October 24, 2022The Intersectioninformation agehigh tech geopolitics

Manpower as India’s siliconpolitik strategy

Government support for semiconductor industry development must focus on transforming the industry towards software

Mint This is an unedited draft of my The Intersection column that appears every other Monday in Mint.

Washington’s wide-ranging sanctions on China’s semiconductor industry will go far in containing the United States’s geopolitical rival. Not only will they set China’s chipmakers back by years, but also restrict the country’s progress in several areas, from personal computers to data centres, from artificial intelligence (AI) systems to hypersonic missiles. This may well be America’s most consequential move in the ongoing contest among global powers.

To be sure, a country as large and resourceful as China can overcome the hurdles that have been thrown its way, but it will take time. In the five to ten years it might take to catch up or render the sanctions irrelevant through innovation, the US-centric technology ecosystem would have moved ahead several generations. The opportunity cost of the lost time, in terms of both economics and politics, will be massive. China brought this upon itself. In her latest book Overreach”, Susan Shirk explains how Beijing’s decision to confront the United States began as early as 2006, under the collective leadership of the Hu Jintao years. It took the United States over a decade to recognise this reality, overcome denial and decide to fight back.

How will China respond? Once its political system gets back to work after the grand spectacle and event management of its Party Congress, it is likely to explore a combination of sanctions evasion, circumvention and indigenisation, reaching out to countries such as Germany, Israel, South Korea and perhaps Malaysia and Singapore. Since most of these are US partners, Washington will retain leverage over Beijing’s progress. This route may allow Chinese firms to satisfy their domestic requirements with import-substituted products, but unable to compete in export markets in high-end and high-value products.

There’s another possibility. It is further-fetched and foolish, but we are in a period where it cannot be ruled out. Beijing could well regard a semiconductor blockade as the trigger and pre-text to invade Taiwan and take over TSMC and other manufacturing facilities concentrated on that island. Even if the invasion succeeds, unless the managers and engineers cooperate with the invaders, the PLA would merely be flying the Chinese flag over some high-tech real estate. But they could well turn off the supply of chips to the rest of the world.

Here is my assessment of the state of global siliconpolitik that forms the basis of this column.

That, in part, is the reason why the US and European governments are providing tens of billions of dollars in fiscal incentives to build fabs outside Taiwan, preferably on their own soil. These will take four to five years to come on stream, after which it would theoretically be possible for Taiwanese professionals fleeing the PLA to take a trans-continental flight to their new workstations. Therefore, on current trends, we should expect higher geopolitical risks across the Taiwan straits around 2027.

India has already benefited from the world’s desire to manage the risks to high technology supply chain. Sensing the window of opportunity, the Union and several state governments have announced policy frameworks to attract some of the global investment in the shifting semiconductor industry. Fiscal incentives of over $10 billion to attract investment in fabs and assembly plants are a step in the right direction. India must have the baseline capability to produce advanced semiconductors on its own soil, both to service its national security requirements but also equip its talent pool with the skills to master the entire semiconductor industry cycle.

India must double down on the fundamental source of its strength in the technology industry — the quality and quantity of skilled manpower. The economics of semiconductor manufacturing mean that the hundreds of billions of dollars required to set up fast depreciating state-of-the-art plants are beyond the reach of most governments. While India shoud welcome private investors to build the most advanced plants in the country, public funds are better invested in manpower development. India’s goal should be to use human resources to dominate the semiconductor industry just as it does the IT industry today. It will be a good idea to set up half-a-dozen industry institutes that create professionals in design, production and management of the silicon industry.

Government support has been critical for the development of the semiconductor industry worldwide. Where incentives have been applied to shift production, as in Singapore, they have met with modest success. Where they have been used to fund upstream scientific breakthroughs as in the United States, or change the industry paradigm as in Taiwan, they have had tremendous impact. India should explore how its strength in manpower can transform the industry. Can we do to software what Taiwan did to foundries?

My paper for the inaugural Sydney Dialogue outlines the bubbles of trust approach that I first outlined in 2020.

Every country that has a semiconductor industry today, China included, has done so while being Washington’s geopolitical partner. The same holds true for us. A bubble of trust” with the United States and its partners such as Japan, Taiwan and the EU is crucial to secure investments, technology, markets, expertise and movement of people. For the past three decades, there was no need for economic and geopolitical eco-systems to be significantly correlated; but in post-pandemic era, such a correlation is necessary for economic growth.



If you would like to share or comment on this, please post it on Twitter Previous
Why an ‘everything app’ is bad news for liberal democracies and free markets
Next
If it is a polycrisis, then there is hope

© Copyright 2003-2023. Nitin Pai. All Rights Reserved.