The US government is doing everything it can to boost domestic semiconductor production, pumping billions of dollars into this limited sector, and putting in place every political mechanism to give it a competitive edge in Asia.
When the pandemic broke out in 2020, companies initially cut back on ordering these essential micro-components needed in smartphones, computers, cars and many other products. As people started working from home, the demand for information and communication technologies – and the chips that power them – grew. There was a shortage of chips and car factories had to stop production because they couldn’t get chips. This contributed to a skyrocketing price of new and used vehicles, one of the main factors in the painful inflation Americans were feeling at the time.
In a statement earlier this year, Commerce Secretary Gina Raimondo called the semiconductor shortage a “national security” issue as it exposed the US manufacturing industry’s reliance on semiconductor imports from abroad. Chips are also used in essential military services and are needed in cybersecurity tools.
The Biden administration and lawmakers mobilized to respond by passing a bill in August called the CHIPS and Science Act. The diploma includes $52 billion to boost semiconductor manufacturing in the United States. Of this amount, $39 billion is earmarked for manufacturing incentives, $13.2 billion for worker research and development and training, and $500 million for international security in information technologies and movements at the semiconductor supply chain level.
In this regard, several major companies have announced significant investments in the US manufacturing industry. Taiwan Semiconductor Manufacturing Company (TSMC), an industry-leading company, has invested at least $12 billion in the construction of a semiconductor plant in Arizona, with production expected to begin in 2024. Earlier this year, Intel said it plans to build a $20 billion semiconductor plant in Ohio, where the launch ceremony for the new chip factory took place last month. And this month, Micron said it will invest up to $100 billion over the next two decades to build a massive semiconductor factory in upstate New York.
In a series of tweets earlier this month, President Joe Biden said: Prometheus: “America will be a leader in microchip manufacturing.”
But the US has a lot to do to catch up. US-based chip factories currently account for just 12% of the world’s modern semiconductor manufacturing capacity, according to data from the Semiconductor Industry Association. About 75% of the world’s modern chip production is now concentrated in East Asia – the majority in geopolitically vulnerable Taiwan. And even with these renewed efforts, the United States does not currently have the same practice and production capacity as some Asian markets that can support a robust domestic industry.
To complicate matters further, the wave of public and private investment comes at a controversial time as concerns over global chip supply shortages have subsided. Pandemic-related blockages in the supply chain are easing somewhat and deteriorating economic prospects have hampered demand.
During an earnings call last week, CC Wei, chairman of TSMC, warned that he expects the “semiconductor industry to likely decline” in 2023. “TSMC is also not immune,” Wei added, but said he expects to be “more resilient than the general semiconductor industry.”
The promotion of semiconductor manufacturing in the United States can now lead to overcapacity and oversupply. And with demand falling, it’s not immediately clear whether government subsidies will be enough to overcome other obstacles the country faces in developing a competitive semiconductor manufacturing hub.
Understanding American Problems
To understand the latest US efforts, it’s important to get a sense of where the country stands — not just in the chip industry in general, but in relation to specific high-value sectors.
“It is highly unlikely that the US will increase its share of global manufacturing because while the US is increasing its manufacturing capacity, TSMC, Intel and others are announcing the opening of factories elsewhere and building them even faster,” said Scott Kennedy, senior consultant. at the Center for Strategic and International Studies.
“But I don’t think this is necessarily a big problem,” he added. He further noted that the production capacity assessment based on gross production, which aggregates low-end chips and high-end chips, is a more realistic and meaningful measure for evaluating chip manufacturing success. “The US needs to increase chip production to the level of a specific type of chip that is directly related to US national security,” he said.
Last Friday, the Biden administration imposed new export restrictions to restrict China’s access to advanced semiconductors made with US equipment, in an effort to manufacture advanced weapons systems.
Although only “about 10% to 14% of chips sold” [globalmente] come from factories in the US,” according to Dan Wang, a professor at Columbia Business School, the US has other strengths. “In terms of design experience, a large part is still in the US”.
Yet the shortcomings are palpable. “In the field of steel mills, which are responsible for the production of semiconductors, the US has not played a leading role for years,” Wang said. While that was the case in the past, production began to migrate to Asia in the 1980s and 1990s, Wang said. “One of the main reasons for this is that labor costs are lower and it is much cheaper to manufacture integrated circuits and chips on a large scale in these parts of the world,” Wang added. Morris Chang, founder of TSMC, said that making chips in the US is 50% more expensive than in Taiwan.
The fact that they already have factories to produce or raise chips gives Asian countries a big advantage. Wang thinks that may be why we see the US “putting so much money into companies that build factories in the United States.” It’s not just to meet demand and become self-sufficient, “but also because you have to get these factories up and running very, very quickly if you want to get into the race.”
What it takes to build a national chip industry
Building new chip factories is an expensive and time-consuming undertaking. “A modern factory is about 100,000 square feet,” said Bob Johnson, an analyst at Gartner, and needs “monstrous cleanrooms with incredible air handling capacity.” He added that these massive buildings need “exceptionally strong foundations.” According to him, “there can’t be any vibration in the factory because it can destroy the production process.” In addition, a single extreme ultraviolet lithography machine needed to map chip circuits costs about $150 million, and Reuters reports that “a modern chip factory needs 9 to 18 of these machines.”
In addition, semiconductor manufacturing requires a variety of specialized raw materials, including pure chemicals such as fluorinated polyimide and inscription gases, chip inscription machines, and more. In places like Taiwan and Fukuoka, Japan, supply chains have developed in areas where suppliers of these products are located close to semiconductor factories. There are also one or two companies that produce important raw materials and have long been trusted suppliers to companies in Asia. This is not yet the case in places like Arizona and Ohio, where there are already plans to build huge chip factories.
It also requires staff who are available and able to do the job.
In the United States, there is a shortage of recent graduates and experienced workers with the technical and engineering skills needed to manufacture semiconductors. Many of those who can have enough experience prefer to work in more popular industries, Kennedy said.
“If we snapped our fingers and came up with 10 new factories with the best chips in the world, we probably wouldn’t have enough people to run them,” Kennedy said. “This is the biggest obstacle to increasing America’s productive capacity, not capital.”
Intel has tried to establish privileged relationships with Arizona State University to recruit engineers, but it’s not clear whether this and other companies that build factories in America are able to hire enough trained engineers and technicians. If they fail, the billions of dollars invested by the private and public sectors may not be enough to recover semiconductor manufacturing.