America’s 78-year-old president, Joe Biden, might not seem like the world’s most tech savvy leader. But some of his administration’s biggest policies have focused on a small but fundamentally important technology: the semiconductor. Known more prosaically as “computer chips”, semiconductors power technology of all types from vast data centers to the simplest consumer electronics. In the age of the Internet of Things, almost every electrical device has a semiconductor inside. And as societies gather and process more data, it is chips that provide the computing power on which we rely.
The Biden administration is thinking about chips in part because of a semiconductor shortage caused by the economic dislocations of COVID-19. Car factories worldwide have been delayed or stopped because crucial semiconductors are on back order. But today’s semiconductor shortage will eventually pass. The bigger issue is geopolitical, as China invests heavily in chip technology, seeking to upgrade its economy and its military. The United States, which has historically dominated chip design and production, is fixated on retaining its leading role. Backed by bipartisan majorities in Congress, the Biden administration wants to use financial incentives to bolster Americas chip industry.
The first chips emerged out of the Cold War arms race. The first computers to use chips on a large scale guided the Apollo space spacecraft and the Minuteman ballistic missile. In the last two decades, though, the semiconductor industry has thought mostly about technology and business efficiency, not politics. Supply chains have become more global than ever. A typical advanced chip is designed by people in the US, UK, or Israel; fabricated in Taiwan or South Korea using machines made in the US, Japan, the Netherlands, and Germany; and then tested in packaged in Malaysia or Singapore. It is hard to think of a product that is more globalized.
Now China wants to become a major player in chips. Beijing has poured tens of billions of dollars of state funds into its chip industry to domesticate advanced technology. Today China cannot produce advanced chips. Its production of moderately advanced chips relies on foreign technology. But China has set the goal of self-reliance in semiconductors. Its track record of rapid economic growth and its manufacturing prowess means this goal should be taken seriously.
Moreover, China realizes that its reliance on foreign chip technology is a fundamental vulnerability. Over the past few years, the US has restricted the transfer of advanced chips to several dozen Chinese firms that bolster China’s foreign policy, most notably Huawei but also more specialized military producers. There’s no way to produce an advanced semiconductor today without using US software and manufacturing equipment, so facilities in countries like Taiwan and South Korea have cut off Huawei. Even some of China’s own fabrication facilities have reportedly stopped selling to Huawei, to ensure they can continue accessing crucial American equipment.
A digital cold war isn’t something to be avoided; it is reality. Yet like in the original Cold War, when Europe bought Soviet gas and the USSR used IBM computer software, the policy options today are not simply “do nothing” or “completely decoupling.” China will continue to rely on the global chip supply chain for the foreseeable future, and the US will permit the sales of most semiconductors to China. But decisions that used to be made purely on business grounds now inevitably involve calculations about geopolitics, too. Attention is focusing on “chokepoint” technologies that are crucial to chip production but are produced by only one or two countries. The United States today can control multiple chokepoints, including design software and production equipment; China controls none.
Other countries will insist they want to stay neutral. They have every incentive to say this, but in reality each major country is jockeying for its position in the supply chain. All players fear Beijing’s state subsidies and the risk of intellectual property theft. South Korea, Europe, and Japan have all announced major new government spending on their chip industries to bolster existing firms. Controlling a chokepoint isn’t only economically valuable—it also provides countries with newfound geopolitical influence, a useful bargaining chip in relations with Washington and Beijing.
Amid this global competition for leverage in the semiconductor industry, the Biden administration is trying to catch up. The US government is re-learning how to think strategically about semiconductors after two decades of mostly ignoring the industry. Facing the global politicization of semiconductor supply chains and state subsidies from other chip producers, Washington has concluded it has no choice but to device an industrial policy strategy to support semiconductors.
The Biden administration’s strategy is threefold. First, it wants to use existing chokepoints to secure important objectives (e.g., impair Huawei's 5G rollout globally) without overusing them in a way that limits these chokepoints' future efficacy. This requires more subtle communication with allies than the Trump administration’s tactic of twitter diplomacy. Second, the Biden administration wants to coordinate multilateral controls with allies over semiconductor chokepoints. Countries like the Netherlands, Germany, Japan, and South Korea all produce advanced manufacturing equipment or could potentially produce it soon. Multilateral export controls would be less leaky than unilateral US restrictions.
Multilateralism sounds nice. Is it possible? Getting buy in from South Korea is difficult given that country’s economic dependence on Beijing. Support from Germany, meanwhile, requires Berlin to think clearly about geopolitics, something that the German elite is hard-wired not to do. Companies in all countries howl when told their sales to China will be restricted, though off the record many semiconductor executives admit they understand what is at stake.
America’s chip industry would much rather have the third leg of the Biden administration’s strategy, a major government spending program to back the construction of fabrication facilities and R&D. Congress looks likely to pass up to $50 billion in subsidies, a sizeable sum, though far smaller than the hundreds of billions that some Chinese sources speak of. There is a real risk that too much of this sum supports established chipmakers rather than small startups. As countries jostle for position in semiconductor supply chains, government funds will only be one variable among several. More important, in the end, will be the companies that build the innovative technologies and disruptive business models that define the industry in the future.