March 11, 2024 ☼ The Intersection ☼ Information Age ☼ geopolitics
Taking a step back to understand why national security types are concerning themselves with semiconductors
This is from The Intersection column that appears every other Monday in Mint. This is an unedited, pre-publication draft.
The United States is going after the Chinese semiconductor industry with a ferocity that has very few precedents. Driven by a national security doctrine aimed at denying China the ability to exploit American technology to threaten America’s interests, Washington has been tightening the screws on its own industry and that of its allies since the summer of 2022. In addition to export restrictions and employment controls, the US government has been pushing Taiwan, the Netherlands, Japan, South Korea and Germany to squeeze the sale of manufacturing equipment, critical parts, raw materials and on-going service contracts with mainland Chinese companies. Wang Yi, China’s foreign minister recently called the sanctions “reaching bewildering levels of unfathomable absurdity.”
No one likes the prospect of cutting themselves off from the Chinese market — which used to purchase half the world’s chip production — and Washington’s policy is unpopular, painful and costly. Yet the United States is doubling down on what it calls the “small yard, high fences” strategy.
Why has geopolitics become obsessed with semiconductors? It is because chips are the most important physical manifestation of the currency of power in the Information Age. Data, algorithms and intellectual property are abstract, only hardware and people are palpable. Computers, networks, vehicles, equipment and armaments are also physical manifestations, but chips are in everything and everything depends on chips. They are zero-sum goods in a massively non-zero sum economy, so controlling them is seen as a way to remain powerful in this era. Whereas power once came from controlling land, sea routes, gunpowder, factories and nuclear weapons, it now comes from chips. At least in the eyes of world leaders and top policymakers.
To what extent are chips the source of geopolitical power? When launching the crackdown on China’s semiconductor industry, the US national security advisor said that “These restrictions are premised on straightforward national security concerns. These technologies are used to develop and field advanced military systems, including weapons of mass destruction, hypersonic missiles, autonomous systems, and mass surveillance.” Yes, there is a military dimension but it is a narrow one. Having access to the most advanced chips allows the United States to dominate the global economy. In denying such chips to China, Washington is not merely limiting a military rival, it is containing a strategic geopolitical competitor. At least for a while.
Washington’s actions set China back at least five but potentially more than ten years. Although Huawei has demonstrated 7nm microprocessors and GPUs in recent months made with pre-sanctions Western equipment, the Chinese industry will at best be capable of 14nm chips this year, when the US is at 2nm and about four generations ahead. China will also have to build its own domestic eco-system and external supply chains. A spokesman for China’s parliament declared that “For any technology known to man, the US cannot choke China’s development. It is merely a matter of time before we prevail.” Some of this is bravado, but it is not beyond reasonable imagination. What you can do with a chip depends not only how fast the chip is but also on how you do it. Innovations in computational methods, processing algorithms and applications can offer alternative paths to development.
But it will cost the Chinese government hundreds of billions of dollars over the next few years. Beijing reckons it has the money and a new vice-premier has been put in charge of the mission after the previous team was put behind bars for corruption.
We should expect the global semiconductor industry to split into two competing eco-systems over the next decade. They will intersect at some levels but sharply diverge at the cutting edge. In other words, while the leading manufacturers will be compelled to isolate themselves from each other, a large number of global suppliers at the middle and lower levels of the value chain will be able to work with both Western and Chinese technologies if they choose to and their governments permit. As you might have noticed by now, the tech industry will operate in a world of intensifying political and policy risks. Even open source technologies have become risky, as Washington’s unwise moves against the RISC-V ecosystem indicates.
India’s public support for building basic semiconductor manufacturing capability is well-considered. No country has done so without massive government support. Similarly, no country — including China before Xi Jinping — did so by being on the wrong side of the United States. India is in a good position but New Delhi must figure out how to take greater advantage of East Asian supply chains. One of the conclusions in my colleague Pranay Kotasthane’s book on semiconductor geopolitics is that “dependence on East Asia is unlikely to go away over the next two decades despite the massive investments happening across geographies.”
Tailpiece: The massive opportunity costs and weaker network effects are not the only price the world is paying for this fight over chips. To the extent that the turmoil in the semiconductor sector affects the pace of energy transition, the damage to global climate might be more significant.
There are many more The Intersection columns here
If you would like to share or comment on this, please discuss it on my GitHub
Previous © Copyright 2003-2024. Nitin Pai. All Rights Reserved.