In 1985, the U.S.-Japan Plaza Accord ended Japan’s semiconductor hegemony. Forty years later, South Korea now stands at a similar crossroads.
Recently, the U.S. planned to revoke the Validated End-User (VEU) system for South Korea, requiring Korean companies to apply for a license for every single piece of U.S. equipment they import. What used to be a green channel without case-by-case licensing now requires review for each use. The tightening of U.S. regulations has left South Korean firms feeling suffocated.
Over the years, South Korea’s semiconductor industry has become deeply entwined with the U.S. While Korean companies can use American technology and equipment, they cannot iterate or upgrade independently. On the surface, the U.S. grants Korea prosperity, but behind the scenes, it enforces cruel technological control.
01 Korea vs. Japan: 65 Years of “Competing for Favor”
Initially, South Korea’s ability to develop semiconductors owed a lot to its “lifelong rival” Japan. After the war, the U.S. heavily supported Japan, which became a major technological power globally.
By 1972, Japanese companies like Toshiba and Hitachi improved upon American technology, enabling DRAM (Dynamic Random Access Memory) chips to surpass U.S. yields. With wings hardened, Japan began exploring boundaries.
In 1982, Japan experienced its most severe “betrayal”—the Königsberg Incident. Against the backdrop of U.S.-Soviet competition, Japan secretly sold critical military equipment to the Soviet Union.
At the time, both the USSR and Japan had pressing problems. For the Soviets, submarine propellers were too noisy, allowing the U.S. Navy to track them from 200 nautical miles away. They urgently needed new propellers but lacked domestic manufacturing equipment.
Meanwhile, Japanese companies like Toshiba faced a cliff-like drop in export orders and severe business crises. In desperation, the Soviets proposed buying MBP-110S nine-axis five-linkage CNC machines from Japan at ten times the price to produce these propellers.
Toshiba then secretly supplied four CNC machines to the USSR via Norway’s Konsberg company (Königsberg in Russian), improving the stealth of Soviet submarines tenfold and falsifying export documents to bypass inspection.
This incident caused significant trouble. At the time, Japanese cars had a 22% market share in the U.S., and Japan controlled over 48% of the global semiconductor market, having replaced the U.S. as the world’s largest semiconductor producer. Japan climbed the ladder, but the U.S. nurtured a tiger that could bite back.
At the same time, South Korea had completed its integrated circuit technology accumulation and gradually shifted its industry focus from packaging to wafer manufacturing.
In the 1980s, Korea launched the “Semiconductor Industry Promotion Plan”. Samsung paid $3 million to Micron for 64K DRAM design rights. SK Hynix imported equipment and processes from Texas Instruments. When IBM sold chip production technology to Samsung, Wall Street capital behind it invested in Samsung via preferred shares, becoming 53% of its foreign capital.
With these shifts, Japan’s DRAM chip orders began to decline. Amid fierce competition between Japan and Korea, the U.S. developed 256K DRAM chips in 1982 to consolidate its technological dominance. Japan responded by launching lower-priced, same-tier chips, overwhelming U.S. companies with high cost-performance.
Humiliated, U.S. capital retaliated in the name of national security, implementing semiconductor trade countermeasures against Japan in 1985:
-
Launching the largest anti-dumping investigation on Japanese chips in history.
-
Forcing restrictions on Japan’s chip export share.
-
Using the U.S.-Japan Semiconductor Agreement to compel Japan to raise chip prices by 100%.
Under the influence of the Plaza Accord, Japan adopted aggressive monetary easing to counter the yen appreciation shock, plunging into its “Lost 30 Years.” By 1995, Japan’s global semiconductor share had dropped to less than half of what it was a decade earlier.
Meanwhile, South Korea’s semiconductor industry expanded rapidly. Samsung launched the world’s first 64M DRAM in 1992 and mass-produced 1GB DRAM in 1996, boosting Korea’s global DRAM market share from less than 5% in the 1980s to over 30% by the mid-1990s.
American assistance and Korean loyalty created a short-term positive cycle of cross-border industrial cooperation—but this situation would soon be disrupted again.
02 The Global Game Built by the “Chinese”
South Korea’s semiconductor industry today finds itself in a difficult position, partly due to Western influence, but largely because of global trends.
Consider one day in 1988 when Intel CEO Andy Grove visited TSMC’s Hsinchu factory. Grove had worked with TSMC founder Morris Chang at Texas Instruments. He visited at Chang’s invitation to inspect the factory and prepare for a major chip foundry deal.
At Texas Instruments, Chang recognized that chip manufacturing was technically demanding, no less than design, with astronomical upfront costs. He decided to separate upstream and downstream semiconductor processes and offer foundry services to multinational giants, seeing immense potential.
But he needed a super-factory like Intel’s support. Chang led TSMC through extensive upgrades, raising yields from 50% to 80%. From then on, “pure foundry” TSMC became one of the most unique entities in the global chip industry.
On one hand, it only does foundry work, never selling chips itself—doing two things in life: producing chips and maintaining secrecy. On the other hand, TSMC became the “cost master” of the chip industry. By replacing the traditional IDM model with a foundry model, it saved tech companies money, improved quality checks, and significantly enhanced procurement cost-performance.
Within a few years, TSMC became like a backbone wedged firmly into semiconductor production. Imagine countless clothing brands worldwide outsourcing to a single factory that keeps all secrets while quietly excelling at production—that is TSMC’s role: the “Sweeping Monk” of the chip world.
TSMC drove innovation across the supply chain, taking responsibility for all manufacturing, while U.S. tech companies only needed to handle design and marketing. Qualcomm, NVIDIA, Apple, Tesla, and even frontier AI companies like OpenAI all depend on TSMC’s foundry services.
Under Western oversight, TSMC succeeded by providing new rules to iterate old rules, reshaping semiconductor cost structures and showing industry leaders new possibilities. When Koreans review semiconductor history, the parallels are striking: the target of Western dominance has shifted—from neighbor Japan to Korea itself, and now Korea’s position must be shared with TSMC.
TSMC also bears another strategic threat to the West: the rise of China’s semiconductor industry.
In 2020, the U.S. pressured TSMC to stop producing 7nm Kirin chips for Huawei. By 2025, restrictions escalated, banning TSMC from producing 16nm and below chips for Chinese firms, requiring these orders to go through U.S.-designated packaging facilities. Moreover, the U.S. forced TSMC to build expensive fabs in Arizona to shift the semiconductor supply chain to the U.S.
The goal of these actions is clear. Japan, Korea, and TSMC face the same dilemma: the U.S. can provide mature technology and equipment worldwide, but allies trying to rise to prominence are blocked. “When the birds are gone, the bow is hidden.”
Asian powers have long seen the situation clearly. Sony, Tokyo Electron, Samsung, and SK Hynix all hope for a safe exit if the U.S. turns against them.
Consider a comparison easily overlooked: in 1969, Korea’s semiconductor industry was just starting, with a GDP of only $7.676 billion—less than 5% of Japan’s. Yet this tiny country, 1/98 the size of the U.S., became a key player in global semiconductors. Today, Korea’s GDP has reached $1.86 trillion, a 242-fold increase. How can such a powerful nation be constrained so tightly?
Decades of hard work, yet Korea cannot break the technological blockade imposed by Western powers. What Korea lacks is not technology, but the determination to break historical cycles and achieve independence.
Globally, the semiconductor ecosystem has a few major players, each with advantages:
-
Netherlands: ASML, the only producer of EUV lithography machines, nearly monopolizing high-end chips.
-
Japan: Despite setbacks, remains a major supplier of semiconductor materials and photoresist, with 70% of chip manufacturers relying on Japanese photoresist.
-
China: Although late to start, possesses rare earth resources and can even counter U.S. chip sanctions.
-
Korea: Lacks natural advantages. It is technologically constrained by the U.S., overlaps with TSMC, and is not irreplaceable.
Korea’s main asset is decades of market share: around 14% globally, and over 50% in DRAM and NAND flash. Yet U.S. policy could easily strip this away.
-
Without U.S. technology and equipment, Korean semiconductors cannot survive independently. Applied Materials, Lam Research, and KLA-Tencor provide much of Korea’s technology and equipment. Even Samsung’s Xi’an fab and SK Hynix’s Wuxi fab rely on over 70% U.S. technology.
-
Chinese orders sustain roughly half of Korea’s semiconductor industry, which the U.S. does not welcome. Since China-Korea diplomatic ties in 1992, trade has soared. By 2013, China overtook the U.S. as Korea’s largest semiconductor export market. In 2020, exports to China exceeded 35%, with DRAM reliance at 42%.
Even in the 2024 U.S.-China tech rivalry, Korea maintains exports to China through indirect routes like Vietnam. Technology comes from the U.S., but orders mainly come from China.
Korea’s economy is highly dependent on China too: in 2024, bilateral trade reached $328.08 billion, 21% of Korea’s total foreign trade. China has been Korea’s largest trading partner for 21 consecutive years.
In recent years, Korean firms have tried to navigate U.S. policies against Chinese chips. In October 2023, the U.S. extended equipment supply exemptions for Samsung and SK Hynix’s China fabs, allowing continued production of advanced DRAM with U.S. equipment. The exception: EUV lithography machines, which must be used in Korea’s Icheon fab.
SK Hynix ingeniously flew wafers from its Wuxi fab to Korea for EUV processing, then back to China for packaging—adding transport costs but remaining cheaper than fully localized production and bypassing U.S. scrutiny.
As a critical part of the semiconductor industry, Korea walks a tightrope between East and West, incurring avoidable waste. Yet facing potential VEU revocation, Korean firms are powerless.
Korea’s semiconductor predicament reflects the fate of a technological vassal: relying on foreign countries for prosperity ultimately risks becoming a pawn in great power games. History serves as a warning.
Meanwhile, China’s semiconductor industry is rapidly advancing along another path.
In 2023, the U.S. still aimed to maintain a >5nm lead, yet Huawei achieved nearly 4nm performance at 14nm using the Ascend 910B chip’s 3D stacking. Last year, Yangtze Memory mass-produced 232-layer Xtacking 3.0 NAND chips, giving China pricing power in storage.
Even high-end equipment has seen breakthroughs: Shanghai Microelectronics’ SSX800 lithography machine achieved 28nm domestic production, and 5nm quantum chip equipment has entered Alpha testing.
Technological breakthroughs require high investment and risk. Only by embracing risk, committing resources, and enduring growing pains can one break historical cycles and remain independent.
Under equality, pursue moderate competition and rational cooperation. Don’t seek to dominate the world, but insist on independence. This is the new rule suited to today’s international environment.