Steven Rattner

What is a semiconductor chip? It’s a small electronic component containing many transistors that can process or store large amounts of data. The transistor was invented by Bell Labs in 1947. The image on the left is a 1956 model; the version on the right, which is depicted roughly to the scale of the one on the left, is a current state-of-the-art chip that powers Apple’s newest laptops.

It contains 57 billion transistors, each measuring just 5 nanometers (There are 25 million nanometers in an inch).

Here’s another example of the miniaturization that has occurred: Intel’s first microprocessor, introduced in 1971, had 2,250 transistors that were each 10,000 nanometers wide. To assemble Apple’s M1 chip using those transistors, you would need 3,300 square feet, a space roughly 50% larger than the median American house.

While chips are made in many countries, the Pacific rim has come to dominate chip manufacturing. In 1980, the United States was the leading manufacturer of chips, with a market share of nearly 60%. (In 1986, Japan overtook the United States as the world’s largest producer.) Today, the U.S. share is down to 13%. That shift occurred for a variety of reasons, but among them was a desire – including of ours – during the Cold War to encourage chip making in Asia to bolster the economies of the countries that surround China.

But our competitive situation is even worse than these figures suggest because none of the most sophisticated chips are made in the U.S.: 92% of them are made in Taiwan by TSMC, the balance by Samsung in South Korea. The most advanced chip made in the U.S. or Europe is 12 nanometers. (For context, every iPhone made since 2017 has used a 10 nanometer or smaller chip.) China has also been far behind Taiwan, with a 14 nanometer chip capacity but there have been reports that China is close to advancing to a 7-nanometer chip. (In October, we banned sales to China of logic chips smaller than 16 nanometers.)

Belatedly, we have come to realize the national security risks of concentrating so much manufacturing of high-end chips on Taiwan. For example, TSMC makes the chips used in an F-35 fighter, and one Javelin missile uses 250 chips (not all of them cutting-edge). At the behest of the Taiwanese government, TSMC has never been willing to make its cutting-edge chips outside of Taiwan. But fortunately, Apple is TSMC’s largest customer and has used its leverage to get TSMC to make its newest 4-nanometer chips in Arizona and eventually, even smaller ones.

The CHIPS Act, pushed heavily for several years by Senators Mark Warner and John Cornyn, provides $52 billion in appropriations and loans and billions more in tax credits to lure chipmakers back to America. In addition to TSMC, Intel and Samsung have both announced plans to build state-of-the-art facilities, in Ohio and Texas respectively. According to consulting studies, the $52 billion would provide for the construction of enough new facilities (known as “fabs”) to meet our critical national security needs, like defense, telecom, and data-center applications.

However, achieving self-sufficiency in chips is unlikely; that would take more than $400 billion in incentives.