Therefore, when U.S.-based semiconductor companies observe (rightly) that their movement to Asia significantly reflects efforts to benefit from Asian government subsidies and tax breaks that Washington has failed to match, they’re sending a message that they’ve moved from a country that has long seemed pretty blasé about nurturing that industrial commons to lands for which it’s a top priority. And the same can be said about the vast gap between how manufacturing overall has long been valued in East Asia versus its status in the United States—where for decades political leaders, influenced by the economics and finance establishments, came to insist that industrial manufacturing no longer matters. Indeed, as early as 1985, then-President Ronald Reagan declared that “the progression of an economy such as America’s from agriculture to manufacturing to services is a natural change.”
Intel, moreover, has been no exception. As the company’s own records show, the United States’ share of its global manufacturing production has fallen from nearly 70 percent in 2004 to a “majority” in 2019. In 2004, 60 percent of its 85,000 employees worldwide were located in the United States. By 2019, its worldwide headcount had grown to 110,800 but the U.S. share was down to 49 percent.
EVEN SETTING aside the current conventional wisdom in Washington that government can’t pick industrial and technological winners remotely as well as the private sector, the reasons for doubting America’s ability to reinvigorate the semiconductor manufacturing industrial commons are easy to tick off. There’s the sheer complexity of producing defect-free (and therefore profitable) quantities of ever more powerful chips, each of which contains billions of transistors that are ever smaller micro-fractions of a human hair wide and thus able to handle a ballooning number of increasingly challenging tasks. There’s the need to recreate an entire end-to-end supply chain (including customers) as opposed to assuming that the industry can survive as a series of multi-billion-dollar chip fabs existing in relative isolation. There’s the short-term and strategically indifferent focus of America’s current dominant shareholder-oriented form of capitalism. There’s the deep ambivalence about any measures smacking of enduring de-globalization on the part of both by Intel and its fabless counterparts. The former, after all, remains jittery about any U.S. government decisions that might endanger its access to its enormous Chinese customer base. The fabless companies depend heavily on China sales too, and from their individual corporate standpoints, reliance on foreign foundries like those of Taiwan’s TSMC has so far worked out swimmingly.
Further, as suggested by many of these obstacles, there’s the powerful inertia—in the public sector as well as the private, intellectual as well as organizational—behind the globalist view that as long as the U.S. economy can even access products like semiconductors, it doesn’t matter where they’re made—especially when the foreign suppliers are military allies. When this access faces serious enough potential threats, steps like sending another carrier battle group to East Asia-Pacific waters or issuing reassuring boilerplate diplomatic statements could understandably seem preferable to moving manufacturing back home.
In fact, it is precisely because of this inertia and the real obstacles to promoting adequate domestic semiconductor manufacturing that switching from a foreign policy-dominant approach to a domestic policy-based approach won’t happen overnight, even with a whole-of-economy-and-society campaign.
It’s true that Americans can take some comfort from the emergence of ways other than shrinking circuit size to drive semiconductor innovation. With speculation rife that the physical limits of shrinkage possibilities are uncomfortably close, and the cost of progress becoming stratospheric, tinier and tinier transistors might not even be the most promising way to improve semiconductor performance, at least in the medium-term.
For example, packaging entails producing the cases and other products that protect the circuits themselves from corrosion and other forms of physical damage, and through which pass the signals that operate all the devices that use them. Long dismissed as one of the low-tech “back end” phases of chip-making and largely offshored to Asia, players in this segment are now valued for their own potential to boost semiconductor performance while preventing problems like overheating. Intel itself has started to tout its prowess in this and similar fields.
At the same time, circuit shrinkage efforts aren’t going away any time soon. TSMC is aiming to deliver a chip to Apple this fall that is a generation ahead of the device on which Intel has just whiffed. They’ll be powering the first iPhones that use the much-ballyhooed 5G communications technology. The Taiwanese manufacturer has also unveiled plans to produce even more capable processors by 2022, and has begun developing their successors as well.
YET DESPITE the bipartisan support for the microchip bills, and still-mounting bipartisan alarm about China-related threats, legitimate questions can be asked about how enduring and serious Washington’s interest in domestic chip manufacturing will be. After all, the U.S. government has been fretting about American industry’s prospects since the 1980s, but its domestic manufacturing capacity still shrank to a small fraction of the global total. Maintaining U.S. forces in the Asia-Pacific region to deter Chinese political pressure and military aggression will therefore be essential for the foreseeable future. But these units should be recognized as a wasting, and possibly already wasted, asset, as growing Chinese and North Korean nuclear and conventional military power may have already eroded the escalation dominance that for decades bolstered the credibility of American defense guarantees. In essence, these countries are either amply capable of launching nuclear retaliatory attacks on the American homeland (China), or are dangerously close to creating intercontinental delivery systems (North Korea). As a result, the America First approach might not be simply the most efficient route to regain national semiconductor and tech vibrancy. It might also be the safest.
Alan Tonelson is the Founder of RealityChek, a blog on economics and national security, and a columnist for IndustryToday.com. In 2016, he advised both the Trump and Sanders campaigns on international trade issues.