https://worldview.stratfor.com/article/race-boost-semiconductor-manufacturing-global-powers-take-their-marks?id=743c2bc617&e=de175618dc&uuid=d5de7741-28bd-4058-b4c5-4fb30a887271&mc_cid=8faadc3ae7&mc_eid=de175618dcASSESSMENTS
In the Race to Boost Semiconductor Manufacturing, Global Powers Take Their Marks
13 MIN READNov 11, 2021 | 22:43 GMT
Employees make chips at a semiconductor factory in China's eastern Jiangsu province on March 17, 2021.
Employees make chips at a semiconductor factory in China's eastern Jiangsu province on March 17, 2021.
(STR/AFP via Getty Images)
Increased investments following the pandemic-induced semiconductor shortage will lead to more chip factories being built in more parts of the world. But the sector’s rigid nature and boom-and-bust cycles will still result in countries facing similar supply disruptions in the future. China, the European Union, Japan and the United States are all offering incentives to chipmakers amid the global semiconductor shortage in the hopes of insulating themselves from a repeat in the future. Governments seeking to become largely self-sufficient in the semiconductor industry will find that their investment leads to a dead end. Not all government-backed programs will succeed and some risk wasting billions of dollars of public money on projects that fail. Taiwan Semiconductor Manufacturing Company (TMSC), the global contract chipmaker heavyweight, already plans to invest $100 billion over the next three years. And South Korea, home to Samsung, is targeting $450 billion in investment into semiconductors over the next decade.
China: In its 14th Five-Year Plan adopted in March, Beijing singled out the semiconductor industry as one of seven strategic sectors to prioritize investment in. China is already targeting 70% self-sufficiency by 2025.
European Union: The European Union wants to double its share of global chipmaking to 20% by 2030. In September, the European Commission also unveiled the European Chips Act to promote investment into the sector.
Japan: In June, the Japanese government said it plans to treat the semiconductor industry’s growth as a “national project” akin to energy and food supplies, after Japan’s share of global chip making dropped from 50% in 1988 to just 10% in 2019.
United States: The U.S. Congress is currently weighing the CHIPS for America Act that has passed the Senate and includes $52 billion in support for the semiconductor industry.
The chip shortage during the pandemic exposed how critical the semiconductor industry is to the modern economies, which — coupled with growing U.S.-China competition over technology — will drive government support toward the sector long after the current supply disruptions end. In the short term, the semiconductor shortage will continue to ripple across the global economy, Little respite appears to be on the horizon, with U.S. automakers Ford and General Motors both reporting lower profits on Oct. 27 and saying they see the chip shortage for automakers lasting into 2022. On Oct. 21, Intel CEO Pat Gelsinger also said he thinks the overall shortage will last until 2023. At the same time, tech competition between the United States and China will lead to both countries eyeing further investments into the semiconductor sector as Washington seeks to strengthen its own production capabilities while Beijing tries to build a domestic industry.
The United States has viewed China’s rise in the semiconductor industry as a potential disruptor to the current U.S.-led semiconductor and technology industries, even though some parts of the supply chain (like assembly and packaging) are no longer based in the United States. Many China hawks in Washington fear that Chinese advancements in the semiconductor sector could pave the way for China to eventually replace the United States as the global leader in tech innovation for decades to come.
With a structural overhaul of the global semiconductor industry likely unrealistic, most countries will have little choice but to partner with established players like TSMC, Samsung or Intel. The semiconductor industry is very fragmented and requires high levels of specialization to offset the steep costs of investment. This, in turn, often leads to just a handful of players dominating a few segments of the market. Taiwan’s TSMC and United Microelectronics Corporation (UMC), for example, specialize in contract chip making. These companies neither design the chips they produce nor build most of the equipment they use to manufacture semiconductors. The Netherlands’ ASML, on the other hand, holds a monopoly over some of the world’s most advanced lithography machines needed to build high-end chips that companies like TSMC must buy. The high barriers to entry and the tens of billions of dollars of investment needed to enter the semiconductor industry means that the European Union, the United States and Japan are likely to have the most success in partnering with established chip manufacturers. The alternative can be an expansive long-term investment that those governments are probably unwilling to pay for.
China has struggled to advance its semiconductor sector, despite pouring billions of dollars worth of investment into the industry over the past decade. China is targeting 70% self-sufficiency in chip production by 2025, but only reached 16% self-sufficiency in 2020. One of China’s initially most promising semiconductor ventures, Tsinghua Unigroup, also defaulted on a bond payment in 2020 and is now facing bankruptcy unless it receives a bailout from Beijing.
Though China will make gains in older generation semiconductor technologies, the country will continue to fall short of Beijing’s self-sufficiency goal, especially as the West becomes increasingly nationalistic over the industry amid the competition. China’s overall fab plant capacity growth over the next decade is likely to far outpace that of Japan, Europe and the United States. Chinese companies, however, are largely focusing on older generation technology. China’s SMIC is only just starting to manufacture 10nm or fewer chips. The company also still relies on imported chemicals, raw materials and lithography machines that China cannot produce domestically. This means it will take years for SMIC to invest in the same cutting-edge technologies that competitors like Intel are already researching and investing in. U.S. export controls are impeding China’s ability to import higher-end machines, like those produced by ASML, making it unlikely China will be able to domestically develop those anytime soon as well. But while it’ll struggle to compete with TSMC, Intel and Samsung in the high-end chip market, China can reduce its reliance on other chips, which represent the bulk of its domestic semiconductor consumption for household appliances and Internet-of-Things devices, among other products.
China’s most advanced semiconductor equipment manufacturing company is just beginning to produce equipment that can produce 28nm chips, several generations older than current cutting-edge chips in the market.
Japan will never regain the market share in chipmaking it had in the 1980s, but the country’s unique position in the electronics industry will still enable it to increase its semiconductor manufacturing capacity and seize other market opportunities across the supply chain. Although its share of semiconductor manufacturing has declined, Japan remains an important part of the overall semiconductor industry in certain areas, like chemicals. Japan’s domestic semiconductor industry and human capital, along with its ideal location in East Asia and large customer base, also makes it an attractive destination for nearby chipmakers seeking to diversify their production capabilities, like Taiwan’s TMSC.
On Oct. 4, the Taiwanese chipmaker announced that it would build its first fabrication plant in Japan. The $7 billion factory will use TSMC’s older generation 22-nm and 28-nm processes, which commonly make chips for consumer products that Japan’s electronics industry specializes in, like televisions. TMSC recently announced the plant will be built in partnership with Japanese electronics giant Sony, which is investing $500 million in the project.
Tokyo Ohka Kogyo is the world’s largest maker of photoresist materials, which are needed to produce the extreme-ultraviolet lithography machines that manufacture high-end chips. Japanese companies account for about 90% of the global market share of photoresist materials. Japenese companies Shin-Etsu Chemical and SUMCO also produce more than half the world’s silicon wafers.
The United States’ chipmaking capacity will also rise as Intel, TSMC and others increase investments into fabrication plants within its borders. But the breadth of the U.S. tech sector will keep it dependent on global semiconductor supply chains, even as domestic capacity increases. Despite a decline in logic chip manufacturing share over the last few decades, the United States remains an indispensable part of the overall semiconductor industry. U.S. intellectual property is found throughout the industry; the United States is also home to some of the world’s most important fabless semiconductor companies like Qualcomm, which designs chips that companies like TSMC manufacture. Chip design remains one of the more technologically advanced parts of the industry. The United States plays a large role in manufacturing analog and memory chips.
Nonetheless, the growth of TSMC and Samsung in Asia is now fueling bipartisan support in Washington for increased domestic manufacturing of chips. Although the CHIPS Act has not been fully passed by Congress, some level of federal government support is likely in addition to state- and municipal- level support. This has resulted in several companies announcing plans to build new semiconductor fabrication plants in the United States, including TSMC. But even with these efforts, the United States will remain tied to the global industry, as many U.S.-produced semiconductors will still be shipped elsewhere for the final and more labor-intensive packaging, assembly and testing stages of chipmaking. And most of the electronics where those chips will ultimately be used will also be manufactured in other countries. The United States will still need large imports of materials, chip-making equipment and other inputs for chips as well, further increasing its reliance on the global supply chain.
Construction has already begun on a new $12 billion TSMC plant in Arizona, which is scheduled to come online in 2024.
Samsung is planning to build a $17 billion plant for high-end chips in the United States and is currently evaluating sites for the plant.
Intel is also building new chipmaking capacity in the United States, with plans to introduce new CPU chips every year between 2021 and 2025, as part of its so-called “IDM 2.0” strategy to build more semiconductors for companies that design but don't manufacture them (like Qualcomm and, Apple), in addition to designing and manufacturing its own chips.
Europe’s chip ambitions are facing the steepest uphill battle, as the European market offers few customers (i.e. companies that design but don't make chips) and consumers (i.e companies that build smartphones and personal computers) for high-end chips that chipmakers like Samsung and TSMC transact with. Europe is not tied into the broader electronics industry the same way that the United States and Asia are, with only limited industrial customers for both chip design companies and contract chip makers to target. Still, the European Commission wants to target high-end chips because it is the most innovative part of the industry. The commission’s strategy has been criticized for focusing too much on that aspect of the industry instead of parts of the semiconductor industry that make more sense for Europe to focus on, like the automotive sector. Europe’s large industrial manufacturing base provides a large market for legacy chips. No single European country can compete with the type of aid that China or the United States can give the industry, requiring a pan-European approach — something French President Emmanual Macron has backed significantly — to make meaningful progress. Financial support, however, would still need to come from governments. France and Germany have pledged around 9 billion euros (roughly $10 billion) in funding for the sector, although it remains to be seen if commitments survive German coalition talks and French elections next year. But while Europe will have a harder go in boosting domestic semiconductor manufacturing compared with other regions, U.S.-based Intel has given EU leaders some hope, with the company’s CEO announcing in September that it would build a “big, honking chip fab” in Europe.
Intel plans to build two chip fabrication plants at a mega-site in Europe that could hold up to eight factories as a part of its IDM 2.0 strategy. The plants would be among the company’s most advanced. Intel’s new business model will require building new chipmaking plants.
Intel currently has just four chipmaking plants located outside of North America, including two in Israel and one in Ireland. As the company seeks to build more chip fabrication plants, Intel may see greater growth opportunities in Europe compared with Asia, where it has no existing presence and would face many competitors.
The United States, China, Japan and Europe’s different strategic interests in boosting semiconductor manufacturing could increase the number of high-end chip factories located outside of Taiwan and South Korea by 2025. Having more plants in more places, however, will not defuse the threat of another global semiconductor shortage. Over-investment could also still create a bust in the industry. The geographic concentration of semiconductor manufacturing was only a small part of the chip shortage and it is likely that the shortage would have been nearly as bad over the last two years even if the plants were widely distributed. A quick, unanticipated surge in demand for chips occurred due to the pandemic as more consumers wanted to buy laptops, networking devices, electronics and other products to work from home. The semiconductor industry and the long investment horizons to increase capacity means that the industry will never be structured to deal with sudden shifts in demand. Moreover, the automotive industry’s chip shortage has been exacerbated by the intense certification process that the auto sector undergoes, which makes shifting suppliers particularly difficult. When the pandemic began, auto companies canceled many of their orders, anticipating the economic crisis hitting demand for vehicles. But when they went back to place new orders as demand rose, automakers found themselves behind other customers. And because of the certification process, it takes months for auto companies to shift to other suppliers. The supply issues automakers are dealing with would have thus occurred regardless of whether the chips were made in Taiwan or Germany. The booms and busts the chip industry often sees as new capacity comes online will also continue to be a problem. In fact, the United States, China, Europe and Japan’s increased focus on domestic manufacturing capacity — along with that of established producers South Korea and Taiwan — could lead to an even bigger bust, should many of these new plants be built by the mid-2020s. The high-end chip market at the core of many Western strategies may not be entirely immune to this risk either. At some point, saturated demand for new computers, smartphones and other home electronics that use advanced processors may blunt further demand growth since most people won’t replace as many devices as they did at the start of the pandemic. But the bigger risk may be for older mature technologies, where the scale of China’s investment may overwhelm markets if the country no longer needs to import as many legacy chips. And as seen in other industries, China’s self-sufficiency targets can lead to producers being less responsive to market price signals to reduce production or investment.
