important. 10
Figure 7-The total cost of ownership of US factories is 25%-50% higher than other regions. Source: BCG analysis.
1. TCO includes capital expenditure (early land, construction and equipment) plus ten years of operating expenditure (labor, utilities, materials, taxes).
2. Those multinational companies that choose to enter China for technology sharing can enjoy a wider range of incentives, including equipment leaseback with preferential conditions.
The following factors explain the significant gap in TCO:
Government incentives are the main factor. The United States’ incentive policies are at the bottom of the list, far below those of Asian countries and regions that have large semiconductor manufacturing bases. (See Table 8) Depending on the type of fab and the situation in the relevant countries and regions, these incentives account for 40% to 70% of the cost advantage of other countries or regions relative to the United States.
In some cases, incentives give priority to leading national semiconductor manufacturing companies, thus helping to support the domestic semiconductor industry. But in many cases, multinational companies can also enjoy these preferential policies.
In some cases, the United States is tax-competitive because the effective tax rate in the United States is much lower than the nominal corporate tax rate, and in some places, state and local taxes have been significantly reduced. However, the incentives of these state and local governments are far lower than the subsidies and direct cash incentives provided by other governments.
Table 8: Comparison of local government incentive measures. Source: BCG analysis.
Note: Incentives are for operations in the previous ten years. The countries listed in the table also include a 100% reduction in equipment import costs and a 5% cancellation and extension of R&D projects; it is not exhaustive.
1 The best solution based on current incentives and recent agreements.
2 Excluding China.
4 The effective tax rate is considered separately from general incentive measures and is based on current regulations.
The natural disadvantage of factor costs. The difference in TCO between the United States and other regions is about 15% to 40%, which is mainly caused by structural disadvantages in labor and public utility costs. The median wage in the US manufacturing industry is higher than that in other countries. The US labor cost for fab construction and operation is 40% higher than that of Singapore and Taiwan, and twice that of mainland China. The difference in utility costs between the United States and other countries is not significant, but it is still nearly 25% higher than that in mainland China.
Capital expenditures. Capital expenditure accounts for 15% to 20% of the US TCO disadvantage. Approximately half of the capital expenditure of a new plant is for manufacturing equipment provided by a small group of highly specialized global suppliers, so the situation is expected to be similar across regions. Construction costs account for 20% to 40% of capital expenditures, which is quite different.
In addition, some countries have further promoted the development of their own semiconductor industry ecosystem by building supporting infrastructure around the fab, and semiconductor manufacturers need not pay any fees for this. China’s benefits in this area are particularly comprehensive, usually including housing, telecommunications, public utilities, and logistics infrastructure.
Similarly, other Asian countries and regions, such as Taiwan, Singapore, and South Korea, also provide infrastructure support, usually in the form of special economic zones and technology parks. For example, in Taiwan, China, in addition to providing land, electricity and water, the Science Park also provides space for other supply chain companies to integrate them into the larger manufacturing ecosystem. Similarly, the scope of the Korean government’s cooperation is not limited to public utilities and infrastructure, but also includes identifying and providing convenient locations, simplifying or expediting procedures, and deregulating.
Opportunities to change the trajectory of the next ten years
In view of the strategic nature of semiconductors as the promoters of technological progress, the expansion of domestic manufacturing is extremely important for the US semiconductor industry, which is crucial to enhancing the overall economic competitiveness and national security of the US.
It is estimated that in the next ten years, global demand for semiconductors will grow at an average annual rate of 5%. The driving force will be the large-scale application of new technologies, including artificial intelligence, Internet of Things, edge computing, 5G, and electric vehicles and Vietnam. More and more self-driving cars.
It is estimated that by 2030, manufacturing capacity will increase by 56% on the existing basis, that is, approximately 10 million wpm. As of June 2020, about 50% of the new global production capacity from 2020 to 2030 has not been developed or planned. (See Table 9) This “blank zone”, that is, the solvable part of the demand for new capacity, provides the United States with an opportunity to enable the United States to obtain a higher share of new capacity in the future, surpassing those already under development or 6% achieved in the planning stage. Estimated increase in global production capacity from 2020 to 2030 (M wpm) by development status.
Table 9: The strong growth of global production capacity has brought us growth opportunities. Source: BCG analysis based on SEMI, VLSI and Gartner market forecasts; SIA members’ opinions; BCG project experience.
1. “Under development” includes any status between “breaking ground” and “production”, as well as “planning or announcement” (oral or public confirmation of the construction intention).
2. In order to meet the estimated demand in 2030, it is estimated that additional production capacity will be needed, but there is no specific announcement from any company or country.
3. Discrete, analog and optoelectronic devices.
To realize this market opportunity, it is necessary to bring the TCO of the US fab closer to other countries, thereby making the US a more attractive country for semiconductor manufacturing.
Because the United States has obvious advantages in other criteria that are important for the location of fabs—such as synergy with the existing layout and the rest of the ecosystem, access to a pool of skilled talent, and intellectual property protection—it attracts Semiconductor companies may not need the same total cost to build a larger percentage of new capacity in the United States.
In addition, the ever-changing geopolitical background has made the geographic distribution of manufacturing more diverse, making it more attractive to US and foreign semiconductor companies.
According to our analysis, in order to make the economic benefits of the new American fabs more attractive, the gap in government incentives must be narrowed, because government incentives directly lead to a higher TCO of 40% to 70% in the United States.
New government incentives may also help offset the structural disadvantages we see in the United States in terms of construction and operating costs. In order to assess the potential changes in the current US share of global manufacturing capacity, we developed an analytical model that breaks down the total global new capacity by product type and country.
Then, we used our estimates of the economic conditions of different fabs in various countries to create a “performance ranking” based on TCO. Given the U.S. advantage in other key selection criteria for fab site selection, we assume that the TCO of U.S. fabs needs to be reduced from the current level of 25%~30% higher than that of Taiwan, Singapore or South Korea to only 5%~ The 10% level, not the same total cost, can make the United States an attractive location for new fabs.
To achieve this goal, the US government must develop a new incentive plan. We define this new incentive plan as a fund with a fixed total amount (for example, a grant, a tax credit plan, or both) that can be used for new capacity in the United States from 2021 to 2030. We assume that the existing U.S. state and local incentives remain unchanged and apply to any new capacity built in the U.S. (that is, they are set on a “per plant” basis, rather than capped at a given total amount) .
How much new global capacity the United States can attract depends on the size of the new US incentive plan. We simulated the current situation and two possible additional US incentive schemes. Table 10 shows the expected results for each scenario. ·status quo. We assume that under the circumstance that existing incentives remain unchanged, we can see that the United States accounts for 6% of the projects already developed. This is a good indicator of how much new capacity the United States can attract. This will be lower than the 10% share of the United States in global new capacity in the past decade.
Therefore, the US’s share of global manufacturing will further drop from 12% in 2020 to 10% in 2030.
Scenario 1-New government incentive plan of 20 billion USD. According to our model, we expect that the United States will attract 14 new fabs, 5 more than the current ones, accounting for 14% of the new capacity. The United States will become the third largest location for building new capacity, after China and Taiwan. Therefore, by 2030, the United States will be able to maintain its current share of 12% of global production capacity, which is 2% less than the current forecast. ·
Scenario 2-New government incentive plan of US$50 billion. According to our model, such a plan may make the United States the preferred destination for new semiconductor production capacity outside of China. We estimate that the United States will be able to attract a total of 19 fabs, 10 more than it currently is. This accounts for 24% of new capacity entering the market in the next 10 years. This is a significant increase from the 10% from 2010 to 2020 and the current 6%. This will cause the United States to increase its share of global production capacity from 12% in 2020 to 13% to 14% in 2030, which is a significant increase from the current forecast of 10%.
Table 10: The potential impact of the new incentive plan on the status of the US manufacturing industry. Sources: VLSI research; SEMI 2020 second quarter update; BCG analysis.
1 Assumptions apply to the new capacity in the United States in the next 10 years.
2 For ease of comparison, assume that the fab scale is 75,000 wpm, which is consistent with the average fab scale used in the forecast from 2020 to 2030. According to SEMI data, the number of fabs actually built in the United States from 2010 to 2020 is 19 (excluding experimental and very small fabs, the average scale is about 40,000 wpm.
In our model, we assume that the new additional incentives only apply to new capacity built under the status quo (that is, those “capacity that would otherwise not be built in the United States”).
In practice, this means that only certain fabs that use advanced technologies are eligible for the new incentives. The new government incentives will turn the United States into an economically competitive and attractive place for new fabs.
These potential incentives will mark a real turning point and will reverse the historical trend of continuous decline in US share over the past 30 years. The United States will once again become a competitive semiconductor manufacturing base. In the decades after 2030, the United States will be in a favorable position to continue to increase its participation in global capacity expansion.
In addition, the new government incentive plan will promote the expansion of US manufacturing capacity, which will bring significant benefits to US technological competitiveness, supply chain flexibility and national security. Under the US$50 billion incentive plan, the number of wafer fabs built in the United States in the next ten years may jump from only 9 according to the status quo to 19.
These new fabs headquartered in the United States will bring the most advanced manufacturing technology and sufficient production capacity to meet the semiconductor needs of the US defense and aerospace industries. In addition, we estimate that these 19 new fabs can create about 70,000 direct employment opportunities, significantly increase the talent pool of skilled semiconductor manufacturing technicians in the United States, and strengthen the United States’ capabilities in advanced manufacturing technology. Finally, these government incentives and the construction of new US production capacity will not distort the global semiconductor market. Such incentives are non-discriminatory and can be used for new incremental projects proposed by the company.
Plans of this nature are not intended to select winners or guide market outcomes through government ownership of production capacity or manufacturing companies. In addition, the commercial viability of potential new fabs is supported by the fact that the United States already has all the key driving factors: semiconductor manufacturing technology, talent, supporting infrastructure, and booming semiconductors throughout the value chain Ecosystem, and global market access. This minimizes the risk of global overcapacity, which is clearly not in the interests of the US semiconductor industry.
Reversing the decline of the US semiconductor manufacturing industry
The strategic nature of the semiconductor industry in terms of technological leadership and national security has raised questions about the continued decline in the United States’ share of global semiconductor manufacturing capacity in the past 30 years. The continuing geopolitical friction between the United States and China has exacerbated this concern.
Currently, 75% of the world’s semiconductor production capacity is concentrated in East Asia, and China is actively investing and striving to become the world’s largest manufacturing power by 2030. Given that global production capacity is expected to grow strongly from 2020 to 2030 to meet the growth in semiconductor demand, the next decade will provide the United States with a market opportunity to stop falling or even expand its manufacturing share. The United States already has comparative advantages in some key criteria for selecting factory locations, such as synergy with the existing layout and ecosystem, technical talents, and intellectual property protection, but it is not competitive in terms of cost.
Although the purpose of this report is not to provide policy recommendations, our analysis shows that expanding the current limited state-level government incentives and formulating a new federal plan with a target of 20-50 billion US dollars within ten years can make the United States The incentives are consistent with those of Taiwan, South Korea or Singapore, and re-establish the US as the most attractive country for advanced semiconductor manufacturing.
However, the window for reversing historical trends and expanding the territory of the US semiconductor manufacturing industry is rapidly closing.
First of all, 50% of the new capacity needed to meet global demand in the next 10 years is already under development, so it may be out of reach.
In addition, the current manufacturing powerhouses or regions—especially Taiwan and South Korea, but also more and more smaller places such as mainland China and the emerging Singapore and Israel—benefits from an important cluster effect, which Naturally, it is conducive to the construction of new production capacity in existing manufacturing bases, forming a virtuous circle, making it increasingly difficult for countries like the United States to maintain their global share with shrinking territory.
It is worth noting that although the government is necessary to create a level playing field for American fabs in order to expand the United States’ share in the manufacturing industry, there are other important structural factors that promote the prosperity of the semiconductor manufacturing industry. Support can be beneficial. Basic research in materials and manufacturing science is the foundation of innovation. History has proven that government support is effective in this field. Similarly, another important way for the government to encourage domestic manufacturing is to support training to ensure that the United States has a strong talent pool, including production engineers, operators who can use highly sophisticated computer-controlled equipment, and skilled technicians.
Finally, although the US’s position in the manufacturing industry has attracted strong interest from policy makers, the strength of the US semiconductor industry also requires continued commitment to maintaining the US’s leading position in R&D and ensuring access to the global market. A strong American semiconductor industry that is fully integrated into the global technology supply chain is essential for digital transformation and the creation of a new era of artificial intelligence. Like the mobile revolution of the past 10 years, the huge benefits of this breakthrough will benefit consumers and businesses in all countries, not just the United States.