CHIPS 1, CHIPS 2 . . . CHIPS X

Semiconductor titans are angling to feast on new public subsidies even before they have digested the first tranche. Two weeks ago in Davos, Intel CEO Pat Gelsinger, in effect, made a strong call for a CHIPS 2, referring to the $52 billion CHIPS and Science Act passed by Congress in August 2022. Not to be outdone, OpenAI’s Sam Altman is now lobbying key congressional leaders for a government-backed venture to build new manufacturing factories (fabs in industry parlance) for advanced chips to underpin the burgeoning artificial intelligence industry, independently or in partnership with existing companies such as TSMC and Samsung. All of this comes before the Commerce Department has dispersed large sums ($39 billion) for such construction and operation under the CHIPS Act.

At Davos, Gelsinger gained support for his view that more government subsidies for the semiconductor industry would be necessary from Arati Prabhakar, the Biden administration’s director of the White House Office of Science and Technology, who stated that to catch up on the “leading edge,” semiconductors manufacturing “must be supported by the government.”

As to specific motivation, Mr. Altman is noted for highly ambitious blue sky proposals, while Mr. Gelsinger is in the midst of a heroic effort to return Intel to its position as a top competitor in the design of advanced (3/2 nanometer) semiconductor chips. Intel is also attempting to reenter manufacturing chips for additional companies not dominated by TSMC.

Whatever the outcome here, the idea of a succeeding CHIPS acts raises several issues. Inevitably, political questions are intertwined with economic, technological, and security questions. As noted previously, the view here is that the strongest argument for the CHIPS Act’s advanced fab subsidies is based on national security considerations. At this point, Taiwan’s TSMC manufactures about 90 percent of the world’s most advanced chips. Given its fraught relations with an ever-more aggressive China, it makes sense for the US and its allies to build backups. However, this does not constitute an open-ended commitment to public support for all elements of the semiconductor supply chain. Mr. Altman’s case goes beyond security and at base argues for public support on a “competitiveness” basis.

At this point, nearly eighteen months since the CHIPS Act passed in August 2022, the Commerce Department has received almost 200 applications and concept proposals, but has only dispensed several hundred million of the $39 billion allotted in the legislation. New advanced chips fabs cost $10 billion and more, and the leading companies for these giant projects—Intel, TSMC, and Samsung—have made it clear that they will demand substantial public funds. As always in Washington, political considerations come into play, with the Washington Post noting about the upcoming big dollar grants: the “Biden administration hopes chips will give it an election year boost.” So, it is no surprise that over the past weekend, the Wall Street Journal reported that before the president’s State of the Union, the administration plans to announce billions of dollars of fab subsidies to companies such as Global Foundries, Intel, Micron, TSMC, and Samsung.

It is hard to know how much heft calls for new chip fab subsidies will gather. Given the upcoming election and a divided Congress, it is unlikely that new semiconductor programs will be acted upon this year. From the perspective here, that is all good. Before embarking on large-scale public or public/private programs, the Biden administration (and whoever becomes president in 2025) needs to dig much deeper into the technological and economic rationale for future US semiconductor policy. And more specifically, which chips and how many chips are truly vital for the US defense effort.

As this blog was being completed, Chris Miller, AEI Nonresident Senior Fellow and the leading US authority on the “Chip Wars,” analyzed the extraordinary complexities also facing responses to China’s plans to flood international markets with less advanced chips. While the emphasis in this blog has been the dangers of Washington–based analysis overplaying “market failure” in defense of public intervention, there is also, in keeping with Chris Miller’s compelling analysis, an awareness that “security failure” is real and a comprehensive response to competition-destroying Chinese tech/trade policies is also imperative.