Is America’s chipmaking renaissance real?

Is America's chipmaking renaissance real?

The CHIPS Act: Rebuilding America’s Semiconductor Industry

Image

American chip designers lead the global semiconductor industry, accounting for a third of global semiconductor revenues. Their microprocessors power smartphones, data centres, and are increasingly integral to artificial intelligence (AI) models. However, unlike their Asian counterparts, neither American chipmakers nor their contract manufacturers produce cutting-edge chips within the US. Given the strategic importance of chips in modern economies and national security, policymakers in Washington grew concerned. Their solution was the CHIPS Act, a comprehensive $50 billion package of subsidies, tax credits, and other incentives aimed at attracting advanced chip manufacturing back to America. President Joe Biden signed the Act into law on August 9th, 2022.

At first glance, the CHIPS Act seems to be making an impact. Since its proposal in 2020, chipmakers have announced over $200 billion in investments in America. The goal is for American chip factories, also known as fabs, to produce 18% of the world’s leading-edge chips by 2025. Taiwanese manufacturer TSMC is investing $40 billion in two fabs in Arizona, while South Korea’s Samsung is committing $17 billion to a facility in Texas. In addition, Intel, America’s chipmaking champion, plans to spend $40 billion on four fabs in Arizona and Ohio. As the CHIPS Act celebrates its first anniversary, and with the administration preparing to disburse funds, both Democrats and Republicans view it as a bipartisan triumph.

However, premature triumphalism may be misplaced. The leading-edge fabs being established in America face challenges such as longer construction timelines, higher costs, and smaller capacities compared to those in Asia. On average, it takes American manufacturers 900 days to build a new plant due to the complex regulatory environment at the federal, state, and local levels. Construction expenses, accounting for approximately half of the capital spending on a new fab, can be 40% higher in America than in Asia. While some of these costs can be offset by the CHIPS Act’s financial support, annual operating expenses in America remain 30% higher than in Asia, partially due to higher wages for American workers. However, finding workers with semiconductor industry experience has proven to be a challenge. For example, TSMC had to delay the launch of its first Arizona fab by one year because it couldn’t find enough skilled workers.

The smaller scale of planned American projects further complicates the economics of chip manufacturing. Lowering the unit cost of chips depends on economies of scale. In Arizona, TSMC plans to produce 50,000 wafers per month, which is equivalent to two “mega-fabs” as the company describes them. Meanwhile, in Taiwan, TSMC operates four “giga-fabs” that each produce a minimum of 100,000 wafers per month, along with numerous mega-fabs. Morris Chang, TSMC’s founder, has cautioned that chips made in America will be more expensive.

Image

While TSMC’s CEO, C.C. Wei, has indicated that the company will absorb the higher costs, it is important to note that TSMC and Samsung will continue to produce the majority of their chips more affordably in Taiwan and South Korea respectively. Even Intel is investing more in foreign fabs than in domestic ones. If all planned investments materialize, America will only produce enough cutting-edge chips to meet about a third of its domestic demand. Apple, for instance, will still rely on Taiwan for sourcing high-end processors for its iPhones, as will America’s nascent AI-industrial complex.

Additionally, the CHIPS Act may have unintended consequences. Chip companies that accept state aid are prohibited from expanding their manufacturing capacity in China. This rule diminishes the desire for companies like TSMC and Samsung, which have a substantial customer base in China, to invest more in American fabs. Consequently, Chinese chipmakers are investing in producing less sophisticated semiconductors, causing excess supply and downward pressure on prices for these trailing-edge chips. This situation could harm higher-cost Western manufacturers, potentially driving some out of business. Acknowledging this concern, Gina Raimondo, America’s commerce secretary, stated that China’s focus on the trailing edge is a problem that requires attention.

The CHIPS Act’s impact on the semiconductor industry’s notorious boom-and-bust cycle is unpredictable. Typically, chipmakers increase capacity during periods of rising demand. However, the current landscape is different, with a surplus rather than a shortage of chips due to consumers’ temporary saturation with digital devices. TSMC reported a 10% decline in sales in the second quarter, anticipating a similar drop for the entire year of 2023. Intel experienced a 15% revenue decline in the second quarter compared to the previous year, while Samsung attributed falling revenues and profits to chips’ oversupply. Intel’s stock price has also declined significantly since its peak in early 2021.

Chip executives remain optimistic about the long-term prospects of their industry, anticipating a revival in demand. Nevertheless, the adjustment of inventories to reduce oversupply is taking longer than expected. Moreover, once inventories stabilize, the resulting business environment may be less lucrative. Combined, Intel, Samsung, and TSMC have lost nearly half a trillion dollars, equivalent to a third of their market value since early 2021. Assessing the full impact of the CHIPS Act on American economic security will require further observation, and investors are already drawing their own conclusions.