The Semiconductor Supply Chain Reshuffle Is Not What It Was Sold As

The reshuffle of the semiconductor supply chain is, in the political language that surrounds it, a story about independence. The concentration of chip manufacturing in a few places, exposed by the disruptions of recent years, has been treated as a vulnerability to be corrected, and the correction has been pursued through the subsidies and the politics that such corrections require. The reshuffle is happening, and it is producing something, but the thing it is producing is not the independence the politics promised, and the gap between the promise and the product is the thing the industry will be living with for the next decade.

The concentration that the reshuffle is meant to address is real, and the vulnerability it produced was exposed, brutally, by the disruptions that interrupted the flow of chips to the industries that needed them. The factories that make the most advanced chips are located in a small number of places, and the places are subject to the geopolitical and the natural risks that concentration always invites. The case for diversifying the manufacturing is not in dispute. The question is whether the diversification the subsidies are buying is the diversification the case requires.

The fab that does not make the thing it was built to make

The first thing the reshuffle has revealed is that a semiconductor factory is not a single thing, and the building of one does not produce the chip the politics promised. The most advanced chips require, in addition to the factory, an ecosystem of suppliers, a workforce with specific skills, and a generation of learning that the new factories are, in their early years, without. The factory that is built in a new place makes, at first, the chips that the ecosystem can support, and the chips it can support are not the most advanced ones, and the most advanced ones are the ones the politics of independence was about. The gap between the factory and the chip is the thing the subsidies paid for and the years will have to close.

The closing, if it happens, will take longer than the political cycle that authorized the spending, and the lag is the thing that will test the commitment of the governments that undertook it. The factory that does not make the most advanced chips is a factory that requires, to justify its subsidy, the patience to let it learn, and the patience is the thing the political systems that funded it have not historically shown. The factories that survive the learning will be the ones whose governments sustained them through the years they were not yet what they were built to be, and the sustaining is the thing the reshuffle's success actually depends on.

The economics the politics overrode

The second thing the reshuffle has revealed is that the economics of chip manufacturing do not change because the politics do, and the manufacturing that concentrated where it did concentrated for reasons that the subsidies are working against. The places that make the most advanced chips do so because they accumulated, over decades, the conditions that make the making possible, and the conditions are not transferable by spending. The factory built in a new place is more expensive to operate, for the chips it can make, than the factory that already exists in the place that makes them, and the expense is the thing the subsidy covers and the market, left to itself, would not.

The covering is sustainable only as long as the subsidy continues, and the subsidy is the thing the political systems can withdraw, and the withdrawal is the risk the new factories carry. The factory that cannot survive without the subsidy is the factory that exists at the pleasure of the government, and the pleasure is the thing the industry has learned, over its history, not to rely on. The reshuffle has produced factories that the economics did not justify and the politics did, and the factories are the thing the industry will have to integrate into a market that did not ask for them.

The interdependence that did not disappear

The third thing the reshuffle has revealed is the one the politics of independence most wanted to deny: the interdependence of the supply chain did not disappear when the factories were built, because the factories are one link in a chain that extends in both directions. The materials that the chips require come from places the factories are not, the tools that make the chips come from a small number of suppliers the factories depend on, and the design of the chips is done by companies that may or may not locate near the factories that print them. The independence the reshuffle promised is the independence of one link, and the chain remains as interdependent as it was, and the interdependence is the thing the politics of the reshuffle preferred not to mention.

The semiconductor supply chain reshuffle is, in its honest accounting, a partial diversification purchased at a high cost, and the diversification is real but narrower than the politics claimed. The factories are being built, the learning is beginning, and the chain is, in the link that was relocated, somewhat less concentrated than it was. The rest of the chain is as concentrated as ever, and the concentration is the thing the next disruption will expose, and the disruption is the thing the reshuffle did not actually protect against. The independence was a promise. The diversification is a fact, and the fact is smaller than the promise, and the gap between them is the thing the industry and the politics will be arguing about for as long as the factories the subsidies built are in operation.