- Parag Naik, CEO, Saankhya Labs is of view that setting up a semiconductor fab right now is the wrong strategy
- He noted that investing in fabless design companies makes more sense as more companies can be nurtured with that approach
Parag Naik, CEO, Saankhya Labs, in an exclusive conversation shared his thoughts about the burning topic ‘to fab or not to fab in India’. He is of the view that unless a local ecosystem to support an India-based semiconductor fab exists, the very existence of the fab will always loom in uncertainty.
With this interaction, EFY has been able to unearth ideas of two leaders about building a fab in India. The first in the series was by the CEO and founder of Vann Consulting – Mr PVG Menon. He is of a strong view that a semiconductor fab should be built in India. (You can read the interaction here)
Saankhya Labs is India’s first fabless semiconductor company. It has business ties with ISRO, Indian Railways and tier-1 companies from the USA. Saankhya Labs outsources the manufacturing of its products to TSMC and Samsung.
Excerpts from the interaction
EB – What is Saankhya Labs all about? How has been the experience so far?
Ans – We started our journey in India about 12 years back. Saankhya Labs is a company that builds semiconductors. When we say we build semiconductors we do not refer to building stacks but building chips. We outsource manufacturing to foreign companies. Our chips are manufactured at facilities owned and operated by TSMC and Samsung.
Our specialization lies in the software-defined radio vertical. We were among the firsts in the world to talk about building software-defined networks and radios.
Today we have evolved from just a semiconductor focus to an end-to-end wireless communication solutions company. We have developed solutions for 5G, Broadcast and Satellite Communication based on our SDR Chipsets. We have embarked on the journey of making cutting edge 5G Radio Access Network solutions for the global market. We have already started working on the design and development of 5G Radio products and have also started investing in some future-looking 5G product offerings. We are focusing on building a 5G Radio Access solution based on Open RAN. We are in discussions with some US-based customers and partners to take our product into the market especially for the US market.
I feel that building a semiconductor company in India is a fool’s errand. From raising finances initially to government policies, we have had hiccups at all the stages. For manufacturing chips, you must pay money to the foundry. There was a point in time when the RBI thought that we were running a Ponzi scheme! I think the doubts were a result of us paying millions to our manufacturing partners. In return, we were only getting software files.
EB – Where are your clients located?
Ans – Saankhya Labs is a fabless semiconductor company. That simply means that we own the designs. Saankhya holds 65 different patents. We are technology partners with the Indian Space Research Organisation (ISRO). Indian Railways is one of our customers. We have also developed Satellite-Based Vessel Tracking Terminals for fisheries
We are also working with tier-1 operators around the world. Most of these are in the USA and we are also in talks with some in India. As a matter of fact, we are India’s first fabless semiconductor company.
EB – The debate of having a private sector semiconductor fab has caught fire again. What are your views on the same?
Ans – We have been also seeing major interest from a lot of people advocating to build fabs in India. Building a fab might not be the right strategy now as it requires investment in the range of US$ 30 to 40 billion!
We should always keep in consideration that several governments have tried the same and have given up. In fact, I can remember the same talks originating in as early as 2006.
EB – If having a fab is the wrong approach, what is the right one?
Ans – We must understand that the semiconductor supply chain is global. You cannot do everything in India. The equipment for making semiconductors comes from a couple of companies located in the USA and Netherlands. The sand and other materials come from different parts of the world. The technology needed comes from countries like Taiwan and the USA. Even the USA is lacking in that arena.
The right time to talk about this approach was around 25 years ago. We should rather be concentrating on design skills in India. This is one domain where our expertise lies.
EB – Can you elaborate more, maybe via an example?
Ans – Absolutely, let us look at the example of Apple. The company does not manufacture anything, but it owns the intellectual property rights (IPRs) for all the products. For Apple, everything gets done by Foxconn. But who gets most of the money, Apple or Foxconn?
It’s Apple! Then who gets the most valuation? It is Apple again! Coming back to the semiconductors, I see fabs at the lowest levels of the food chain. What we need is more fabless semiconductor design companies in India. While we stress on made in India, we should also explore the possibility of made in India!
To explain more let’s do a simple calculation. Consider you have US$ 40 billion ready to invest in a fab, why not invest parts of these in 5000 different fabless semiconductor companies?
With this approach, we will start to see companies building chipsets for products like laptops, set-top boxes, mobiles, TVs and whatnot. Once a critical mass of companies originating in India start doing so, then you go and build a fab in India.
Starting with a fab is like trying to reach for a solution from the wrong end. Even if we invest in a fab, where are the companies that will use that fab?
EB – Made by India rather than made in India, what will that look like?
Ans – I have been arguing more and more about why India needs to build more and more companies that want to build IPRs. The reasons include generating and adding value and creating new jobs. Once you have complete control over manufacturing, you can easily choose to go to any of the suppliers.
For example – Apple is moving its manufacturing via Foxconn from China to India. That is how it will be possible for India to have a hand in every chip that is used in the world. IF we look at Bangalore alone, there is a whole of design engineers available there ready to work on the design part.
Local players who can build end products would be the right approach now. We should try to nurture these in India first. Once there is a strong network of these, the fabs, and not just one, then they will come on their own.
Que – How about the comparison in terms of jobs created by fab vs fabless companies?
Ans – Fabless design companies can create 10x more jobs in comparison to what a semiconductor fab can do. That too with lesser investments in comparison to a fab! If you look at the larger side, fabless companies create more jobs in the complete ecosystem.
You need people working on the mechanical side, people of the software side, people on the testing side, on the packaging side and more. I second the thought that eventually we should have a fab built in India, but now the concentration should be building local companies that are going to utilise this fab in future!
If you also look at how Taiwan was built, you will find an interesting approach that they followed. TSMC, the largest semiconductor fab, was built by a guy named Morris Chang after retiring from Texas Instruments in 1986. But before TSMC came into existence, there were tens of companies in Taiwan that were already manufacturing things like motherboards and whatnot. I think Cheng was completely aware of the existence of ecosystems and players required to make a fab successful. Do we have any such ecosystem here?
You cannot plant a tree in the middle of a desert and hope that the ecosystem to support it will come up on its own. It takes strategic thinking. To top that, how many private players do you think are ready to invest US$ 30 to 40 billion?
EB – The government had recently announced schemes worth Rs 50,000 crores for electronics. Would these help in creating an ecosystem?
Ans – These schemes look crafted for big foreign players such as Apple and Samsung. However, to complement a fab in India, we need a local ecosystem and local players. As I earlier said, once you have control over manufacturing you can choose the location as per your needs.
Additionally, just saying things out loud will not do it. With the policies, you need to have the right intentions as well. You must start building from the ground level up and not the other way around. I have completely given up on Delhi (central government). I think five years down the line we will be having similar conversations. I have been in touch with different governments over the last ten years.
The only difference between the current government and the last government I found was in terms of approachability. The current government is more approachable than the last one, but nothing has been materialised on that front as well. These grand schemes, in my view, are announced for the media and nothing on the ground takes place. I do not think any Indian semiconductor company has got any benefit from any policy announced by the government.
At the end of the day, they must understand that it is not just about creating funds; the intentions to work matter more. I do not think they have allocated a single penny out of those big schemes so far!
EB – What approach would you recommend then?
Ans – I am not at all against the government trying to bring in global players with a red-carpet welcome. But where is the intention to develop players in India? Mark my words that a semiconductor fab will be supported by local players first and only then the MNCs will follow. Semiconductors is a long-term and long-strategy business and it takes a longer time to be fruitful.
What’s better – pumping some billions of dollars into a fab or investing half of that amount into a handful of fabless semiconductor companies? If the only fab you are trying to accomplish fails, every penny associated with it goes into a dump, whereas if you invest in some 50 fabless semiconductor companies and even three of them become successful, all the money invested gets bang for the buck.
EB – Talking about support from local players, are you doing business with any local companies?
Ans – The products that we built are hardly used by any company in India. Some of them have started to do it now and we are already in talks with them. We managed to convert ISRO into one of our clients after four to five years of follow-ups with them. We are also doing business with the Indian Railways.
To be blunt, no one in India believed until very recently that semiconductor designs can be done in India. But during the last one year and especially in the last few weeks, after the Galwan incident, people who are associated with electronics in one form or the other have started saying that it can be done.
The ideal approach will include access to money and time. You need billions of dollars to develop products in the semiconductor arena and a lot of time. The third thing is local as well as international customers willing to take risks to try out the new products. If you want to build an Intel out of India, you will need to invest into at least 50 companies and track all of them. It is also natural that only two or three will be able to go up to that level.
If we look at how China built the mobile ecosystem, we might be able to find a lot of answers. About three-four years back nobody knew about the likes of Oppos and Vivos. Now, these companies are ruling the smartphone sales charts all over the world. If you go to big corporate houses in India and talk to them about the same, most will say it is too risky and not possible in India.
You also need financial institutions to understand the value of semiconductors and how semiconductor companies work. What we have in India are engineers who can work on the design part. Lawyers who can write legal stuff and even marketing guys who know how semiconductors as products are advertised are not here.
Finally, I do not know of any bank or a financial institution here that values IPR. If you step into any banks of Taiwan or Silicon Valley, you will be amazed by the kind of knowledge they have about the IPR and its importance.
When the IPR part comes to the semiconductors, forget IPR, banks here do not even know what semiconductors are! If I were to step into a financial institution for getting Saankhya Labs evaluated, all they would do is count the number of machines (laptops, computers, printers, etc.) and servers and give us a value!
Story by Mukul Yudhveer Singh
Do not miss