(p. A15) Intel, with huge profit margins on its Pentium microprocessors, could spend more than its competitors on state-of-the-art fabs, but innovation eventually was pushed aside for predictability. Intel would get one fab working and then “copy exactly” new cookie-cutter fabs. For smaller feature sizes, Intel looked at the new Extreme Ultraviolet technology from Dutch equipment company ASML and thought it too expensive and risky to use. TSMC embraced ExtremeUV and won, especially for lower-power chips for mobile devices. TSMC can now spend more than anyone else on fabs.
With the Chips+ Act chock full of $52 billion in subsidies and tax credits for chip makers, Congress is saying that real countries have fabs. The act also authorizes $1 billion for carbon removal—weird, because chips are made from silicon. Worse, the U.S. is rewarding Intel, which just announced a disastrous quarter, for coming in third place behind TSMC and Samsung.
Nothing is free. Even Commerce Secretary Gina Raimondo admitted “there’s a lot of strings attached” in the 1,054-page law. National Economic Council director Brian Deese endorsed command-and-control industrial policy: “The question should move from ‘Why should we pursue an industrial strategy?’ to ‘How do we pursue one successfully?’ ” This is as wrong as Soviet or Chinese five-year plans. Industrial policy eventually leads to disaster. Japan’s Ministry of International Trade and Industry micromanaged the country’s domestic semiconductor industry and ended up presiding over its decline. Today no Japanese semiconductor company sits in the global top 10. Because China doesn’t have access to ASML ExtremeUV equipment, it has made little progress in advanced chips.
Yes, we need domestic supplies of advanced chips in case China invades Taiwan. But subsidies are the wrong approach.
. . .
Instead, the U.S. could enable suppliers to place large orders for chips for the military, intelligence agencies, whatever. They could even prepay. Silicon Valley was originally built on orders for transistors for intercontinental missiles and the space program.
For the full commentary, see:
(Note: ellipsis added.)
(Note: the online version of the commentary has the date August 14, 2022, and has the same title as the print version.)