As it gets harder to make computer chips, the U.S. and China battle to lead the way

Silicon Valley, the heartland of the U.S. technology industry, takes its name from the chemical element that is the most important ingredient in microchips.

It was there in the 1950s and 1960s where inventions like the transistor and the integrated circuit were refined, helping to transform computers from unreliable machines the size of a room into dependable devices that fit neatly into pockets. That in turn enabled the technology titans of today to prosper.

If data are the new oil, chips are the internal-combustion engines that turn them into something useful.

The ubiquity of chips has led to the growth of a vast global industry. Modern microchips have billions of components and are made in ultra-advanced factories that cost tens of billions of dollars to build. Indeed, that such devices can be built at all is a testament to the power of specialization and trade.

Chips are caught up in an increasingly bad-tempered rivalry between the U.S. and China. This brewing technological struggle comes at a historic moment. For 50 years progress has been driven by Moore's Law, which states that the number of components that can be crammed onto a chip doubles every two years and thus, roughly, so does its computational power. But the law is breaking down, leaving the future of the industry looking messier and less certain than at any time in the past.

The U.S. has long seen its lead in chipmaking as a vital strategic asset. A White House report published in 2017 does not mince its words: "Cutting-edge semiconductor technology is … critical to defense systems and U.S. military strength."

China also sees chips as critical to its future. In 2014, it established the National Integrated Circuit Industry Investment Fund, to channel cash to research and development in the semiconductor industry. One of the targets of "Made in China 2025," a national program to boost high-tech industries, is to increase domestic production.

This year, the Trump administration briefly banned American firms from selling components to ZTE, a Chinese maker of smartphones and telecoms equipment, after it breached the terms of a settlement having to do with illegally exporting technology to Iran. Bankruptcy was only averted when President Donald Trump unexpectedly agreed to lift the ban.

Just last week, the U.S. alleged that the chief financial officer of China's Huawei, who had been arrested in Canada on Dec. 1, led that company's efforts to circumvent sanctions against selling equipment to Iran.

And chips are a front in Trump's tariff-led trade war with China. They are among a range of goods on which the U.S. raised duties in August.

China has toughened up in retaliation. A planned takeover by Qualcomm, a U.S. firm that designs chips, of NXP, a Dutch one, was abandoned in 2018 after heel-dragging by competition regulators in Beijing. Chinese authorities are also investigating price-fixing among U.S. and South Korean manufacturers of memory chips, including Micron.

China has found it harder to make progress in cutting-edge manufacturing, which is the most demanding part of chipmaking. Chinese upstarts must compete with incumbents that have intimidating technological leads and engineers with decades of hard-won know-how.

The result has been consolidation at the forefront of chipmaking. In 2001, there were 29 companies offering the most advanced fab facilities, according to McKinsey, a consulting firm. Today there are five. That could make it easier for Western techno-hawks to slow China's progress. These fabs are owned by companies based in the U.S., or in Taiwan and South Korea, both of which are its allies.

Another lever might be the firms that supply equipment to those fabs. One in particular — ASML, a Dutch firm — has, after more than a decade of trying, finally commercialized "extreme ultraviolet lithography," a manufacturing process needed for the most advanced chips. That could offer Western policymakers a pinchpoint.

But the end of Moore's Law offers China hope, too. One of its side-effects used to be that, as components shrank, chips were able to operate faster. But that effect — called Dennard scaling — broke down in the mid-2000s, which meant that shrinking a chip's components offered fewer benefits than it used to. For that reason, said Linley Gwennap, a chip-industry analyst, being a pace or two behind the leaders in manufacturing matters a bit less than before.