Chip War

Chris Miller masterfully told the story of the spectacular rise of the semiconductor industry (the “chip”) and its ever-growing entanglement with geopolitics (the “war”).  It’s a fascinating narrative, filled with ups and downs, twists and turns, heroes and villains, and a cast of victors and losers — well worth reading for its own sake .  It is a must read if you want to understand the current U.S.-China relationship and the slow-moving crisis hanging over the Taiwan Strait.  Semiconductors have become central to the U.S-China relationship, with one side aggressively playing catch-up, and the other striving to maintain its waning lead.  Taiwan has the misfortune to be caught in the middle of this seemingly inevitable epic-clash, not so much because it offers a beacon of hope for “the free world”, as because it houses Taiwan Semiconductor Manufacturing Company (TSMC), the sole fabricator of the world’s most sophisticated chips.

As I read about the legends of semiconductors unfolding in the book, I came to realize my own ignorance about an industry that has profoundly transformed humanity.

I did not know Williman Shockley who, along with two other scientists at Bell Labs, discovered semiconductors and invented transistors. He also started a company called Shockley Semiconductors Laboratory that counted Gordon Moore (yes, that’s the Moore after which Moore’s law is named) and Robert Noyce among its first hires. The pair would rebel against Shockley later and go on to become the giants of the burgeoning industry. They first founded Fairchild Semiconductor that supplied the computing power to land men on the moon in the 1960s, and then Integrated Electronic, or Intel – a household name in today’s tech world.

I had never heard of Texas Instruments (TI) before I read the book.  But among TI’s earlier employees are Jack Kilby, who won a Nobel prize in physics in 2000 for inventing the integrated circuit (集成电路), Jay Lathrop, who created the first photolithograph (光刻) scanner, and Morris Chang, an immigrant from Mainland China and the founder of TSMC.

Nor could I distinguish between memory chips and logic chips, PC chips and smartphone chips, or deep ultraviolet (DUV) lithography and extreme ultraviolet (EUV) lithography.  What has struck me the most, however, is the incredible difficulty to keep up with Moore’s law, which posits that the number of transistors on a microchip doubles approximately every two years. Indeed, the cutting-edge chips have become so complex that TSMC is the only manufacturer in the world capable of fabricating them at scale.  TSMC does this with “ultra-pure silicon wafers and specialized gases from Japan” and machinery that “can etch, deposit, and measure layers of materials a few atoms thick”. Supplied by only five companies, these tools themselves took decades and an astronomical amount of money to develop, and their core technologies are closely guarded trade secrets.  Take the development of the EUV lithography for example. The project was launched in the early 1990s thanks to a $300-million investment from Intel.  However, it wasn’t until nearly 30 years and billions of dollars in spending later that the Dutch manufacturer ASML finally introduced EUV scanners to the market in 2018, at a price of 100 million apiece for an expected lifetime of four years. For a layman like me, it is mind-boggling to read just how the scanner produces enough EUV light needed for fabrication:

The best approach was to shoot a tiny ball of tin measuring thirty-millionths of a meter wide moving through a vacuum at a speed of around two hundred miles per hour. The tin is then struck twice with a laser, the first pulse to warm it up, the second to blast it into a plasma with a temperature around half a million degrees, many times hotter than the surface of the sun. This process of blasting tin is then repeated fifty thousand times per second to produce EUV light in the quantities necessary to fabricate chips.

It does sound like a miracle, as Miller put it, that something this delicate not only works, but “does so reliably enough to produce chips” that can make lots of money.

This sums up the history of chips. What about war?  The book describes three chip wars that took place between the U.S. and her rivals in different eras.

The war with the Soviet Union, fought mostly in the first half of the cold war, was won with relative ease. The USSR treated its semiconductor industry as a weapons program, similar to its treatment of nuclear and space technology.  In hindsight, this strategy was a huge mistake, as the immensely profitable civilian applications of semiconductors turned out to be such a strong driving force for innovations that no level of government spending could hope to rival.  Faced with the lack of progress, the Russians tried to copy the U.S. technology through espionage. Yet, this did not work either.  For one, even the most skilled spies cannot steal all the technical know-how involved in complex production processes. More crucially, the “copycat” mindset inevitably condemned Russians to a perpetual game of catch-up, rather than allowing them to lead the way.

Japan was a much greater threat.  Thanks to favorable technology transfer and trade policies that the U.S. willingly offered in exchange for the Japanese support of America’s global order, Japan’s semiconductor industry evolved from a niche player specializing in consumer electronics in the 1960s and 1970s to a formidable powerhouse in the 1980s. By 1985, Japan had begun to outspend the U.S. in capital investment for semiconductors, and by the end of that decade, it had become the dominant supplier of the world’s Dynamic Random-Access Memory (DRAM) chips (90% market share) and lithography equipment (70%). Trade disputes soon ensued.  The skirmish started with the U.S. accusing Japan of espionage, double-dealing, and dumping.  It escalated to the point where the U.S. openly threatened tariffs, ultimately compelling Japan to impose quotas on its exports of DRAM chips to the U.S. in 1986.  This did not help Silicon Valley recover their lost ground, however.  Eventually, nearly all American companies, including Intel, were driven out of the DRAM and lithography markets.

Carried away by their astonishing success, the Japanese began to dream about, in the words of Sony Founder Akio Morita, overcoming the United States economically and becoming “number one in the world”.  The U.S. was understandably frightened by the pent-up nationalism revealed in The Japan That Can Say No—which Morita co-authored—and the gloomy prospect of relying on a foreign country to maintain the most important edge of her military. In response, the U.S. launched a campaign to curtail Japan’s dominance in chip-making industry.  The core strategy involves mobilizing South Korea (Samsung), Taiwan (TSMC), and to a lesser extent, Mainland China, to erode Japan’s competitive advantages by enabling cut-throat competition against her companies.  It worked like magic.  In 1998 Japan’s share in the DRAM market fell to 20% from a near monopoly less than a decade ago, while South Korea dethroned Japan to become the largest producer of memory chips. Not only did Japanese firms suffer tremendous share loss in the DRAM market, but they also missed the emerging opportunities in the nascent PC market.  In what Miller dubbed as one of the greatest comebacks in industry history, Intel, under Andy Grove’s leadership, reinvented itself as the king of microprocessors for PCs.  For what seemed like an eternity in this fast-paced industry, Intel was literally the icon of the PC industry, as the blue trademark of its processors emerged as the most recognizable feature on most PC sold globally. Indeed, I remember the first PC I ever owned— which my five college roommates and I purchased in 1995 using pooled funds—simply as a 486, because it was powered by Intel’s 486 microprocessor.  According to Miller, that very computer chip was the first ever with over a million transistors!

This brings me to the latest, still on-going chip war with China.  On the surface, the plot of the Chinese edition bears resemblance to that of Japan: the wary incumbent hegemon, spooked by the rapid ascent of an upstart, is compelled into massive counteractions to neutralize the threat, real or imagined.   However, unlike Japan, China has never really overtaken the U.S. in any high-end technology areas of the semiconductor industry.  Not even close.  According to Miller, toward the end of the 2010s, China had less than 1% of the global chip design software tool market, about 2% of core intellectual property related to “the building blocks of transistor patterns”, 4% of silicon wafers, 1% of fabrication machinery, 5% of chip design, and 7% of fabrication concentrated in the non-cutting-edge chips.  If that is the case, has the U.S. overreacted with her heavy-handed sanctions and embargo against China’s tech sector?

Regarding this, Miller’s insights on the crackdown of Huawei were particularly enlightening. He acknowledged that the charges against Huawei, which included theft of intellectual property, ties with the Chinese military, and violation of U.S. sanctions on Iran, were “ultimately a sideshow” – basically a euphemism for made-up excuses.  The real issue was, Miller wrote,

That a company in the People’s Republic of China had marched up the technology ladder… Its annual R&D spending now rivaled American tech giants…, it was the most successful exporter [of all Chinese tech companies], giving it detailed knowledge of foreign markets. It not only produced hardware for cell towers, [but] it also designed cutting-edge smartphone chips. It had become TSMC’s second biggest customer, behind only Apple.

Therefore, the real question was: “Could the United States let a Chinese company like this succeed?” That is a rhetorical question in case you did not catch the drift. But why?

I can think of several reasons.  

First, unlike Japan, China was not a liberal democracy.  Judged by what was going on in the country since the early 2010s, China absolutely has no interest in becoming one anytime soon. To make things worse, under the current leader, China has repeatedly asserted that perhaps her system, rather than America’s, should be the model that the rest of the world admires, envies, and emulates.  Even when Morita lectured Americans about the superiority of the Japan system, it was seen in Washington as a serious provocation – and he wasn’t even talking about authoritarianism.

Second, unlike Japan, China has never pledged allegiance to the America-led world order.  In fact, in the past decade, China has decisively shifted from biding her time as a halfhearted participant of that order to openly flirting with the idea of challenging it, economically, technologically, and if necessary, militarily.

Third, China has increasingly embraced nationalism as a rallying cry for its people to coalesce around the current regime. However, the inherent logic of this political agenda requires the “unification” of the motherland, regardless of the cost. Whether this stance is a concrete national policy or merely a slogan to appease the little pinks on the internet remains to be seen. Yet it does place China on a collision course with Taiwan and the U.S.  When push comes to shove, the U.S. could find herself in a treacherous standoff with what she now regards as a “peer competitor”. The stakes are incredibly high. Retreating from the American commitment to Taiwan’s security would spell the end of the current global order, potentially plunging the world into chaos.   More importantly, losing Taiwan could hand China a golden opportunity to erode the America’s technological supremacy, which has been a cornerstone of her national security since at least World War II.

As of this writing, China has been denied access not only to high-end chip-making technology but also to the high-end chips themselves. Lacking essential tools (e.g., EUV scanners) and raw materials (e.g., pure silicon wafers), China’s semiconductor industry, as well as her tech sector in general,  is likely to fall behind. Indeed, it has already missed out on the latest gold rush in AI, particularly the triumph of large language models, partly because her access to computing power (GPUs) was severely restricted by sanctions.

Could China break this “neck-strangling” (卡脖子) situation entirely through its own initiatives? Before reading this book, I thought there must be a way, if the nation threw its entire weight behind the challenge.  I am much more pessimistic now. If there’s one thing I’ve learned from the book, it’s that the creation of cutting-edge chips can no longer be achieved within the borders of a single country, not even the U.S. Moreover, the pursuit of technological innovations as a nationalistic project may not be a sound strategy for long-term success, as demonstrated by the failure of the USSR.

Could the chip war have been averted had China chosen a different path in the early 2010s?  What impact will the current conflict have on the “China Dream” and the lives of 1.4 billion Chinese? No one knows the answer.  One can only hope that the war remains confined to the realm of chips and continues to be fought by scientists and engineers with their computers, rather than soldiers with guns and missiles.

 

Marco Nie

Wilmette, IL

3/3/2024