Westerns governments are no longer comfortable leaving leading-edge chip manufacturing to Asian companies. But is subsidizing fabs the solution? It’s a precarious dilemma.
“This is infrastructure,” US president Biden, holding up a silicon wafer, told a group of CEOs. Offering 50 billion dollars in support, he urged the corporate leaders to invest in domestic semiconductor manufacturing. European Commissioner Thierry Breton is having meetings with TSMC Europe president Maria Marced and Intel CEO Pat Gelsinger to discuss the possibility of these companies building a fab in Europe. Japan, too, tried to woo TSMC into building a chip factory on Japanese soil. Even India is handing out cash to anyone setting up a semiconductor manufacturing facility.
Governments around the world have had a change of heart. Going against a two-decade-long trend of going along with companies outsourcing IC production to Asia, they want their ‘own’ fabs now, and they’re willing to pay a pretty penny to get them. A necessary move to secure access to a strategic technology in an increasingly unstable world? Or a fool’s errand? Breaking down the bullet points for both sides of the argument.
Yes, subsidizing those fabs is a good idea
- The Semiconductor supply chain, in general, has become vulnerable to natural disasters and geopolitical disruptions because suppliers have become more concentrated in distinct regions, according to a study by the Semiconductor Industry Association (SIA) and the Boston Consultancy Group (BCG). Leading-edge manufacturing capacity, however, has practically become a single point of failure: 92 percent of sub-10-nanometer production is located in Taiwan. The island is prone to natural disasters, such as earthquakes and droughts. Moreover, China considers it a breakaway province that needs to be reunited with the motherland. These risks can’t be tolerated: fabs need to be distributed more evenly around the world to minimize supply disruptions.
- Chips have become a natural resource of sorts: without them, economies and societies can’t function. As such, they’re an indispensable part of any country’s economic security. Chips are also a matter of national security, in part because they power military technology and digital infrastructure, but also because their production can be used as a pressure point in geopolitical maneuvering. Any country that allows itself to be overly dependent on others for such a strategic resource weakens its clout on the world stage.
- Markets should welcome new fabs because they’ll be sorely needed to meet future demand. AI, the IoT, healthcare – digitalization has only just begun. Ever more products are fitted with chips, and ever more chips are used to enhance products. CEO Ajit Manocha of trade organization Semi estimates that the semiconductor market will roughly double to 1 trillion dollars in 10-15 years. That prediction follows the historical trend: it also took about fifteen years for the market to double to 480 billion dollars, the expected size at the end of this year.
No, governments should stay away from fabs
- What good would having your own fab really do? The semiconductor supply chain is still highly globalized. A substantial part of the EDA tools, intellectual property, designs, materials and equipment would still have to come from abroad. Once the wafers are processed, the dies need to be packaged and integrated into a device, which still may not be the final product. It’s a pipe dream to think that this entire supply chain could be reconstructed in a single country.
- The globalized supply chain is already a pretty good way to ensure access to semiconductors. Each country or region has its own specialties, upon which others rely. Rather than pursuing technological sovereignty, governments should focus on making sure that their leading businesses stay ahead. Should there still be worries about supply vulnerabilities, international collaboration to strengthen or diversify supply chains, rather than techno-isolationism, is the answer.
- If governments push forward with their investment plans without regard for the markets, it will negatively impact the semiconductor industry. At the very least, it will reduce capital efficiency, raise prices and harm innovation rates. Worst-case scenario: so much additional capacity will be installed that it would send the industry into a devastating downturn, endangering the survival of companies as well as the return on investment for the taxpayer.