Pages

9 March 2023

The Strategic Importance of Legacy Chips

Sujai Shivakumar , Charles Wessner , and Thomas Howell

The chip shortage in late 2020 drew widespread attention to the fact that the most advanced semiconductors are no longer manufactured in the United States, and that this represents a strategic vulnerability. Interestingly, this chip shortage was overwhelmingly a function of inadequate availability of so-called legacy chips, which U.S.-based firms continue to make but not in sufficient quantities to meet the needs of domestic manufacturers. The shortfall adversely affected not only traditional industries like automobiles but also other devices that incorporate a complement of semiconductor technologies—including microprocessors, advanced microcontrollers, analog and mixed-signal products, and power semiconductors—and drive displays, decode audio, operate engines, and perform other key functions. The absence of these components created significant disruptions in the U.S. economy, prompting a deeper look at the strategic significance of so-called legacy chips.
What Are “Legacy” Chips?

“Legacy” or “mainstream” semiconductor-based integrated circuits (ICs) are made using established—but still evolving—manufacturing processes, typically with larger transistors etched on each chip. The CHIPS and Science Act of 2022 defines legacy devices as those produced with 28-nanometer (nm) technology or larger, and it tasks the Department of Commerce with formulating a precise definition for other types of chips. That determination is still pending. The official definition of “cutting-edge” chips has also yet to be formulated but can be assumed to apply to process nodes at or below 5 nm. The highly advanced 10 nm and 7 nm chips are in a definitional gray area, at least until the Department of Commerce categorizes them. Even then, this is a static distinction in a technology that is still following a trajectory predicted by Moore’s Law: what is considered “legacy” today was cutting-edge not long ago, and what is cutting-edge today is destined to become tomorrow’s legacy chip.

Legacy chips are ubiquitous. While cutting-edge chips or microprocessors have widespread applications in critical technologies, legacy chips are involved in the production of most automobiles, aircraft, home appliances, broadband, consumer electronics, factory automation systems, military systems, and medical devices. These devices have a central role in the U.S. manufacturing economy, meaning disruptions in the availability of legacy chips negatively impact U.S. manufacturing and downstream economic activities.

Despite the name, legacy chips are not stale technology. The connotations associated with terms like “mature,” “older,” and “legacy” are misleading because these categories of chips are constantly being refined for new requirements and applications. Innovations in these chips, for example, include the use of silicon carbide semiconductors, which are expected to have an important role in decarbonizing the economy. Legacy chips are destined to remain highly relevant to emerging industries and technologies far into the future.

These devices have a central role in the U.S. manufacturing economy, meaning disruptions in the availability of legacy chips negatively impact U.S. manufacturing and downstream economic activities.

Therefore, identifying legacy chips as separate from leading-edge chips in terms of transistor size may limit our understanding of their strategic and economic importance. The term “legacy” is itself a holdover from an era when military applications were a driving force in chip innovation. Reflecting this mindset, recent export controls aimed at limiting China’s potential to manufacture advanced chips—leaving China to innovate in mainstream chips—mask a major vulnerability.

If the United States is to protect its economy from the impact of Chinese industrial policy, U.S. strategic thinkers can no longer categorize chips as “advanced” or “less advanced” purely in terms of the size of their components. Instead, policymakers must think more carefully about the importance of specialized legacy chips and policies supporting their production and continued innovation.

Impact of the 2022 Legacy Chip Shortage

Today’s widespread awareness of legacy chips was spurred by a shortage of chips during the Covid-19 pandemic. Despite operating at historic output at that time, the domestic production base was not able to keep pace with domestic demand. In early 2022, the U.S. Department of Commerce released the results of a survey addressing the chip shortage and found that firms faced their most acute shortages not in cutting-edge chips but in legacy chips at the 40 nm node or larger. Most current U.S. chip production—and most global production—consists of these higher-node devices. By the beginning of 2022, according to Department of Commerce survey, U.S. chip plants were operating at 90 percent capacity or higher, “incredibly high for a production process that requires regular maintenance and very high amounts of energy.” Despite the demand, chipmakers were simply incapable of increasing output over the short term.

A 2021 Biden administration review of the U.S. semiconductor supply chain found that “the United States relies on sources concentrated in Taiwan, South Korea, and China to meet demand for various non-leading-edge memory and logic chips that are used widely in myriad consumer and industrial applications.” In 2022, the Department of Commerce concluded—based on a survey of over 150 U.S. companies that produce and consume semiconductors—that a chip shortage was “threatening American factory production and helping to fuel inflation.” Secretary of Commerce Gina Raimondo said on January 25, 2022, that “it’s alarming, really, the situation we’re in as a country, and how urgently we need to move to increase our domestic capacity.”

Considering semiconductors take many months to manufacture—with typical automotive microcontrollers taking six to nine months—current supplies reflect 2022 boosts in production volume, and the market now faces an oversupply for certain kinds of chips. However, the auto industry still faces a shortage of semiconductors.

Supplying the U.S. Automotive Industry

The experience of the U.S. automotive industry in recent months provides important lessons regarding the importance of legacy chips for car and truck production and related economic activity. The U.S. automobile industry relies almost entirely on legacy chips, which account for 95 percent of its total semiconductor consumption. This is because chips used in the automobile industry must achieve “automotive-grade” function: the chip’s ability to endure punishing environments characterized by extreme temperatures, moisture, dust, chemicals, vibration, and electronic interference. This requires far more robust and near “zero-defect” performance than chips found in advanced consumer electronics.

Accordingly, the auto industry has suffered the most from the recent legacy chip shortage. Starting in 2020, shortages and long wait times for automotive-grade legacy semiconductors crippled U.S. auto production. Despite claims that the chip supply problem was due to pandemic disruptions to global supply chains, surging global demand for chips during the pandemic and “just-in-time” optimized automotive supply chains were more to blame. U.S. automakers, expecting a sharp pandemic-induced decline in demand, had canceled their existing contracts for chips. To their surprise, the pandemic brought about an unprecedented swell in the demand for consumer electronics and work-from-home devices, leading to a global shortage for the chips that drive them. As a result, between the beginning of 2021 and mid-2022, North American auto producers were unable to access additional devices and were obliged to cut 4.3 million vehicles out of their production schedules. Prices of available semiconductors also rose substantially, aggravating inflation. The price per unit of common varieties of microcontrollers, for example, quintupled from the fall of 2020 to the fall of 2021.

While this is not always appreciated, today’s automobile industry incorporates semiconductors into the most basic functions of a vehicle. The average number of chips per vehicle doubled between 2017 and 2021 to about 1,700, and this number will only continue to increase as cars and trucks incorporate new safety features and automated driving capabilities. According to one estimate, the total value of the global automotive semiconductor market will grow from $38.7 billion in 2020 to $116.6 billion by 2030, or a compound annual growth rate of 11.7 percent. However, keeping aside cyclical supply issues, the output of semiconductors for automotive applications is simply not keeping pace. Legacy chip production capacity is only expanding at roughly 2 percent annually. For comparison, production capacity for the semiconductor industry as a whole increased by 10 percent per year on average from 2010 to 2020.

Underinvestment and Lost Capacity

For the auto sector, this underinvestment has been particularly acute with respect to analog devices, which process information along gradients rather than just according to on and off signals. STMicroelectronics NV, one of the world’s largest makers of analog chips, forecasts that its backlog of automotive orders will “continue to exceed existing and anticipated manufacturing capacity through 2023.”

More generally, the U.S. chip industry is losing the capability to produce some types of legacy devices altogether. This is not a reflection of technical hurdles but rather of the fact that older fabs are closing down and not being replaced given the associated difficulties. According to a report by the Congressional Research Service, many legacy chips are produced in fabs that are only able to process 200-milimeter (mm) wafers, not the current-generation 300 mm wafers made in newer fabs. To complicate matters, the report mentions that “equipment to make 200 mm chips is no longer readily available.” As a result, firms are hesitant to make new investments in fabs making “old” chips, which can only achieve the required profit margins if produced on depreciated equipment.

This diminution in capacity means that there is no “silver bullet” solution to the chip shortage in the auto sector, not least because the needs of the industry are growing rapidly. Going forward, virtually all the forecasted growth in automotive chip demand will be with respect to devices needed for electrification, autonomous driving, and connectivity—and over two-thirds of the chips required for these applications will consist of legacy nodes through 2030.
The Impact of the Shortage

Supply constraints have disrupted U.S. automakers’ operations as well as affected many other U.S. manufacturing sectors. Medical device producers, for example, suffered from the shortage of mature chips needed to power their machines. The limited availability of chips forced “companies into the spot market to buy electronic components . . . making it difficult for patients to get some critical devices.”

The shortage of mature chips also affected makers of electronic consumer products. In 2021, Apple was forced to curtail its production plans for iPhones and other devices because of shortages. With respect to the iPhone 13 Pro Max, a highly complex device incorporating over 2,000 components, the shortfall did not occur with respect to the most advanced chips. Instead, shortages in peripheral components costing just a few cents—such as power management chips made by Texas Instruments, transceivers from Nexperia, and connectivity devices from Broadcom—held up production. Importantly, “such chips are not unique to the iPhone, to smartphones, or even to consumer electronics, but are used across computers, data centers, home appliances, and connected cars.”

Lagging Investments in Legacy Chips

A major source of chip shortages is simply a lack of investment. For several years, investment in the capacity to manufacture legacy chips has lagged far behind demand. The explanation is straightforward: returns are much lower on investments in the production of legacy chips. At present, only about one-sixth of all semiconductor investment is directed toward legacy chip manufacturing. Despite these risks, several major semiconductor device makers are now pursuing investment strategies aimed at producing higher-node chips, including Infineon, Analog Devices, Texas Instruments, and NXP Semiconductors.

A tension exists between public investments in manufacturing capability for the most advanced chips and the need for additional capacity for more mature devices, forcing policymakers to reconcile conflicting demands for the allocation of public resources. The dilemma has been most evident in the European Union, which has committed €43 billion for advanced chip research and production. In March 2022, Intel announced plans to invest €30 billion, with substantial public support, to build a fab in Magdeburg, Germany, using the most advanced chip manufacturing technology. European manufacturing executives have been critical of these investments in fab construction. They argue that devices made at the 5, 3, and 2 nm process nodes contemplated for the Intel project are best suited for “high-volume, highest performance uses such as smartphones and servers.” According to the executives, these chips “would not match the needs of European industry,” which requires legacy chips for applications in sectors such as automobiles, machinery, and process industries (e.g., chemical or pharmaceutical). One European executive commented that “the narrative that everything will converge to less than five nanometers is a false statement. The main innovation for the auto industry is happening on mature nodes. They need to be very energy-efficient and safe.”

France and Germany have reportedly secured a compromise agreement that will allow older chips to qualify for state aid, as long as they bring an “innovative element” to manufacturing or the final product, which would make Intel’s legacy chip fabs eligible for EU subsidies.

The same tensions were evident in the United States during the 2021–2022 policy debates over legislation that ultimately became the CHIPS and Science Act of 2022. The act became law in August 2022, committing $52 billion in federal funds to promoting investments and research in the domestic semiconductor industry. A coalition of senators secured a cut-out of $2 billion in incentives explicitly meant to subsidize the production of legacy chips. The Biden administration also later clarified that the Department of Commerce would allocate approximately $10 billion of CHIPS funding to expand “new manufacturing capacity for mature and current-generation chips, new and specialty technologies, and for semiconductor industry suppliers.” Such incentives are necessary to incentivize private investment in lower-margin fabs producing higher-node chips, but whether the public funds allocated will prove sufficient to meet the growing need is an open question.

In Japan, virtually all current demand for semiconductors is for the higher-node chips needed by the auto and consumer electronics industries. In 2021, Taiwan’s TSMC and Japan’s Sony Group announced they would invest $7 billion to build a chip plant in Japan to address the need for older-generation chips. The plant is expected to become operational in 2024. Auto parts maker Denso and Sony will take minority stakes. TSMC is reportedly considering building a second fab in Japan to support the transition of the auto industry to electric and autonomous vehicles, which will require access to increasing numbers of cutting-edge chips. Japan is also reportedly working with the United States to establish a fab capable of producing 2 nm devices beginning in fiscal year 2025. Current partners include IBM and Rapidus, a Japanese consortium that includes Sony, NEC, Toyota, Softbank, Kioxia, and MUFG Bank.

The China Risk

Apart from low margins, another major risk factor facing U.S. companies considering legacy chip investments is China. The Chinese government recently announced a $143 billion investment in its chip industry. The Chinese semiconductor industry is expected to have spent $12.3 billion and $15.3 billion on capital expenditure in 2021 and 2022 respectively, accounting for 15 percent of the global total. China is projected to nearly double its installed wafer capacity over the next 10 years to reach approximately 19 percent of worldwide installed chip capacity—assuming it can acquire the necessary manufacturing equipment. Analysis from the Semiconductor Industry Association (SIA) notes that “Chinese EDA firms now have increasingly robust offerings for legacy chips, and domestic Chinese equipment firms are on track to provide strong capabilities for mature node (40/28 nm) production over the next few years.”

Given the Western embargo on advanced chip technologies to China, most of the new investments will likely be in the production of older (28 nm and above) devices. An unintended consequence of U.S. export controls on advanced chip technology to China may be a wave of state-backed investment leading to overproduction and, potentially, Chinese dominance of global legacy chip production.

U.S. experts have already highlighted the risk of this potential Chinese dominance. Former deputy national security adviser Matt Pottinger observed that a Chinese buildup of legacy chip-making capacity would “give Beijing coercive leverage over every country and industry—military or civilian—that depend on 28-nanometer chips, and that’s a big, big chunk of the chip universe.” Dan Hutcheson, another semiconductor expert, observed that “the Chinese could just flood the market with these technologies. Normal companies can’t compete, because they can’t make money at these levels.” Of course, state-owned and state-financed companies benefiting from new equipment do not have to make a profit. They can and likely will focus on market share, cutting foreign companies’ revenues and the resources they need to invest. This will leave authorities in market economies with little choice but to block exports of “dumped” chips or see the industry suffer the fate of the photovoltaic industry, with a loss of sales, revenue, jobs, and innovation.

An unintended consequence of U.S. export controls on advanced chip technology to China may be a wave of state-backed investment leading to overproduction and, potentially, Chinese dominance of global legacy chip production.

Looking ahead, the importance of legacy chips for the operation of the modern economy will only continue to grow. Maintaining a robust and resilient supply base able to make the investments and to produce and improve constantly higher node chips is essential for the nation’s competitiveness and economic security. Moreover, innovation in the higher-node chips is expected to serve as the foundation for a variety of emerging technologies, including those necessary to bring us to a greener, healthier planet. More advanced, higher-node chips will be needed to produce green energy technologies, a domain that neither Europe nor the United States can afford to surrender.

No comments:

Post a Comment