Semiconductors run everything from our smartphones to our cars, and a massive bill just signed by Biden will help innovation, says Tufts author of book “Chip War”
Semiconductors—the digital machinery running everything from our smartphones and computers to cars and refrigerators—are a vital linchpin to the economy. But only 10% are manufactured in the U.S., while the majority of are produced in Taiwan, South Korea, and China.
On August 2, President Joe Biden signed the CHIPS and Science Act, a bipartisan bill to provide $280 billion to support both manufacturing and innovation in semiconductors, among other scientific efforts. CHIPS stands for Creating Helpful Incentives to Produce Semiconductors.
“These tiny little computer chips are the building blocks of the modern economy,” Biden said at the signing ceremony.
Chris Miller, an assistant professor at The Fletcher School, is the author of the forthcoming book Chip War: The Fight for the World’s Most Critical Technology, coming out in October. Tufts Now spoke with him to learn more about the bill, and what its sponsors hope it will accomplish.
Tufts Now: What are semiconductors, and why should we care about them?
Chris Miller: A semiconductor—more formally known as an integrated circuit—is a network or set of transistors, which are little digital switches that are either on or off, the basis for all the computing in our smartphones and computers.
Semiconductors—which are etched into a silicon wafer—were first invented in the late 1950s. The first one had four transistors on it. The main chip in your iPhone now has over 10 billion transistors—that’s just one of multiple chips in your phone.
Making them is the most complicated manufacturing process in human history. The semiconductors in your smartphone have transistors that measure only a handful of nanometers wide—smaller than a coronavirus. We manufacture these things by the billions every day, and they are everywhere, from microwaves to cars.
I’ve read about semiconductor shortages—and how that shortfall can adversely affect every product with a chip inside.
It’s made people realize that we are fundamentally reliant on computing power in everything we do. We hear a lot of talk about big data, but big data is only as big as the semiconductor capacity you have. I think people have begun to realize that the cloud—the digital world—all of this resides on silicon wafers.
What’s in the CHIPS and Science Act, and what is it intended to do?
The key provision in terms of dollars spent is a program designed to provide tax incentives that reduce the cost difference gap between manufacturing semiconductors in the U.S. versus in other countries—above all in Taiwan, South Korea, and China, which are the three biggest producers of semiconductors today.
There are also a number of new programs designed to bolster semiconductor innovation in the United States, looking at different parts of the process from chip design to the manufacturing tools you need to produce a chip, to the final packaging of a chip and putting it in an electronic device.
If we were to look back in 10 or 20 years, we’d probably see that the biggest impact was in programs designed to promote innovation, rather than the financial support for chip fabrication, even though in dollar terms, the biggest chunk of money is going toward fabrication.
“We hear a lot of talk about big data, but big data is only as big as the semiconductor capacity you have. I think people have begun to realize that the cloud—the digital world—all of this resides on silicon wafers.”
Where is the money for innovation going?
What’s really needed is funding for early stage commercialization of ideas produced by basic research. You might say, well, wait a minute, don’t we have a massive ecosystem of venture capital firms and startups to do that?
We do in the software space, where startup costs are quite low. But if you want to build an advanced semiconductor fabrication facility, that’ll cost you $20 billion.
So now there’s a lot of new thinking going into how you take ideas in the semiconductor sphere coming out of basic research and put them on pathways towards commercialization, because those pathways aren’t nearly as developed as they ought to be. And the CHIPS and Science Act will fund those sorts of efforts.
What sort of U.S. companies would benefit, big chip firms like Intel and AMD?
The more innovation-focused provisions are going to be for smaller companies that most people haven’t heard of—or even companies that don’t currently exist—giving them a better shot at having pathways from taking an idea to building a prototype, to finding the first customer.
What was a driving force behind this bill?
One is that there’s a belief that the semiconductor shortage of the past two years happened because semiconductors are largely not manufactured in the U.S. There’s no basis for that belief, but that’s part of the political discussion.
There actually was a shortage, but not because supply constricted—it was because demand grew. In fact, the world produced vastly more semiconductors in 2020 and 2021 than it had produced in previous years.
The share of semiconductor fabrication happening in the U.S. has declined from about a third of chips produced in 1990 to slightly over 10% today. This isn’t the story of low-value production being off offshore. This is high-value, really complicated stuff. U.S. firms today can’t do what the most advanced Taiwanese semiconductor manufacturer can do, so there’s concern that we’re losing not just manufacturing capacity, but also innovation capacity.
What else are concerns?
The biggest producers of these vital components are Taiwan, South Korea, and China, which are not the most geopolitically stable parts of the world. I think people are very worried that if China invaded Taiwan, it would be very damaging to basically every sector of the economy, to have Taiwanese production knocked offline.
So building chip fabrication plants in the U.S. would be an insurance policy against these concerns?
That’s the idea. Intel would be able to build a couple new facilities in the U.S. I think the realistic best case for U.S. manufacturing semiconductors is that Taiwan’s TSMC, which is the most advanced producer of processor chips, continues to invest in a new facility they’re starting to build in Arizona. Samsung, which is the world’s second most advanced producer of processor chips, is investing in Texas.
That doesn’t change everything overnight, but it starts to move the needle in terms of having a bit more fabrication in the U.S., and the entire ecosystem around fabrication—the specialized chemicals, machine tools, and software used in fabrication.