When I spoke with the Synaptics CEO in 2022, he said that the semiconductor manufacturing ship has sailed already, even for a country like the US as TSMC’s expertise would be hard to replicate due to various reasons at the cutting edge. It would be better to help Taiwan rather than compete. You yourself have said that the US has no intention of becoming like Taiwan either. But has the ship sailed for India as well?
It takes a long time to build a semiconductor manufacturing ecosystem. If you look at how long it took Taiwan to go from zero to being the world's leader – it took almost half a century. To say that it is too late for India would be unfair (I do not know where the world is going to be in half a century). It is impossible to build a semiconductor ecosystem very quickly as it takes a lot of time, money, and training. But I would be surprised if there aren't a couple of new chip factories in India in a couple of years because there is interest from the private sector coupled with government incentives. Currently, I see lot of interest coming in at the assembly and tests level in India and there will be lot of focus on design. This will be a long process, but on a lot of the key early metrics, India is doing relatively well.
Does this imply that manufacturing the most advanced chips is currently beyond India's reach, potentially leading to a focus on lower-end nodes and direct competition with China (which dominates this segment)?
Yes, pursuing the lower-end nodes is part of the natural progression when you are exploring manufacturing semiconductors.
During a panel discussion, you said that "Taking China head-on isn't the best strategy." Could you elaborate?
To succeed in manufacturing, it is essential to target specific markets and competitors. Targeting the global market and competing with China on cost is challenging. The substantial investments China is making in chip manufacturing pose a risk of overcapacity, making it difficult to profit from less technologically advanced nodes. Investing in such areas simultaneously with China might not be advisable.
A more prudent approach involves identifying underserved niches in manufacturing and recognising areas with significant growth potential, such as power semiconductors for electric vehicles (EVs). For instance, despite the current scarcity, the demand for power semiconductors is expected to surge with the increasing prevalence of electric vehicles in the next 10 to 20 years. While government incentives are essential, a critical evaluation of the business model is equally important to ensure long-term success.
What is your take on the idea that, despite considerable semiconductor research in India, a prevailing notion exists that published papers tend to fade away and don’t manifest into bigger things? A noted BITS-Pilani professor suggests a foundry could help India and prevent this. Your thoughts?
There are different types of foundries. TSMC produces both the most advanced smartphone chips and it produces lots of less advanced chips with huge cost differences between them. If you want to make a foundry for the most cutting-edge chips, your price tag starts at about USD 20 billion. That is a lot of money. But for less advanced chips, you can do it for much less. Is there merit in establishing a foundry with a focus on less advanced, cost-effective experimentation rather than the most advanced and expensive technology? Possibly. Other governments are considering that too. If you look at what the US is doing to catalyse more prototyping, more experimentation, it is trying hard to make sure that researchers and startups have more access to fabs early on.
A lot of experts, especially from the West, have told me that India must build on its design expertise – get better at something it already has considerable human capital and expertise in. And Michael Hurlston (Synaptics CEO), during my 2022 interview, said that the focus on IP is the direction to take for India. Is manufacturing not the optimal route, or should there be a balanced approach?
To some extent, design and manufacturing must happen in tandem. Just building one firm or one factory alone is not enough. It is crucial to build a comprehensive manufacturing ecosystem. I do worry that political leaders, not just in India, prefer and like the idea of something getting built successfully. But history suggests, that is not how things work in the semiconductor space. If you build an ecosystem, it can sustain itself but you do not want to be betting on just one.