sourcegraph
March 29, 2024

In September, chip giant Intel convened officials on a piece of land near Columbus, Ohio, where it pledged to invest at least $20 billion in two new semiconductor factories.

A month later, Micron Technology celebrated a new manufacturing facility near Syracuse, New York, where the chip company expects to spend $20 billion by the end of the decade and could end up spending five times that amount.

In December, TSMC hosted a gala event in Phoenix with plans to triple investment to $40 billion and build a second new factory to make advanced chips.

The pledges are part of a surge in U.S. chipmaking plans over the past 18 months that rivals Cold War-era investments in the space race. The boom has implications for global technology leadership and geopolitics, with the United States aiming to prevent China from becoming an advanced power in the chips that power innovative computing devices such as smartphones and virtual reality goggles.

Today, chips are an essential part of modern life, even surpassing the tech industry’s creations, from military gear and cars to kitchen appliances and toys.

Since the spring of 2020, more than 35 companies across the country have pledged to spend nearly $200 billion on chip-related manufacturing projects, according to the Semiconductor Industry Association, an industry group. The money will go to 16 states, including Texas, Arizona and New York, for 23 new chip factories, nine factory expansions, and investments in companies that provide equipment and materials to the industry.

The move is one aspect of the Biden administration’s industrial policy initiative, which has at least $76 billion pending in grants, tax credits and other subsidies to encourage domestic chip production. Along with broad funding for infrastructure and clean energy, the efforts constitute the largest U.S. investment in manufacturing, arguably since World War II when the U.S. federal government splurged on new ships, pipes and factories that make aluminum and rubber.

“I’ve never seen a tsunami like this,” said Daniel Armbrust, Sematech’s former chief executive. Sematech is a now-defunct chip consortium formed with the Department of Defense in 1987, with funding from member companies.

President Biden is betting a significant part of his economic agenda on stimulating U.S. chip production, but his reasons go beyond economic gain. Many of the cutting-edge chips in the world today are made in Taiwan, over which China has territorial rights. This has raised concerns that, in the event of a conflict, semiconductor supply chains could be disrupted — putting the U.S. at a technological disadvantage.

New U.S. production efforts could correct some of those imbalances — but only to a point, industry executives say.

New chip factories take years to build and may not offer the industry’s most advanced manufacturing technologies when they start operating. Companies can also delay or cancel those projects if the White House doesn’t give them enough subsidies. A severe skills shortage threatens to dent prosperity, as complex factories need far more engineers than students graduating from American colleges and universities.

The vast amount of funding for U.S. chip production “is not going to try or succeed in becoming self-sufficient,” said Chris Miller, an associate professor of international history at Tufts University’s Fletcher School of Law and Diplomacy, who wrote a recent book on the chip industry’s battles.

White House officials have argued that investing in chip manufacturing would sharply reduce the proportion of chips that need to be purchased from abroad, thereby improving U.S. economic security. At TSMC’s event in December, Mr. Biden also highlighted the potential impact on technology companies such as Apple that rely on TSMC for their chipmaking needs. “This could be a game-changer,” as more companies “bring more of their supply chain home,” he said.

Beginning in the late 1950s, American companies led chip production for decades. But the country’s share of global production capacity has gradually slipped from about 37 percent in 1990 to about 12 percent as Asian countries offered incentives to move manufacturing there.

Today, Taiwan accounts for about 22 percent of total chip production and more than 90 percent of the most advanced chips, according to industry analysts and the Semiconductor Industry Association.

The new spending will improve America’s standing. According to a 2020 Boston Consulting Group study commissioned by the Semiconductor Industry Association, $50 billion in government investment could spur business spending, with the U.S. share of global production reaching 14 percent by 2030.

“It really puts us in the game for the first time in decades,” said John Newfer, the association’s president, adding that the estimate may be conservative because Congress approved $76 billion through a piece of legislation called the CHIPS Act subsidy.

Still, such growth is unlikely to eliminate U.S. dependence on Taiwan for the most advanced chips. Such chips are the most powerful because they pack the greatest number of transistors per silicon chip, and they are often seen as a sign of a nation’s technological progress.

Intel has long led the race to shrink the number of transistors on a chip, usually described in nanometers, or billionths of a meter, with smaller numbers indicating the most advanced production technology. Then TSMC has made great strides in recent years.

But at its Phoenix factory, TSMC may not bring in its most advanced manufacturing technology. The company initially announced it would produce 5-nanometer chips at its Phoenix facility, then said last month it would also produce 4-nanometer chips there by 2024 and build a second factory, which will open in 2026 for production 3 nanometer chip. It does not discuss further progress.

In contrast, TSMC’s factory in Taiwan by the end of 2022 Started production of 3nm technology. International Business Strategies CEO Handel Jones said that by 2025, Taiwan factories may begin to supply two nanometer chips to Apple.

TSMC and Apple declined to comment.

It’s unclear whether other chip companies will bring more advanced cutting-edge chip technology to their new bases. Samsung Electronics plans to invest $17 billion in a new factory in Texas, but has not revealed its production technology. Intel is making chips around 7nm, though it says its U.S. factories will be producing 3nm chips by 2024 and more advanced products soon after.

The spending boom will also reduce (but not eliminate) America’s dependence on Asia for other types of chips. Domestic factories produce only about 4 percent of the world’s memory chips — needed for computers, smartphones and other consumer devices — and Micron’s planned investments could eventually boost that share.

But there could still be gaps in various older, simpler chips, with supplies so short over the past two years that U.S. automakers have had to close factories and produce some finished vehicles. TSMC is a major producer of some of those chips, but it is focusing new investments on more profitable advanced chip factories.

“We still have a dependency that hasn’t been impacted in any shape or form,” said Michael Hurlston, chief executive of Synaptics, a Silicon Valley chip designer that relies heavily on TSMC’s old factory in Taiwan.

The chip-making boom is expected to create 40,000 new jobs at the factories and companies that supply those chips, according to the Semiconductor Industry Association. This would add about 277,000 US semiconductor industry employees.

But filling so many tech jobs isn’t easy. Chip factories typically require technicians to run the factory machines, as well as scientists in fields such as electrical and chemical engineering. According to a recent survey of executives, the talent shortage is one of the industry’s toughest challenges.

The CHIPS Act contains funds for workforce development. The Commerce Department, which oversees disbursements from CHIPS Act funds, has also made clear that organizations wishing to receive funding should have programs in place to train and educate employees.

In response, Intel plans to invest $100 million to advance training and research for universities, community colleges and other technology educators. Purdue University, with its new semiconductor lab, has set a goal of producing 1,000 engineers a year and has attracted chipmaker SkyWater Technology to build a $1.8 billion manufacturing plant near its Indiana campus.

The training, however, may only go so far as chip companies compete with other workers-hungry industries.

“We’re going to have to create a semiconductor economy that attracts people to have a lot of other options,” Mitch Daniels, then-Purdue president, said at an event in September.

Because training efforts can take years to bear fruit, industry executives want to make it easier for highly educated foreign workers to obtain visas to work in the United States or to stay in the United States after earning their degrees. Officials in Washington are aware that rhetoric encouraging more immigration could draw political fire.

But Commerce Secretary Gina Raimondo was blunt in a speech at MIT in November.

Attracting the world’s best scientific talent is “a lost edge for the United States,” she said. “And we’re not going to let that happen.”



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *