OPINION:
The COVID-19 crisis has brought into stark contrast the need to secure our supply lines and bring them closer to home. While most of the focus has been repatriating manufacturers of medical equipment, pharmaceuticals and other critical supplies to fight this pandemic, ensuring our technological independence will prove similarly important in the coming months and years.
Semiconductors, the chips that power everything from cellphones to fighter jets, have become so important to national security that the Department of Defense recently announced it was shifting its top tech priority from hypersonic missiles to microelectronics. The reasoning is simple: This technology is a part of almost every American weapons system, and we are in danger of losing our edge in this critical field.
China has made it a key part of their national strategy to grow their semiconductor industry while capacity in the United States has stagnated. From 2015-2019, China’s semiconductor manufacturing capacity has as a share of global output that has increased by one-half to 12% of all global output, threatening to overtake the United States, which accounts for 13% of all semiconductor manufacturing capacity.
It seemed to be good news therefore, when the Taiwan Semiconductor Manufacturing Co. (TSMC) recently announced plans to break ground on a $12 billion factory in Arizona in 2021. Touting more than 1,600 high-tech jobs created in America, this initiative has been portrayed as a means to reduce America’s reliance on foreign producers and chip manufacturing in Asia, all while apparently boosting national security.
Unfortunately, this narrative starts to unravel when digging a little deeper into the facts of the deal. Constructing this facility will not help America win the semiconductor race against China nor lessen our dependence on foreign manufacturers. A better deal must be negotiated to truly realize these goals.
For starters, advances in semiconductor world move very quickly and the technology to be produced will likely be obsolete before a single chip is ever manufactured. While the projected TSMC facility in Arizona will construct 5 nanometer (nm) chips, next generation 3nm chips are just a few years out. This U.S. TSMC facility will likely be a generation behind by the time the first chip comes off the assembly line (projected for 2024) but ironically, TSMC is scheduled to start installing 3nm equipment in Taiwan in October of this year.
A second point of concern relates to volume of production. Not only will the facility be a generation behind when it comes online, but its capacity will be significantly lower (20,000 wafers/month vs. 100,000 per month) than the output of TSMC’s so-called Gigafab manufacturing facilities in Asia. This indicates that it will not be able to meet the scale needed for the U.S. to achieve true semiconductor independence.
Furthermore, this facility as currently proposed will not be enough to meet America’s national security needs. Majority-owned foreign semiconductor firms are rightly not considered a “trusted foundry” by the Pentagon and are limited in how they can help fulfill the research and development needs of the military. This is why the Department of Commerce negotiated the TSMC deal independently from the Defense Department and why the Pentagon is expected to continue its own efforts to bolster the supply chain for microprocessors.
In short, the current deal is insufficient for technical, economic and national security reasons. A new, better deal with an American firm at the helm of a joint venture with TSMC is what is needed to ensure our technical leadership and truly shore up America’s national security.
TSMC is one of the most advanced chipmakers in the world, and a partnership makes sense. But the Trump administration should look to negotiate a better deal that includes cutting-edge technology and sufficient production volume. Ensuring some level of investment from an American company will also be crucial to guarantee long term alignment with U.S. interests, regardless of what may happen politically in Taiwan or with TSMC management in the future.
The right partnership with an American company could also help America maintain its semiconductor edge against China. Under the right circumstances, this facility could serve as a new Defense research center that would allow government and private corporations to cooperate on R&D and help restore critical American technological leadership.
The United States has slowly lost most of its manufacturing capabilities to China over the last 30 years and is nearing a “point of no return” when it comes to losing our ability to manufacture cutting-edge semiconductors as well. America needs to craft a solution to end our dependence on Asian semiconductors. The current TSMC Arizona deal, while promising on the surface, will not accomplish those goals.
President Trump and his administration should aim higher for a better deal, truly bringing critical semiconductor manufacturing “onshore” to satisfy essential economic and national security imperatives.
• Bob Dees, a retired U.S. Army major general, is president of Resilience Consulting LLC.
Please read our comment policy before commenting.