Chandramouli Soorian
Executive Summary
- Technology war between US & China and export restrictions
- India projected as the new manufacturing hub
- Chips are the new oil
- China is only 5 years away from cutting edge
- Risk of China making a breakthrough
- India as a source of competitive R&D labour force
- Complementary to the allies in competing with China
- India as the next big semi market
- Non fab opportunities – design, R&D, Innovation, component & chemical supplies
Short window of opportunity
Like many other industries, the willingness of China to bankroll and subsidise its semiconductor ambitions on a large scale has doomed many a Fab project across the world. This brief technology war opens up a world of opportunity for India. Even if it does not materialize into a fab in India, India can still capture value from its allies.
Apart from the technology war, the transition to a post covid economy has created a great demand for semiconductors that exceeds current supplies. Investment in industry 4.0, 5G, AI, Robotics, autonomous vehicles, greater adoption of Big data has already created a shortage in semiconductors. Currently Nvidia & AMD have not been able to meet the demand for graphic cards, and the same is the case for other consumer electronics but are less visible because they do not deal with the leading edge.
A third opportunity is semiconductor being the new software revolution. As many of you must have observed in the last few years, the clock frequencies of our CPUs have stagnated (in many cases slowed down) and we have become used to multicore. What this meant for software is that it doesn’t necessarily run faster with newer processors.
As we go to lower node sizes, 5nm and 3nm the cost of chips are getting greater, new challenges arise, new designs and innovation are required to overcome this. Also with IoT and ML, custom accelerators, interconnects and packaging innovation are required. We are at the cusp of an explosion of new innovation opportunities in semiconductors, which were thus far restricted to established players. This provides a short window for India to enter into the semiconductor value chain.
Cutting edge Fab
TSMC (Taiwan) has proposed to build a limited scale 5nm (currently cutting edge in production) in the US, it would be operational in 2024. The US has enormous preexisting talent and experience with semiconductor design and manufacturing (Intel). In fact, 50% of the value of cutting edge equipment and design software are produced in the US.
It is unlikely that India would be able to get a cutting edge fab that is protected by the parent country. However, cutting edge nodes provide the best profit margins, so it is profitable to build one. India has several compelling reasons to be an attractive destination, if we are willing to go all in.
- An emerging electronics manufacturing hub
- A large consumer market
- A large and stable democracy, willing to ensure availability of critical supplies
- Provide a first mover advantage
- Willing to provide a semiconductor city with a concentration of critical talent
- Leverage its scale in metallurgy, chemical, manufacturing, energy & water to provide competitive input materials
- Build a large fabless capacity and customers who can benefit from proximity
- Build a large pool of suppliers and their R&D that can work in close proximity
Not so cutting edge fab
India currently has a lone fab of 180nm, any node closer to the current state of the art can still prepare us for the future and not leave us behind. Even older nodes in capacity will help us with value add to 90% of the electronic goods produced within the country. The age of electric cars, solar energy requires power electronics which does not require cutting edge nodes.
Though the margins here are slimmer, the huge captive market capacity would be a compelling reason to invest in building one (or more) fabs. This can be a stepping stone to upgrade to better technology nodes, building a competitive supply chain and an innovation ecosystem that can export globally.
No fab
Even if no fabs are built, the government can build an ecosystem by building a “microcity” with this specialisation. Several countries notably Sweden use this model to build a critical mass of expertise in a city (telecom city).
The way to go about it is to build one or more microcities away from expensive cities, around a university or college campus dedicated to research in this area. Polytechnics, training schools, middle & higher secondary schools with a specialisation in the area of semiconductors, VLSI, Analog design or Fabrication related skills. It is crucial to develop a pipeline of talent from school, to college and onto research to build a sustainable & competitive ecosystem.
Inviting multinationals, Indian companies & startups to set up shop here, once a critical mass is created, it will automatically draw more talent, companies and investment.
Why semiconductors & why now
As China has realised semiconductors are the new oil and is mostly imported. As India aims to become a global electronics manufacturing hub, it too will depend on expensive semiconductor imports. The limits to India’s growth will depend on the import inputs.
For instance, to undertake a massive road and bridge infrastructure projects, the government can take a huge loan to fund the projects. After all it will end up back in the government’s coffers as long as the money spent productively & stays inside the country. However, at some point in time it will be used to import crude, which is the basis of tar and fuel for construction machinery.
Semiconductors unlike oil can be produced in the country, there is no intrinsic reason it can not be. At the moment like software India also has a low end semiconductor ecosystem, mostly engaged in low value addition like testing and verification. It would take at least a decade to build a mature ecosystem that is globally competitive.
Why now & will the rest of the world collaborate
We are at this crucial juncture where improvements in computing would require innovation in several areas. Custom accelerators, Domain specific chips, device innovation, quantum, neural networks, architecture and hybrid systems.
The emergence of China as a global competitor is quite worrying, this poses a strategic threat. As with many other industries, if China makes a breakthrough, it would monopolise this crucial industry. Unlike oil which can be substituted with other sources or forms of energy, semiconductors are the very basis of our modern economy and has no substitutes. He who dominates this sector will have an unfair advantage in every other sector.
The west missed a very good advantage it had in the 5G space, the telecom sector especially the wireless standards has been dominated in the past by western companies like Ericsson. By failing to partner with an ally like India which has the capacity to supply (cost) competitive engineers and scientists, the west lost their edge to the Chinese.
Semiconductors are national interests projects, where the countries often hold substantial stake in these industries. And it can thus be a very protective domain. If the dominant players are willing to open up their ecosystem to India, India has the capacity to produce the pipeline of talent that is competitive with China.
Unlike China it is not India’s goal to dominate semiconductors but to be a significant player and not to be left behind.
Chandramouli Soorian is an engineer by education and a software developer/data engineer by profession. Apart from growing up in India, he has also lived in Sweden, Germany & currently in Denmark. He would like to share his insights in the areas, (View More)