Washington has rushed to save computer chipmaking in the United States and has found $280 billion for the project. Even by the standards of modern Washington, that’s a sizable sum. And as is typical of federal practice, the new legislation is about a lot more than manufacturing semiconductors.
Chipmaking will get a little less than a fifth of the total outlay. The rest will go to a wide range of activities favored by Congress. However the money is spent, the U.S. taxpayers will foot the bill.
The balance of the spending will go to a raft of projects, almost all being at least vaguely connected to technology and science and aimed at increasing Washington’s control over research and technological directions.
It isn’t so much favoritism—although that isn’t unheard of in Washington—but rather that these companies do most of their manufacturing domestically, while others, such as Advanced Micro Devices (AMD), Qualcomm, and Nvidia, tap foreign partners to fabricate their chips. Management at AMD has argued that the law should be written more broadly to give these companies credit for the research and design work that they do domestically.
AMD’s point has merit, but the legislation was done to secure the supply of chips, and that would seem to demand domestic manufacturing, wherever the research and design are done. If others set up U.S. manufacturing operations, they would presumably also get the subsidies and tax breaks, and Intel would get relatively less of the total. Moreover, domestic supplies would be much less concentrated and hence more secure. Perhaps this logic captures the thinking of those who wrote the legislation.
Meanwhile, more than four-fifths of the allocated funds would go to activities other than chipmaking. About $100 billion—almost twice the share allotted to the chipmakers—would go to the National Science Foundation to set up technology hubs in regions of the country with little technology activity. Funds would also go to the Department of Energy for green energy initiatives. It may be a bit of a stretch to link green energy to chip security, but there it is in the legislation.
Funding would also go to establishing a Directorate for Technology, Innovation, and Partnership with what appears to be a broad mandate to provide support for all sorts of technology.
The National Aeronautics and Space Administration (NASA) would receive substantial funds for its exploration of Mars. Other monies would go for research on blockchain, low-emissions steel manufacturing, and the production of more efficient, quieter airplanes. The legislation emphasizes STEM (science, technology, engineering, and mathematics) education at all levels, from high school through post-graduate work. This way, perhaps, the effort can produce the staffing for the new technology hubs without the need for a great migration from existing hubs to new ones.
As with all Washington spending legislation, this one includes a long list of conditions before any entity can receive funding. Much of this focuses on the by now familiar issues of inclusion and diversity. More than 30 percent of the legislation’s language concerns diversity and sexual harassment issues, while 60 percent of its language dwells on requirements generally, including how products should be shipped.
The worth of all this is, of course, debatable. It’s unclear whether the effort will do much to create more domestic chip manufacturing. After all, Intel was already planning new facilities. Now, it might just substitute government for private funding. On the many initiatives that make up the bulk of the legislation and the spending, the detail is so great that even the government scorers have refrained from drawing conclusions.
What’s sure is that Congress has just put the U.S. taxpayers on the hook for an additional $280 billion.