An “industrial policy” occurs when the federal government attempts to pick a technology that will be important in the future, and then subsidize it. T.J. Rodgers is an insider in the computer chip industry and has written a credible op-ed analyzing two attempts at industrial policy in his industry, the Sematech consortium in 1987, and the Chips and Science Act in 2022.
He makes a compelling case that Sematech did not work and that the Chips and Science Act is not working either.
Rodgers gives us more evidence that Secretary of the Treasury and Harvard President Larry Summers was right when he wrote “the government is a crappy venture capitalist.”
If the feds’ industrial policy on computer chips has failed, why do we think its industrial policy on cancer cures will succeed? (Nixon and Biden’s cancer “moonshots” amounted to picking, and then subsidizing, what they hoped would be cancer cures.)
By the way, when Rodgers stood up against welfare to his industry, he appears to have been an exemplar of Robert Nozick’s ideal CEO: maximizing profits subject to ethical side-constraints.
(p. A15) My mother was a fifth-grade teacher in Oshkosh, Wis. She earned $25,000 a year. Why should chip companies, some of the wealthiest corporations in the world, take money from her and other ordinary citizens? Today’s massive $280 billion Chips and Science Act of 2022, the latest semiconductor welfare program, is even less justified than Sematech was.
. . .
To my knowledge, Sematech contributed nothing of note to U.S. semiconductor technology. Its Final Report in 1997 served up platitudes about “catching up with Japan” and fostering “industry cooperation.”
Decades later, the Semiconductor Industry Association—now a group of lobbyists in Washington—began “saving” the chip industry again. This time the target is China, despite the fact that its best wafer foundry is SMIC, an also-ran in the foundry business. China is less a competitive threat today than Japan was in 1987.
. . .
Today, 100 high-performance computers can be put on one chip. The companies that know how to design 100-billion-transistor chips for a critical function such as artificial intelligence are much more valuable than the companies that carve commodity chips out of silicon wafers, like those in China. While the biggest chip-manufacturing company in China is worth $54 billion, a single U.S. chip company, Nvidia, is worth $3.4 trillion—58 times as much—and it doesn’t even make its own chips. Why are we doing this again?
For the full commentary see:
(Note: ellipses added.)
(Note: the online version of the commentary has the date June 3, 2025, and has the same title as the print version.)
Robert Nozick discusses ethical side constraints in:
Nozick, Robert. Anarchy, State, and Utopia. New York: Basic Books, Inc., 1974.