‘Chip Pe Charcha,’ an initiative by the VLSI Society of India, aims to unite global leaders in semiconductor and VLSI design. Dr Satya Gupta, President of VSI, engaged in a thought-provoking discussion with Chris Miller, the acclaimed author of the 2022 bestseller ‘Chip War.’ This insightful exchange promises valuable perspectives on chips and semiconductors. Do not miss it.
SG: How has the semiconductor industry changed since your book came out two years ago?
CM: The semiconductor industry has undergone significant changes. First, the surge in artificial intelligence has intensified the demand for AI-related chips, and companies are leveraging AI to enhance design and manufacturing processes. Second, geopolitical tensions, particularly between the US, China, and Taiwan, have fragmented the industry, resulting in distinct markets focused on China and the rest of the world. Third, numerous governments, including those of the US, Europe, Japan, India, and South Korea, are heavily investing in local semiconductor manufacturing, leading to substantial global investment in this sector.
SG: With numerous new announcements for fabs and ATMP factories worldwide, will supply soon outpace demand?
CM: This depends on the type of products. For GPUs, there will be a continued severe shortage due to high demand from AI applications and the limited availability of innovative fabs. Semiconductors for electric vehicles are also in high demand as the industry transitions from combustion engines. However, an oversupply of more common logic semiconductors and utility chips may occur with the influx of new factories. Additionally, political factors are causing market fragmentation, with governments imposing restrictions and tariffs on chips from rival countries, affecting supply chains and creating both opportunities and inefficiencies.
SG: Many fab projects, including those in Ohio and Arizona, are delayed. What are the reasons, and how might this affect investments?
CM: There are two primary reasons for the delays. First, the consumer electronics market, including PCs and smartphones, has been slow for the past year and a half, leading businesses to postpone wafer fab projects. Second, there is a shortage of skilled workers needed for new fab construction and operation, particularly in the US, where the number of new fabs has increased significantly. This shortage of specialised workers and the need for additional training are causing further delays.
SG: Regarding China, with numerous US sanctions on advanced semiconductor manufacturing technology and equipment, can China develop indigenous capabilities to produce sub-3-nanometre chips by 2030?
CM: Catching up in the chip industry is challenging due to its rapid advancement. TSMC currently leads in advanced fab manufacturing, while ASML is ahead in advanced lithography. By 2030, China might develop technologies similar to EUV for producing 3nm and 2nm chips, but their effectiveness in terms of yield and quality remains uncertain. Meanwhile, TSMC and ASML are likely to advance further, making it increasingly difficult for China to catch up.
SG: Can China develop 2nm chips using their own equipment and technology by 2030?
CM: It is plausible that China could develop their own 2nm technology and manufacturing equipment by 2030. SMIC has been approximately five years behind TSMC. If TSMC starts producing 2nm chips in 2025, SMIC might reach 2nm by 2030, provided they maintain their current pace. However, developing new tools and processes will be crucial, and other regions may advance further technologically by then.
SG: If China were to develop a competitive mobile app processor and decided not to use, for instance, Qualcomm chips in their smartphones, what would be the implications?
CM: If China were to design, manufacture, and scale up production of smartphone processors with cutting-edge capabilities, they would need to avoid outdated technology. Currently, it is unlikely that China has the capacity to meet such demands with their existing technology.
SG: What would the impact be if this scenario occurred?
CM: Such a scenario would be significant. Chipmakers who primarily sell to Chinese brands would be particularly concerned. As a result, there is a shift towards moving production outside of China. For example, Apple plans to assemble a quarter of its smartphones in India starting next year. Non-Chinese chip design firms are also interested in expanding their presence outside of China. India, with its large smartphone market, offers opportunities for American, Taiwanese, and Japanese firms to collaborate with Indian companies, thus reducing the market power of Chinese firms.
SG: Considering that approximately 76% of Indian smartphones are Chinese brands, what impact would it have on US semiconductor companies if Indian manufacturers chose MediaTek processors over Qualcomm due to trade tensions?
CM: This is a significant concern. US semiconductor companies like Qualcomm and Broadcom are mindful of these risks. They are working diligently to maintain strong relationships with Chinese firms and the government while also seeking to diversify their supply chains away from China to mitigate potential risks.
SG: Considering the current geopolitical situation, could India function as a strategic advantage for US semiconductor companies? For instance, if a company like Qualcomm collaborates with Indian firms to create an Indian brand of mobile phones using their processors, potentially capturing 10%-20% of the market share from Chinese brands, what is the feasibility of such a collaboration?
CM: This scenario is increasingly feasible due to growing geopolitical pressures and investments in the electronics ecosystem. US companies have a strong incentive to support the development of alternative products and brands in India. There is no reason an Indian-based smartphone company could not produce a world-class product, sourcing components from the US, Taiwan, Japan, and India, and compete effectively against Chinese brands. This opportunity extends beyond smartphones to other products where Chinese brands are also vulnerable, making it plausible to build competitive Indian products with the assistance of US semiconductor companies.
SG: China is heavily investing in compound semiconductors like GaN and silicon carbide, which are valuable for EVs and other power electronics applications. Given that compound semiconductors do not require advanced manufacturing technology, can China dominate this sector?
CM: Chinese companies and the Chinese government view compound semiconductors as a significant competitive advantage, with substantial investment driven by the growing EV market. China is likely to see increased protectionism and tariffs, which could provide guaranteed markets for Chinese firms, even if their quality does not match that of European and Japanese counterparts.
SG: China is focusing on lower-end chips like microcontrollers and Wi-Fi, producing them at lower costs using older technology. Could this lead to the US and others concentrating on high-end chips while China dominates the low-end market?
CM: Without intervention from policymakers in the US and other countries, China could indeed dominate the lower-end chip market. US, Japanese, and European leaders are concerned about China’s subsidies impacting Western firms’ profits and may impose more tariffs and restrictions to limit the presence of Chinese chips in global markets. This could result in complex geopolitical and regulatory battles.
SG: Companies such as Google, Amazon, and Microsoft are developing their own custom chips. Could this lead to a transformative shift in the industry similar to the fabless model of the 1980s?
CM: The key challenges for custom chips are cost and complexity. Currently, only major tech companies with substantial resources can afford to develop them. To make custom chips more accessible to other industries, such as automotive and medical, we need cheaper, simpler design processes. AI-driven design software could help reduce costs and simplify development. Additionally, chiplet-based architectures, which involve assembling pre-designed components, could be an easier initial step for non-tech companies. However, further simplification and better design tools will be necessary for broader industry adoption.
SG: India introduced a comprehensive semiconductor policy around the same time your book was released. How do you assess the progress made in the past two and a half years?
CM: Considerable progress has been made under the India Semiconductor Mission. Major companies like Micron have started establishing facilities, and there have been announcements for OSAT and wafer fabrication facilities from Tata and others. Electronics product assembly in India is also expanding, with Apple planning to assemble 25% of its phones in India, contributing to the overall manufacturing ecosystem. Additionally, India’s strength in chip design remains notable, and semiconductor product design should continue to be a key focus.
SG: Some believe India is not advancing as quickly as possible. Considering India started with minimal semiconductor manufacturing, what are your thoughts?
CM: Advancing in chip manufacturing and assembly is a challenging, capital-intensive process that requires several years to complete a single project. Given this context, progress has been solid. It is crucial not to rush the process but to develop a comprehensive ecosystem to support long-term growth.
SG: By 2030, is it realistic for India to have two semiconductor wafer fabs, two compound semiconductor fabs, and two OSAT plants?
CM: This is very realistic. It is certain that India will have at least two silicon fabs, although the specific types of semiconductors, technology nodes, and capacities will need to be clarified. The establishment of two OSAT/ATMP plants and two compound semiconductor fabs is also feasible.
SG: You often mention that large Capex manufacturing projects are like ribbon-cutting opportunities and that India should focus on design. What should be the focus and strategy for India to become a ‘semiconductor product nation’ rather than just a provider of design services?
CM: In the semiconductor industry, most financial gains accrue to companies producing semiconductor products, not merely providing design services. The challenge for India is to create brands and companies headquartered in India that sell globally. This has been achieved in the software sector, and there is no reason it cannot be replicated in semiconductors. India’s venture capital ecosystem needs to become more hardware focused. Talent exists; it is about connecting it with business expertise and financing, and there are encouraging signs in this direction.
SG: The Indian diaspora has established many successful semiconductor companies in Silicon Valley. How can collaboration between Silicon Valley and India be improved to make India a ‘semiconductor product nation’?
CM: Venture funding for hardware in India needs to increase and extend over longer periods, similar to Silicon Valley. While software ventures receive more VC attention, hardware projects require support over a five-year or longer horizon to design and launch products. For every ten software companies funded, at least one hardware company should receive support. Indian companies can also leverage the large domestic market to test and refine products before global expansion. This approach has not yet been fully utilised but offers a promising model for Indian firms.
SG: India is expected to leverage its nearly $200 billion domestic electronics market. However, around 90% of this market is served by global product companies that decide on chip types outside India. How can India grow its electronics product companies and brands to create demand for local chips?
CM: The success of non-US chip design companies like MediaTek and Huawei illustrates the potential of focusing on electronics system-level products. Huawei grew its chip design through its telecom and smartphone systems business, while MediaTek integrated itself into the global smartphone ecosystem. India should focus on emerging electronics products such as EVs and transportation to create leverage for domestic chip companies.
SG: If India aims to create one or two companies like MediaTek, what should be the strategy and focus areas?
CM: MediaTek’s success stemmed from timing its investment in smartphones. For India to replicate this model, it should target different markets such as EVs and medical devices, leveraging its unique and large domestic market in transportation and healthcare. The application of AI to healthcare, wearable monitoring devices, and innovative medical equipment presents significant opportunities for semiconductor innovation.
SG: Lastly, after semiconductors and AI, what could be the next global battleground where countries will seek supremacy, and would you consider authoring a book on it?
CM: Although I have not started a new book yet, I am interested in biotechnology and genetic engineering, particularly with the application of AI. Major innovations in these areas seem imminent. In 10-15 years, biotechnology and genetics could advance as significantly as chip technology does today.
This interview was transcribed by Nidhi Agarwal, who is a journalist at EFY. She is an Electronics and Communication Engineer with over five years of academic experience. Her expertise lies in working with development boards and IoT cloud. She enjoys writing as it enables her to share her knowledge and insights related to electronics, with like-minded techies.
