The White House is intent on outcompeting China on technology. The ground on which this competition is taking place is chip making. But the Biden administration shouldn’t sit back and savor this accomplishment for one reason: What if its core belief — that advanced semiconductors are one of the critical fronts in the contest — is wrong?
Over the past six years, the U.S. government has relentlessly targeted China’s semiconductor industry. The Biden administration extended a Trump-era practice of placing Chinese tech companies on trade blacklists. The White House then declared supercomputing chips all but off limits to Chinese companies, saying that they advance China’s military modernization and human rights abuses. It diplomatically engaged with the Netherlands and Japan to jointly deny advanced chip-making equipment to China.
Then advanced artificial intelligence tools like ChatGPT, which require state-of-the-art chips, appeared on the scene. Now the White House is considering creating an investment screening mechanism that could block American investments in China’s semiconductor companies that could advance A.I. Most recently, the Biden administration is reportedly considering a further tightening of A.I. chip sales to China. Powerful chips are at the heart of A.I. development. And the U.S. government is vigilant about closing off China’s means of acquiring them
These efforts have certainly bruised some of China’s largest tech companies. China’s semiconductor prowess — shaky even at the best of times — is now dealing with major stresses as chip makers start to lose access to leading production tools. Most strikingly, more than half a year after Americans have begun to play with A.I. chatbots and image generation tools, Chinese consumers are still waiting for broadly available homegrown alternatives.
America’s actions are driven by the assumption, articulated by the national security adviser, Jake Sullivan, that computing chips are a force multiplier technology, staking it as critical to continued U.S. leadership. But what if the U.S. government is too focused on the most novel technologies rather than the most important ones? I believe America is in a great power contest with China, one that will be multidimensional and protracted, making it unlikely that success hinges solely on who can stay ahead in a few advanced technologies.
And while there’s no denying the potential significance of large language models, it remains far from obvious that America’s mastery of A.I. would really be a decisive advantage over China. In fact, it’s not even clear that Beijing views the present applications of A.I. as being of great importance. China’s leadership, which recently issued regulations demanding that A.I. chatbots must promote “socialist core values” and not challenge the doctrines of the ruling Communist Party, appears to be in no rush to allow this technology to proliferate among its people.
An excessive focus by the United States on A.I. — and on the advanced chip-making capabilities it requires — may represent a failure to appreciate China’s broad technology strengths. While China has suffered serious setbacks in chip production, its companies are vaulting ahead in other sectors. Last year China overtook Germany in automobile exports, and it is on track to overtake Japan as the global leader this year. While most of these exports consist of foreign brands produced in China, the numbers reflect the deep expertise that Chinese companies have built in the next era of automotive technologies, particularly in car batteries.
It’s not just cars. Industry estimates put Chinese companies at owning around 80 percent of the supply chain for solar manufacturing. Chinese electronics makers have produced a rising share of the components in Apple’s iPhone. And increasingly in less glamorous products — such as industrial machinery and basic household equipment — Chinese brands are joining European and Japanese competitors in the world’s top league tables.
The passage of the Inflation Reduction Act will help build up domestic production capacity in solar and in car batteries. Likewise, Congress and the White House would do better to boost other sectors like biotech manufacturing. China produces much of the supply of active pharmaceutical ingredients, and American manufacturers had trouble making masks and cotton swabs early in the pandemic.
The U.S. government is right to withhold technologies that aid China’s military modernization. Only a small percentage of advanced chips, however, go to military uses. And although Commerce Department officials have severely limited China’s access to the most sophisticated chips, they have taken pains to say that Chinese manufacturers can continue making less-sophisticated chips. One of the revelations of the pandemic was the necessity of having a wide range of chips, both cutting edge and basic. Power management chips, for example, are simpler to produce than the graphics-processing units that power A.I. Yet without them, production lines for everything from automobiles to medical devices could come to a halt.
With one hand, the U.S. government is blocking China’s progress on A.I. and supercomputing, but with the other, it is ushering Chinese companies toward concentrating their efforts on chips for products of daily use. And a world in which Chinese companies dominate the production of mature chips — driven directly by American policy — hardly looks like a victorious outcome for the United States.
In spite of so much talk of economic decoupling with a strategic competitor, America is still enormously dependent on Chinese goods, with a goods trade deficit with China in 2022 that was the second-highest on record. If there is ever a serious disruption to trade, it’s far from obvious that American prowess in A.I. will overcome China’s strength of a large and adaptive manufacturing base.
The problem with trying to regulate fast-moving technologies is that there will always be new loopholes to close. That’s especially the case since American chip makers are constantly dreaming up clever ways to maintain their access to the enormous Chinese market. Administrative agencies risk being bogged down in games of Whac-A-Mole while losing sight of strategic objectives.
We need to spend less time making ever more marginal refinements to restricting an emerging technology. Rather, we should take a more holistic view of a long-term contest with a peer competitor. That means broadening the strategic focus to a wider range of sectors and following through on plans to build unglamorous technologies, too.