Economic Scope Of Semiconductor Fabs In India Go Far Beyond Chip Production: IESA Chief

India Electronics and Semiconductor Association (IESA) President and CEO K Krishna Moorthy spoke to BW Businessworld's Rohit Chintapali on the supporting supply chain which could mushroom around a semiconductor fab to cater to the demands of the facility and the global fab ecosystem from India. Defining the true scope of a fab, Moorthy explained how he was confident of India’s potential in training the presently unavailable talent needed (10,000-15,000 people) to run such a facility that could initially rely on skilled workforce from outside the country. He also spoke about Government of India’s Design Linked Incentive (DLI) scheme and how it gives respite to chip design startups in the country, which have traditionally struggled to find a footing in India over the last two decades. Excerpts:

Global leadership in the semiconductor space is of opinion that India could be focusing on IPs - encouraging design innovation - instead of going with investment-heavy manufacturing. They say that an ecosystem for manufacturing would take decades to develop. So, what is your input on that? 

Look, even if the government gives a clearance for a fab today, it will take three to four years for it to reach its full volume production. We are in FY23-24, by the time a fab becomes ready for full volume production with high-yield numbers – it’s going to be FY27-28. 

During this gap, the design ecosystem is not going to sit quiet. Since we have leveraged design capabilities significantly from our past experience and the Design Linked Incentive (DLI) scheme is now kicking-in with three or four startups, the companies which are going to come up as startups or small enterprises in the design space of semiconductors will all start doing IPs. 

Right now, the question is – ‘Do we stop with IP development?’ or, ‘Do we start the manufacturing now and have a fully manufacturing-compliant fab in about four years’ time?’ I would rather think that we should do the IP development, which starts with the DLI. Also, a lot of service companies that are pure-play service companies are now getting into IP development.

A lot on the public medium is said on having fabs in the country and its correlation to solving supply chain issues and meeting domestic demand. But manufacturing chips for domestic consumption doesn’t exactly make sense for profits and business. So, why would they cater to local demand?

It works both ways. No fab will be created in India which will be planned to cater only to domestic demand. It'll not be viable. So, the products coming out of a fab will definitely cater to a global demand. And the moment we have to cater to a global demand, there will be a requirement for a supporting supply chain all around the fab as there are about 200 different chemicals, 150-odd gases, 30-odd metals required to manufacture semiconductors. The semiconductor manufacturing ecosystem itself is complex. So, when the fab comes, this ecosystem is going to develop. Some of them are available today under different quality levels and standards levels. But they are not semiconductor grade. They all will upscale. And once they upscale, they are not going to supply just the Indian fabs. We are going to supply to global fabs where semiconductor manufacturing is happening today, or wherever they are building new fabs. The economic impact of a fab is not just limited to the fab supplying components catering to the domestic demand. And remember, semiconductor manufacturing, much like semiconductor design, also has a huge service component to it and that industry is yet to even start. So, the economic impact of a fab in India is not just limited to producing chips and selling it – it has multiple dimensions to it.

During a panel discussion recently, a MeitY scientist revealed that India may not have adequate pool of talent to handle semiconductor plants and this talent will come from abroad initially. Is this true?

When we talk about talent, even the design talent was not available when we started designing. If you go back to the early days of chip design, the senior management or the senior technologists were expats and it took about 10 years for the Indian design industry to become mature for Indian designers. It took a little while, but it happened. Today, the design ecosystem is well-established. What's going to happen in the manufacturing is probably the same thing but in a different way. For example, today at the operator level – the kind of technicians you need to operate a semiconductor fab, where a very high level of automation exists – we should be able to get these people trained very quickly. I don't see that as an issue. But yes, we will require international expertise for first one or two years at the senior technical level – people who know how to operate the fab, control the processes, improve the yield and  make sure that the safety of the operation is properly established and managed.

Semiconductor manufacturing industry is a highly process-driven industry and there are numerous processes and processes within process of running. Hence, somebody with a very good understanding of that is essential for one to three years. But our dependency is not going to be beyond that. We will need about 10,000-15,000 people to run a fab. I think India can achieve that.

How many fabs can we expect in the near future? What is your estimation?

My expectation for silicon fabs – probably we will have two. There may be a display fab coming up and then maybe one or two compound fabs coming later. Not more than two or three put together. But we will have multiple assembly and test houses. That number may be more at four to six before they require less capital. And they can become very efficient assembly and test lines to meet demands.

Semiconductor demand in 2022-23 was USD 450 billion worldwide. And it will touch USD 600 billion by 2025-26. That's two years from now. It will touch USD 1 trillion by the end of the decade. Now, if these many parts are going to be built, they need to be assembled, tested, qualified, warehoused and dispatched. So, twofold capacity will be required in the next six to seven years. I would expect maybe two or three fabs and maybe five or six assembly test and packaging houses.

Chip design is something that India has already tried over the years but finding backing and funding has been a problem. How do we solve this problem? 

You are right. They have suffered and struggled to find funding. The chip design startups need only two things: one is funding and another is finding venture capitalists who would stay invested. VCs look at how quickly they can make an exit with a twofold to fivefold valuation. With startups developing mobile apps and software, VCs can generally assess this in six months to one year. But with chip design industry VCs need to stay invested with a doubt about whether they will make any money or not.

In this context, early funding coming through DLI via government is a big boon. Of course, we have heard people say that it is taking a lot of time and that it should have been decided long ago. But believe me, the government is doing a very good job because the kind of fund they are putting on the table is to tune of Rs 50 crore ‘per product’ – not ‘per company’. That’s Rs 15 crore for the design proof-of-concept and Rs 35 crore for marketing, sales, productisation and more. So, if you are a startup and you can do two full products end-to-end, you stand to get Rs 100 crore in funding. Naturally, the government will not spend Rs 100 crore of funding without due diligence. It takes time. People have to be patient. But they don't have to go elsewhere searching for funding.

Also Read: Can’t Have Exit-Focused, Liquidity-Driven Cos During India’s Nation-Building Phase: Zoho’s Vembu