The world’s largest chipmaker, Taiwan Semiconductor Manufacturing Co., revealed last month that it will be increasing its investment in the U.S. to $40 billion to build a second factory in Phoenix, Arizona, in addition to the initial factory announced in 2020.

President Joe Biden and CEOs of TSMC’s major American clients, including Apple, AMD, NVIDIA and Micron, gathered to hear the announcement of reportedly the largest foreign direct investment in the U.S.

Tim Cook, Apple’s CEO, said at the event that the progress his company made switching from Intel Core and Xeon processors in 2022 to Apple Silicon systems manufactured by TSMC transformed Apple products like Mac computers, iPads and iPhones. 

“And now, thanks to the hard work of so many people, these chips can be proudly stamped ‘Made in America,’” Cook said.

TSMC is the ninth largest company in the world, says W. Paul Chiou, director of MathWorks Lab for FinTech and Quantitative Analytics at Northeastern and associate teaching professor of finance. It manufactures 55% of all contract chip fabrication globally and more than 90% of the most advanced processors.

“That makes OPEC’s 40% share of global petroleum production look unimpressive,” Chiou says.

TSMC’s knowledge and management of chip production processes will advance the U.S. industry, he says. The Arizona facility will employ 4,500 permanent workers, while creating 21,000 construction jobs.

TSMC will be able to enjoy subsidies from the U.S. government, provided in the bipartisan CHIPS and Science Act, signed by Biden in August 2022. The CHIPS Act aims to strengthen American manufacturing, supply chains and research and development with a $52.7 billion investment into semiconductor manufacturing and $24 billion worth of tax credits for chip production.   

Some critics rushed to declare that TSMC’s investment won’t be enough to build a self-sufficient semiconductor industry in the U.S. and only “buys the bragging rights.” Even Morris Cheng, who founded TSMC in 1987, said that America’s attempt to grow its domestic chip production will be “a wasteful, expensive exercise in futility” during a talk at the Brookings Institution in April 2022.

According to Chang, chips manufacturing in the U.S. is 50% more expensive than in Taiwan and America lacks manufacturing talent.

But the U.S. does not need to become 100% self-sufficient, Northeastern experts say, and its relationship with TSMC is driven not only by economic efficiency.

In the early 1980s, the U.S. semiconductor industry was competing only with Japan, says Ahmed Busnaina, William Lincoln Smith chair and university distinguished professor of mechanical and industrial engineering at Northeastern and director of the Center for High-rate Nanomanufacturing. 

“All this technology was invented in the U.S.—transistor, even the memory chips,” Busnaina says. 

The main concern in the industry in the 1980s and early 1990s was the manufacturing cost, including the escalating cost of building a foundry, or a fab, short for “fabrication plant,” the terms used in the semiconductor industry to refer to a production facility.

“In the 1990s, the cost of [building] a fab increased to $1 billion, and everybody panicked,” Busnaina says. “That started the decline of all the fabs globally because companies were concerned about the cost.”

That’s when the idea of having a contract manufacturer emerged, he says, and the TSMC’s business took off with tremendous support from the Taiwanese government that invested hundreds of billions of dollars in semiconductors. 

Many American companies delegated chip manufacturing to the offshore foundries to save money. But the number of leading-edge foundries globally decreased from about 30 in 2001 to five—TSMC, Global Foundries, Intel, Samsung and Micron—by 2011, Busnaina says.

The cost of building a cutting-edge semiconductor manufacturing facility currently amounts to $20 billion to $24 billion, he says. It costs $1 billion per year to run a fab. 

The complicated manufacturing process, conducted in a vacuum and at high temperatures, includes hundreds of steps. It uses huge amounts of water, as much power as 50,000 homes per year and many different chemicals.

Unlike Intel or Samsung, TSMC does not manufacture its own devices but rather makes chips for other companies under a contract. Companies like Apple and AMD develop prototypes of the chips they need and TSMC makes circuits for them, Busnaina says.

In the third quarter of 2022, the U.S. imported 70% of the chips made by TSMC at a total amount of $13.1 billion.

TSMC reportedly will introduce the 3 nanometer and 2 nanometer chips to the mass market in Taiwan this and next year, respectively, while it plans to start manufacturing 3 nm in Arizona in 2026. This raises the question of whether the company will be supplying the latest technology to its U.S. site or will hold it a step behind its Taiwanese facilities.

One of the reasons TSMC decided to invest in a fab with the U.S. government’s encouragement is the pressure that its American clients must have been putting on the company to diversify geographically, says Stephen Flynn, professor of political science and founding director of Global Resilience Institute at Northeastern.

The COVID-19 pandemic disrupted global supply chains and demonstrated the U.S. vulnerability due to a great reliance on foreign imports, Flynn says, especially for the automotive industry.

“Things got disrupted and there was no backup,” Flynn says. “This obviously has implications for the defense industry and our national security.” 

At the same time, the worsening of the U.S.-China relations, China’s aggressive moves to assert more control over Hong Kong and investments in its military fleet and weapons signaled that Taiwan might be in a more vulnerable situation with China potentially trying to reclaim the island in the future, he says. That could not only endanger U.S. imports of chips from Taiwan but also lead to TSMC losing its fabs and China gaining the lead in advanced chip manufacturing.

These factors along with Biden’s campaign promise to stimulate the economy have encouraged his administration and Congress to step up aggressively on trying to address the dependency on foreign chips, Flynn says. 

“It’s a step in the right direction, whether it turns out to be sufficient or not,” he says.

Flynn believes that it is unrealistic that the U.S. will become completely self-sufficient manufacturing semiconductors due to the nature of the modern global economy and the need for specific materials. But it can still tackle the excessive dependency and the need to diversify. 

From a resilience perspective, companies do not need 100% of supply or capability available at all times. There should be a distinction between normal supply and critical and essential need, Flynn says. 

“There’s a threshold that you want to maintain that assures you some level of criticality that you can prioritize, when things are disrupted, because you will get at some point to recovery and get back to a place that looks more like normal,” he says. “Overinvesting in having 100% assurance all the time is not feasible.” 

Besides TSMC, more than 35 other companies have pledged to spend about $200 billion for chips-related manufacturing projects since the spring of 2020. 

Busnaina believes that continued innovation is the only way the U.S. can stay ahead of competitors.

“And innovation is actually happening,” he says.

With the financial support from the Defense Advanced Research Project Agency, National Science Foundation and the industry, Northeastern was able to develop a new technology that allows to manufacture a chip on one machine in one day at a 100 times lower cost and with the carbon footprint reduced by 20 times.

“We can print a whole wafer with nanoscale and microscale in one layer in one minute,” Busnaina says.

He believes that the cost of making advanced electronics will be 100 times lower in 20 years.

But for now, the U.S. needs to invest money to make chips cheaper and affordable.  Although the CHIPS Act provides for large investments in the semiconductor industry, Busnaina says, it is still much less than China, South Korea and Taiwan invest individually.

One undoubtedly crucial advantage that TSMC currently has is its employees.

Since the chip manufacturing process is so complicated, foundries need professionals with doctoral degrees, advanced chemists and physicists, who are able to solve problems at the atomic level.

“That is why workforce development is an essential part of this. We have to not only train enough engineers, but inspire enough kids at the high school level, to actually go into engineering, and then go into electronics,” Busnaina says.

He blames it on manufacturers who continued to ship jobs overseas in the 2000s that America does not have enough manufacturing talent right now. 

“We definitely had the manufacturing talent until about 2010 when TSMC got ahead of us,” he says.

Another important factor for rebuilding a sustainable chips industry, Chiou says, is the development of an extensive cluster of related manufacturers that provide horizontal supplies and services to the semiconductor fabs. 

In accordance with the CHIPS Act, the U.S. government wants to set up nine hubs across the country to propel ideas and innovation in the industry. Massachusetts is currently working on its application to become one of them, Busnaina says. The effort involves about 30 electronics-related companies and universities such as Northeastern, Massachusetts Institute of Technology and the University of Massachusetts.

“People are now accepting that they really have to look for a cheaper way to make electronics. We don’t have to do it the old fashioned way,” Busnaina says.

Alena Kuzub is a Northeastern Global News reporter. Email her at a.kuzub@northeastern.edu. Follow her on Twitter @AlenaKuzub.