October 11, 2022

Why U.S. Efforts to Regain Chip Dominance Are Uphill

The American semiconductor industry ceded the lead to Asian rivals long ago, and now it is scrambling to catch up.

Reported by Larry Light

Are recent U.S. ambitions to recover lost ground in the vital semiconductor sector too little and too late? They may be, according to a host of skeptical strategists.

In the 1950s, two Americans, physicist Robert Noyce and electrical engineer Jack Kilby, invented the microchip—a tiny silicon wafer that today runs much of civilization’s underpinnings, from supercomputers to cell phones to cars. It packs legions of transistors onto something that’s the size of a grain of rice. The U.S. once dominated the chips business globally, in 1990 making 37% of them. Today, that share has fallen to 12%.

Because of a lack of investment domestically, U.S. semiconductor production has slipped, to leave America the world’s fifth largest. Lockdowns early in the pandemic didn’t help, causing a chip shortage. The chip-making action now is centered in Asia. Its decline in the U.S. mirrors that of domestic manufacturing generally.

Taiwan, despite its small size, is the leader, with around 60% of the globe’s output. The island’s Taiwan Semiconductor Manufacturing Company dwarfs anyone else in chip making. After Taiwan come South Korea, Japan and China. The Chinese are making a huge push to improve their ranking and churn out the most advanced chips, an area where they have been lacking.

But now a major effort is underway to restore the U.S.’s prominence in chips. What are the odds of that happening? While many expect the nation to improve its position to a degree, amid some encouraging signs of growth, a lot of doubt exists that the United States can return to anywhere near its erstwhile standing.

Why is the task of bringing back the glory days of U.S. semiconductors such a tall order? Matt Clark, chief investment officer of the South Dakota Investment Council, applauds efforts to improve the U.S. position in this vital industry, but is skeptical how successful the campaign can be. Clark, an astute economic observer, said he was “not sure we are willing to sacrifice enough to produce our own things, since we got spoiled by just printing money and having others make the sacrifices and do the hard work.”

Noting the high cost of doing business in the U.S., Clark observed that “it may be hard to compete without some domestic preferences, which would raise costs to consumers.” What’s more, he added, “we would have to sacrifice to pay more to make [chips] here. I’m not sure we are up to that.”

Throwing Money Around

Regardless, throwing money at the problem is the strategy that Washington has adopted. The Chips and Science Act, passed earlier this year, provides $53 billion for production and research, which comes atop costly plans from U.S. chip companies such as Intel and Micron Technology to expand their reach beyond what the federal subsidies will provide. Intel is building new fabs (chip factories) in Ohio and Arizona. Micron intends to construct the country’s biggest semiconductor plant in upstate New York.

“America invented the semiconductor,” President Joe Biden remarked at a signing ceremony for the measure in August. “And this law brings it back home. It’s in our economic interest and it’s in our national security interest to do so.”

The U.S. strategy is not totally about promoting domestic chip production. It’s also about hindering the competition, in particular China, which seeks to supplant the U.S. as the world’s foremost economic power, in part by propelling tech advances. To hinder Beijing’s plans to forge ahead in semiconductors, the Biden administration is blocking sales of chip-manufacturing equipment and top-tier semiconductors to China.

Still, two large countervailing forces are at work to possibly derail the White House’s objectives: the allure of the massive Chinese market, where many foreign businesses still yearn to be, and President Xi Jinping’s willingness to spare no expense in strengthening its chip capability. So other nations may find a way to sell U.S. chip-making gear to China anyway.

In addition, China has a large and growing cohort of scientists and engineers focused on boosting the nation’s chip endeavor. Recently, Chinese chipmaker Semiconductor Manufacturing International surprised the U.S. by producing 7-nanometer chips, an advanced wafer—and doing so without extreme ultra-violet lithography. This is a process that the U.S. had barred China from getting, a technique many considered crucial to make such small semiconductors.

China’s geopolitical threat to the U.S. has provoked a sotto voce worry that American dependence on TSMC and its brethren is a huge vulnerability to the U.S. economy in general. To wit: What if China invades Taiwan and take over its fabs? That possibility spurs “a desire to diversify away from Taiwan,” noted Chris Shipley, chief investment strategist, North America, at Northern Trust Asset Management.

There’s also a structural weakness that retards a surge in onshore chip making. The U.S. chip industry is not tilted toward manufacturing. AMD, Nvidia, Qualcomm and many other American chip providers design the semiconductors, then ship out the production overseas, often to TSMC. On top of that, American engineers prefer to work in software, not fabs, said Sean Sun, a portfolio manager at Thornburg Investment Management.

Another headwind, at least for the moment, is that the American chip industry could be hindered by the much-anticipated recession lurking ahead. Some customers are cutting back their semiconductor orders, which presents a problem for chipmakers’ finances.

Micron, for instance, has estimated that third-quarter revenue could slide 45% compared with the year-before period. One signal of slackening demand: Memory chip prices are plunging, with DRAM dropping 15% and NAND flash 28% in the just-completed quarter.

At the same time, higher interest rates boost the cost of borrowing to underwrite capital spending. The stock market also is not a propitious place to raise chip capital lately: The Philadelphia Semiconductor Index is off 41% this year, worse than the S&P 500’s downdraft.

Too Little, Too Late?

While sporadic new fabs will pop up in the U.S., the thinking goes, any massive reshoring of capacity runs into the harsh reality of high expenses in America. The cost to set up a new fab in the U.S. is roughly 30% higher than in Taiwan, South Korea or Singapore, and 37% to 50% higher than in China, per a Boston Consulting Group study.

The $53 billion from the Chips Act “won’t be enough to solve our dependence on Taiwan,” said Ken Crawford, a portfolio manager at Argent Capital Management. TSMC and other Taiwanese companies have spent years building out their chip-making prowess, committing financial resources that tower those of U.S. semiconductor providers.

Known in Taiwan as “the sacred mountain, protector of the nation,” TSMC has committed $100 billion over the next three years to ramp up production, far overshadowing Washington’s $53 billion (spread over five years) or the capital spending of any competitor.

One odd feature of the Chips Act that has sparked criticism is that not all the money will go to benefit U.S. companies. Federal grants could be sent to South Korea’s Samsung Electronics and TSMC, for use to build U.S. fabs they plan to erect, in Arizona (cost: $12 billion) and Texas ($17 billion), respectively. To take federal bucks, the foreign firms need to promise they won’t expand their already-existing facilities in China. Further, the Chips Act distributes its largesse widely “to less-efficient geographically dispersed fabs away from R&D centers,” a Credit Suisse report found.

All that said, the U.S. has been counted out in the past—in the 1950s, it lagged behind Russia in the space race—and emerged on top. American ingenuity and entrepreneurial spirit have a storied track record. A big reason for optimism is that America is where a big chunk of the market is, and fabs located here mean fewer supply-chain snarls to inhibit deliveries. The U.S. remains the biggest chip buyer, with half the world’s purchases.

Intel, once the paradigm of inventiveness and the world’s chip kingpin, is spending enormous amounts to scrap its way back. Disastrously, it had spurned smartphones and saw its business erode. Patrick Gelsinger, who took over as CEO in 2021, believes that the U.S. can climb to a 30% market share in chips. He said in a conference appearance in August that he sees Intel developing a chip that holds a trillion transistors, up from 100 billion now.

For certain, wherever they are made, the future is bright for semiconductors, as chips increasingly are used in the Internet of Things, to run everyday activities from lighting systems to medical devices. The industry minted 1.1 trillion chips last year with $590 billion worldwide in revenues, according to Fred Alger Management. That market, the firm forecasted, is poised to hit a projected $940 billion by 2030. With all this turbocharged demand, BCG thinks doubling the U.S. market share is possible.

“The growth area is in the U.S. for chip” demand, said Thornburg’s Sun. “That’s in good shape.”