As discussed here repeatedly, semiconductors are at the center of the debate in Washington about whether the United States should (once again) embrace industrial policy. In particular, numerous elected officials and wonks support appropriating tens of billions of new taxpayer dollars, initially approved under the 2021 National Defense Authorization Act, to subsidize the construction of semiconductor manufacturing facilities (“fabs”) on U.S. soil. Among the primary reasons for why such subsidies are urgently needed is the claim that American chipmakers have lost their edge and either can’t or won’t build new fabs here without the same kind of subsidies that the Chinese government has lavished upon its industry (lackluster results notwithstanding).
Not everyone agrees, however, including not just libertarian skeptics like me but also… the new CEO of Intel, the United States’ largest chipmaker. In announcing the company’s new $20 billion plan to build two new fabs in Arizona, in fact, Intel’s head honcho expressly said this week that the company doesn’t need taxpayer help (though of course, they won’t object to free money!):
While specifics were not given out, Gelsinger said that Intel was expecting to receive incentives from the Biden administration for building out indigenous fab capacity here in the United States and also incentives from the state of Arizona for the $20 billion foundry buildout revealed today. Not that Intel needs it, apparently.
“This is the Intel strategy, period. Full stop,” declared Gelsinger. “It does not depend on a penny of government support or state support or any other investments to make it successful. It is the right strategy for us going forward,” he said, adding that “of course, governments – we want incentives. We want investments because it is the right thing for them to accelerate manufacturing and this imbalance of the global supply chain. But we’re making these commitments without any commitments from the governments to accelerate them.” Intel is, as Gelsinger punned it, “putting our chips on the table” first and then anticipating that eager governments will up the ante some.
Despite subsidy supporters’ claims, Intel’s position here shouldn’t be surprising. First, research has repeatedly shown that subsidies rarely tip the scales of corporate decision‐making and are most often considered after a recipient’s plans have been made. (Intel rival Samsung has reportedly said similar things in private about U.S. subsidies for a new semiconductor plant in Texas.)
Second, despite recent and well‐publicized struggles producing top‐end chips, Intel is still incredibly profitable and flush with cash. According to the Wall Street Journal, for example, “Besides the $20 billion earmarked for the new Arizona fabs, to be spent between now and 2024, Intel… expects to have between $19 billion and $20 billion in total capital spending this year—a record amount that is about 45% above the company’s average annual capital expenditures over the past five years.” Last year, in the middle of a global pandemic, the company spent only (only!) about $15 billion on capital expenditures, leaving it with almost $18 billion in free cash flow.
Intel’s problems, it seems, have had nothing to do with a pressing need for more capital and much to do with recent decisions about how to spend it. Indeed, the Journal previously reported that “Intel has fallen behind Taiwan Semiconductor Manufacturing in the most‐advanced production processes, even though the American chip maker has outspent its Taiwanese rival by an average of nearly 6‐to‐1 on research and development annually over the past 10 years.” There’s little reason to think that billions of taxpayer dollars, untethered from market discipline, will improve Intel’s new approach (and, in fact, there’s plenty of evidence that it could do just the opposite).
The federal government should therefore pocket any planned subsidies to Intel — just as Intel’s CEO himself has suggested.