--- Created On: 2022-10-25 10:15 Creator: Nitin Pai publish: true --- --- Status: [[Brewing]] ## 2022-10-21 1. The latest US offensive in the ongoing US-China geopolitical contest has thrown the semiconductor industry into a state of flux 1. It is aimed at containing China's geopolitical power by blocking the leading edge of power: advanced semiconductors; and this will slow down everything from new consumer devices to AI to weapons systems 2. China will be severely affected in the short-, medium-term as it lacks immediate options; and constraints in medium-term options. Money cannot buy many of the ingredients. 3. US, Western and other tech companies will suffer in the short-term; loss of market, supply chain disruption, lower economic activity 4. All semiconductor firms in major economies are likely to benefit from favourable government policy; this is globally sub-optimal in the short-term but still not so bad. 5. It is quite possible that there will be two or more technology eco-systems, from silicon to content, over the next ten years. 2. China's options: 1. Indigenize 1. Extremely hard to do at the leading edge given industry economics 2. Potential to innovate around (disruptive) in the medium-term 3. Will be behind the West for 10 years and a few industry cycles 4. While this may allow China to satisfy its domestic requirements with "import substituted" products, it may not be able to compete in export markets in high-end and high-value products 2. Break through sanctions through third party engagements 1. Germany, Israel and some ASEAN countries might provide loopholes; depends on how much pressure the US is able to exert on these allies; and on Washington's resolve to keep the pressure. 2. Situation is more or less like Indigenize 3. Take over Taiwan: sanctions being a trigger pretext; Cff [[The questions of Taiwan]] 1. Taking over Taiwanese facilities alone, without management, people and business arrangements will give China the factories, and allow it perhaps to shorten the indigenization cycle, while causing disruptions for the rest of the world. 2. TSMC and other vendors are building out in other countries to hedge against this; 5 year timeframe to build fabs. Such locations will attract the managerial & industrial talent from Taiwan. 3. Aggression of this scale will trigger a different level of geopolitical, geoeconomic and military response that might force China to confront the West without the chips (which will be available only in the future); fabs can be sabotaged etc. 4. Compromise 1. Unless China creates more pain for Washington, it will be handing out the olive branch, and will thus face unfavourable terms 2. What is it that China can do to coerce Washington to negotiate? Cyberwar, Taiwan? The path back to cooperation is unclear. 3. A compromise that allows China access to advanced technology will require massive political-economic changes in either Washington or Beijing. 4. Even under such a compromise, global supply chain de-risking will continue 3. Semiconductor sanctions on China and broader sanctions on Russia will affect their arms industries 1. China 1. Impairment of artificial intelligence and supercomputing capability has downstream effects on advanced military technologies such as missiles, drones, autonomous systems, C4ISR 2. Russia 1. Faces shortage of chips and subsystems; with China left as the primary supplier. Unclear to what extent they can collaborate to circumvent sanctions. 2. Russian companies will heavily rely on Chinese semiconductors, sensors and sub-systems. At the top end, these will be a couple of cycles behind Western equipment. 4. India has a window of opportunity of around 5-10 years to become a significant player in the global semiconductor industry. This will generate jobs, growth and global influence. It will also provide opportunities to cover some techno-strategic vulnerabilities. 1. A basic ability to manufacture semiconductors is necessary in an uncertain world. But for marginal benefit = marginal cost, overstretch should be avoided. Union and state governments intend to spend $10 billion or so on fabs. 2. Even for basic ability, it is abundantly clear that the key success factor is to be plugged into the US-centred ecosystem within [[Bubbles of Trust]]/'friendshoring'. Or within the Chinese ecosystem, but that's not a serious option for us. 3. Investing in human talent is the best strategy: they can bolster the domestic production; or work in industries abroad. 4. The Second-ORR [^1] option, where India becomes the world's semiconductor services hub just like in software services 1. convert capex to opex; fixed cost to variable costs; The Morris Chang model applied to design services. 2. [^1]: A metaphorical or literal Second Outer Ring Road (ORR) of Bangalore. The first ORR belt is a global hub of IT services companies; the second ORR can be the same for seminconductor services. 5. Government investment in high technology industrial institutes that cover everything from design, production and management of semiconductors; we can build 50 such institutes for the price of a wafer fab 1. Rs 2000 crore capex, Rs 100 crore/year opex; may be an opportunity to reform some of our public universities beyond the IITs & IIITs 5. The greatest success has been when government funding has either funded an upstream scientific breakthrough or backed an industry-changing concept. Merely bringing production home through fiscal incentives has had modest success, and quite a number of failures. 6. The Intersection column 2022-10-25 based on this assessment [argues that India must double down on manpower investments](https://www.nitinpai.in/2022/10/24/manpower-as-india-s-siliconpolitik-strategy) to succeed in the siliconpolitik game.