Of all modern industries, semiconductors — i.e., computer chips — are the most jealously guarded and sought-after by national governments, for many reasons. First of all, they’re militarily important — pretty much all weapons are computerized now, and in a war you need to be able to make your own chips. On top of that, semiconductor manufacturing is a very high-value activity that generates a lot of revenue and a lot of good-paying jobs. And since computer chips are key inputs into a huge variety of other high-tech industries, having semiconductors located in your country supports an entire ecosystem of high-value economic activity.
Thus it’s no surprise that the Democrats’ newfound embrace of industrial policy would focus first and foremost on semiconductors. Of the $250 billion allocated in the competitiveness bill just passed by the Senate, $52 billion is to boost domestic semiconductor production. Meanwhile, President Joe Biden’s new review of U.S. supply chains singles out computer chips as a key area in need of reshoring.
U.S. leaders have their work cut out for them. They’re up against China, which has rolled out a huge program aimed at boosting its own semiconductor industry to world-beating position. Japan and Europe are trying to claw back some of the market share they’ve lost. South Korea, whose companies have taken some market share from the U.S. in recent years, is also making its own big push. And Taiwan is home to Taiwan Semiconductor Manufacturing Co., TSMC, which is now regarded as the global technological leader in chip manufacturing after leaving U.S. champion Intel Corp. in the dust.
In other words, everyone is going to be throwing money at chip companies, helping to finance the construction of giant fabricator plants (called “fabs”), doing more research and development, and so on. If the U.S. is able to simply preserve its market share — currently still the world’s biggest at around 47% — it should be counted as a victory. And if the U.S. can persuade TSMC to locate more of its factories in the U.S., that will count as a major strategic success, even though TSMC is a Taiwanese company.
But what would it take to decisively win the semiconductor wars, rather than just running harder to stay in place? To answer that, the past provides a number of instructive examples — most importantly, the battle over semiconductors in the late 1980s and early 1990s.
In the 1980s, Japan’s memory chip industry caught up with America’s technologically and started taking away market share. The US, which was much more protectionist then than now, freaked out and started a trade war. In the 80s, America used high tariffs to force Japan to agree to set aside a portion of its domestic market for U.S.-made memory chips, and to share some of its technology. Meanwhile, the U.S. set up Sematech, a government-organized consortium of private semiconductor companies dedicated to maintaining the U.S.’ technological edge.
None of this managed to save domestic producers of memory chips. The U.S. market share in that industry never recovered from the plunge it took in the early 80s. Japan retained its lead for a few years, then was outcompeted by Korea, which spent big to dominate what by then was an increasingly commoditized and low-margin market. Nevertheless, the U.S. did recover its pole position in the global semiconductor market overall. Because while the world was fighting over memory chips, U.S. companies like Intel switched to making something much more valuable — microprocessors. American CPUs raked in the cash while Asian countries fought over the scraps in the memory industry.
This episode illustrates the importance of innovation. Intel is suffering now, but its multi-decade dominance — which translated into national market dominance — came not via cheap government financing or dogged competition, but by focusing on innovative new types of products. Similarly, later shifts in the industry have mostly been due not to competition in existing markets, but to novel products: GPUs, low-power chips, mobile chips and TSMC’s foundry model.
As the U.S. heads into the latest semiconductor war, it would do well to remember this basic principle. There will certainly be a temptation to throw money at Intel and other U.S. chipmakers to design the next iteration of their existing products. But while that probably doesn’t hurt, history suggests that U.S. dominance of the industry will depend on new companies creating new kinds of chips. This means the government needs to support startups in addition to national champions, make research results broadly available instead of letting companies hoard them, and put money toward disruptive innovations with the power to create whole new markets.
If it comes down to a pure contest of subsidies, the U.S., with its deep political divisions and its traditional ambivalence about public-private cooperation, is likely to lose to China — or at best, to spend a lot of money just to fend off the challenge. We innovated our way to victory in the last semiconductor war; we’ll probably have to do the same this time.