The semiconductor value chain is simultaneously global and highly concentrated. Fabs — the large factories that create chips out of semiconductor wafers — exist in only a handful of countries. South Korean firms and the Taiwanese company, TSMC, account for most leading-edge chip manufacturing. Companies that don’t manufacture their own chips feel the effects of disruptions to the global value chain because of that market concentration. Natural or other issues that create disruptions can have outsized effects on firms up the supply chain.
Many countries have now moved to increase chip supply chain security by onshoring more semiconductor production capacity. China has long aimed to increase domestic manufacturing, while the US 2022 CHIPS Act contains provisions aimed to bolster domestic capacity. It will take years for the planned capital investments to bear fruit, and risks will remain even then.
Fabs are the single greatest expression of semiconductor production, but they are wholly dependent on other parts of the global supply chain. No fab can operate without access to specialised and often esoteric inputs. The concentration in Japan of suppliers of several critical chemicals used in semiconductor manufacturing led Tokyo to try to exercise leverage over Seoul in a political confrontation over wartime reparations in 2019.
The war in Ukraine has similarly threatened the global supply of several critical materials used to make chips. The response by some firms has been to diversify their supplies, but the necessary natural resources are not evenly distributed across the planet. The same constraints exist among the suppliers of semiconductor capital equipment. Only a few companies — primarily in Japan, Europe and the United States — produce the machines used at different stages of chip production.
There is precedent for the criticality of these nodes. The Soviet electronics industry lagged behind its competitors because there was a coordinated effort to keep them from accessing internationally state-of-the-art equipment. By the end of the cold war, the USSR was about a decade behind technologically.
Chip manufacturing requires myriad designers with access to libraries of intellectual property (IP). Some of the largest semiconductor companies don’t make any physical items themselves but send their IP to be made by fabs owned by other firms, often in other countries. Without this, firms can’t manufacture the leading-edge chips that power smartphones and other advanced devices.
Chinese firms ZTE and Huawei felt the impact of this once they lost access to US IP. Replicating these resources is difficult, as it requires a pipeline of talent from universities into the industry that feeds into an ecosystem of state-of-the-art research and development.
The urge felt by governments to invest in domestic semiconductor manufacturing is a natural instinct. But it will prove difficult, if not impossible, for the vast majority of countries to achieve complete independence based on their material limitations. Semiconductor investments take years, if not decades, to pay off.
Top-to-bottom onshoring is unlikely to mitigate supply issues in the short term, nor would it be feasible for all but the wealthiest states to even attempt. For most countries, a cost-effective way to reduce risk across supply chains could be to form more robust multilateral partnerships. A constellation of states with similar policy goals might be better positioned to shore-up shared bottlenecks and deficiencies.
Participation in such arrangements would require giving up their virtual monopolies over aspects of the supply chain. New sources of raw materials and additional manufacturing capacity spread across multiple countries would help prevent a disruption in one part of the world from shutting down manufacturing elsewhere. Enabling such diversification would require greater standardisation to ensure materials are compatible across the group.
Individual states could also analyse their supply chains for the points of greatest vulnerability. Such analysis would show where the requisite materials, capital and capacity are located and any deficiencies that may arise due to disruptions. The next step would be to isolate the greatest risks and determine the potential for action. If the resources are present and the cost is not exorbitant, domestic investment may make sense. If alternative global suppliers exist, diversifying supply chains may prove sufficient.
Such arrangements could create friction. Diversification costs money. The global industry may resist policies that create new costs for risk mitigation. Market access has been a critical goal for many of these companies over the past several decades, with large markets like China’s providing a massive incentive to resist policies that would restrict their ability to do business globally.
The increased competition and delineation of the market that would be caused by a unilateral or multipolar push to control chips top-to-bottom could be perceived as a threat. The remaining market would have to be large enough to make losing access to it a worse proposition.
It is difficult to imagine the titanic shifts in the semiconductor sector implied by complete onshoring. The present configuration of the global value chain has inertia measured in hundreds of billions, if not trillions, of dollars. In this context, the diversification of supply, risk mitigation and closer partnerships with other like-minded players in the global ecosystem might provide the best insurance.
Samuel Goodman is a researcher and author based in Washington, DC.