American Innovation, from smartphones to search engines to gene sequencing, is built on foundations of impossibly complex, perfectly incised silicon. But few of these semiconductors are actually manufactured in the United States. Only 12 percent of chips sold worldwide were made in the U.S. in 2019, up from 37 percent in 1990.
This has not been seen as a problem for decades. American companies have been world leaders in designing high-end chips, the most prized and important part of the process.
That is changing. Supply interruptions caused pandemic and growing technological rivalry with china they urge industry leaders and policymakers to say that the United States actually has to make, not just design, chips.
“It’s a national security risk if we don’t start producing more semiconductors in America,” Gina Raimondo, the U.S. Secretary of Commerce, said Tuesday at an event in Washington, DC.
Speaking of Global Emerging Technology Summit, sponsored by the National Security Commission for Artificial Intelligence, Raimondo said the overall market share tells only part of the story. “Another statistic, which I personally think is more alarming, [is that] zero percent of the leading chips are currently produced in America, ”she told an audience of policy makers and executives.
That could be a big problem. The most complex and powerful computer chips are driving progress in areas such as artificial intelligence i 5G, which in turn will unlock a huge economic value and competitive advantage. “There are no entrepreneurs or large companies here that can do what they do without semiconductors,” Raimondo said.
The Biden administration hinted at its intention to strengthen the domestic chip industry. The CHIPS Act for America, which would fund the $ 52 billion semiconductor industry over five years, was passed as part of the National Defense Authorization Act; the measure to begin the distribution of funds was passed by the Senate and awaits action in the House.
The most advanced computer chips are made using fabrication techniques that operate at the very limit of physics, using extreme engineering feats to make nanometer-sized components (the nanometer is about 100,000th the width of a human hair).
The number of companies producing advanced chips has declined in recent years, and peak production has shifted to the east. An April report by the Semiconductor Industries Association and Boston Consulting Group found that all chips made using the most advanced methods (better known as processes below 10 nanometers) are produced in Asia – 92 percent in Taiwan and the remaining 8 percent in South Korea.
Semiconductor production began migrating from the U.S. in the 1980s, he says Dan Hutcheson, CEO of VLSI Research, an analytical firm, with the advent of electronic design automation that has made it possible to automate most of the tedious work of setting up a circuit design. Among other things, a new kind of so-called companies that produce powerless semiconductors have appeared Qualcomm,, Broadcom, i Nvidia, who designed the chips but did not produce them. At the same time, companies known as foundries that specialized in the production of chips appeared.
“You could design a chip without being a semiconductor engineer,” Hutcheson says. “At the same time, the cost of fabs was getting so expensive that you couldn’t do it if you’re a small company.”
Around 2015, Hutcheson says, after China announced plans to invest large sums of money to improve its own chip production, the U.S. Department of Defense began to worry about what that might mean for America. “That year I went to Washington, DC more than I had in my entire career,” he says.
The US of course still has some big chip companies, especially Intel. But a series of stumbles on advanced manufacturing methods, along with a failure to predict the rise of mobile computing and AI, saw Intel lag behind TSMC rivals in Taiwan and Samsung in South Korea.