297 / January 28, 2025
Nvidia’s Success, Chip Race, India’s Semiconductor Mission, & Hardware Vs Software | Raja Manickam
How Semiconductor Runs The World
In this episode, we take a deep dive into the fascinating history of semiconductors, their evolution over the years, the rise of old giants and new disruptors, and explore why India has yet to produce a leader in one of the most important technological races of all time.
Raja Manickam, a serial entrepreneur in the semiconductor industry, shares from his decades of experience about the business and science of semiconductors and offers a glimpse into the future.
Watch all other episodes on The Neon Podcast – Neon
Or view it on our YouTube Channel at The Neon Show – YouTube
Siddhartha Ahluwalia 1:20
Hi, this is Siddhartha Ahluwalia, your host today at Neon Show and co-founder of Neon Fund, a fund that invests in the most promising enterprise AI software solutions coming out of India and built for the globe. Today, we’ll discuss about the history of semiconductors in India and what really is all the gung-ho-ness in the government about building semiconductors in India competing with the world. But today I have with me Raja Manickam.
Welcome, Raja, to the podcast.
Raja Manickam 1:47
Thank you.
Siddhartha Ahluwalia 1:47
Raja, you have built two great companies in semiconductor space.
The one you are representing right now, IVP Semiconductors, and the previous one, Tissol, that was acquired by the Hero Group. So, big congratulations on your success.
Raja Manickam 1:59
Thank you.
Siddhartha Ahluwalia 1:59
And you are the very few entrepreneurs that have made hardware in India, succeeded in it, put India on the map for hardware.
Raja Manickam 2:07
Thank you. Thank you.
Siddhartha Ahluwalia 2:08
So, what I want to discuss is because our audience, you know, some of them might not know about semiconductors. So, let’s discuss about the history globally of semiconductors, and then we’ll come to the history of semiconductors in India.
Raja Manickam 2:20
Sure, sure. Even I think Sundar Pichai said that hardware is hard. Okay, so it’s very unique in that sense, right?
It is hard because it involves a lot of multiple players that you’ll have to interact with versus, you know, software, right? You pretty much, you can create your own software package and then, so the engagements that you’ll need to have within the ecosystem players, in multiple locations, geographies, make it very difficult. So, just to, you know, let you give you a little bit of a history of semiconductors in my own journey.
So, I started at Texas Instruments as a test engineer after my graduation.
Siddhartha Ahluwalia 3:06
In which year?
Raja Manickam 3:08
1982,I am from IIT Kharagpur, ECE. Those days, you know, the moment you graduate, you look for going outside of the country.
There’s not too many opportunities here. But, you know, so fortunately for me, one thing that happened was I actually started working in a factory, right? Most people, you know, end up going into a design or something like that.
So, I had a lot of practical exposure, right? And then later on, with Texas Instruments, and I went to Fairchild. Fairchild, as you know, is…
Siddhartha Ahluwalia 3:41
All in India or in the US? No, no.
Raja Manickam 3:43
I was in Singapore and then I was in the US. So, most of my life, my career was in the US. I was in the East Coast with Fairchild.
Fairchild, as you know, is the mother of semiconductors, right? And then I went to the West Coast in Silicon Valley. Fairchild became national semiconductor.
Siddhartha Ahluwalia 4:01
In 82, you moved out of India to Singapore?
Raja Manickam 4:04
Yes.
Siddhartha Ahluwalia 4:04
And then which year you moved to the US?
Raja Manickam 4:07
88.
Siddhartha Ahluwalia 4:08
You moved to Fairchild.
Raja Manickam 4:10
Boston, yeah. Fairchild, Portland, Maine, actually. And then I was there for a few years.
And during that time, I was also sent to Malaysia to run the factory, Fairchild factory. I was running the engineering department for them. And then I came back again to the US.
Then I was based in Silicon Valley.
Siddhartha Ahluwalia 4:30
And which company? Fairchild?
Raja Manickam 4:32
Fairchild got acquired when I was there, acquired by National Semiconductor. So, I became part of National Semiconductor. And during that time, I was actually pretty good.
I ran the telecom business unit. So, I got into the business side of things. And at that time, I had the fortune of coming to that. India was liberalizing telecom and all that in the mid-90s. So, we had a lot of chips that we were trying to sell. So, I actually came to India to work with some of the CDOT engineers in Delhi and Bangalore, very young engineers.
And so, I used to work with them. So, that was also a good experience for me. And then I was back in Silicon Valley.
I was in some few startups there, but nothing became big. So, then I started my own company in 2004.
Siddhartha Ahluwalia 5:30
In the Valley or in India?
Raja Manickam 5:31
2003, I started Tessolve in the US. In 2004, really took off in India. And really, it was an India-based company.
I wanted to do manufacturing. Believe it or not, in 2004, I had all these manufacturing plans that are being funded by the government today. We had that plan in 2004.
But the country was not ready at that time. So, I ended up doing engineering services, which was Tessolve, very, very successful. We created probably more than 10,000 engineers that have gone through our system.
And they’re all over the world. I’m very proud to say that you will find a Tessolve engineer in almost every chip company in the world. And then after I sold it to the Hero Group, I was running it for a while, for four or five years.
And then I joined the Tata Group and Tata wanted to get into semiconductor. I was there for a couple of years. And then I said, hey, time to build a chip company out of India, right?
You know, we have a lot of chip designers, but we don’t have a chip company. And I kind of figured out why and so forth. And I believe that the timing was good.
Siddhartha Ahluwalia 6:51
Sure.
Raja Manickam 6:51
That’s how IVP was born.
Siddhartha Ahluwalia 6:53
And which year IVP was born?
Raja Manickam 6:55
Last year.
Siddhartha Ahluwalia 6:56
Okay.
Raja Manickam 6:57
2024.
Siddhartha Ahluwalia 7:00
This is your history as an entrepreneur. Let’s discuss the history of semiconductors.
Raja Manickam 7:05
Okay. So semiconductor originated like Fairchild and TI. And then the Fairchild guys created Intel and kind of mushroomed out, right?
And so semiconductor was a very innovative, you know, the transistor was the basis for it, right? Where you could make smaller and smaller transistors. And that is how, you know, the industry flourished.
And semiconductor was very vertically integrated. So when I was in TI, we had a design. We had our design tools.
We had made our own design tools. We had our own factories. We had our own fabs and backend packaging.
We had, you know, even making equipment for this. All were all done vertically integrated. So if you didn’t have all of these things, you could not be a semiconductor company, right?
So it was, you know, very few people were involved because, you know, you needed all of this to make a chip, right? Now, TI was the first company that set up shop in India. Believe it or not, to make design software tools.
And the reason was also because the CTO of TI at that time, you know, I’m talking about the mid 80s, was an Indian guy. And he was going towards, you know, to his retirement and say, you know, I need to start. But anyway, so that’s how TI started.
TI was the first company set up in Bangalore to do, you know… So fortunately for me, my first job was TI, right? So I learned a lot from TI on, you know, how they were doing this.
Siddhartha Ahluwalia 8:50
And semiconductor means chips or what does semiconductor mean?
Raja Manickam 8:52
Yeah, so semiconductor by the name, right? You know, when you say conductor, right? Conductor is like…
Siddhartha Ahluwalia 9:02
Conducts electricity.
Raja Manickam 9:02
Conducts electricity, right? So if you put a copper wire or aluminum wire or something, right?
You know, so you can send electricity to one. Semi is where you can control it, right? So you dope a certain material and then you can control how much, you know, and when you want to switch and when you want to switch it off, right?
So controlled, that’s what a transistor is, right? It’s a transistor you can switch on and off and bias it, right? So as you want it.
So I think that’s the difference, right? So that’s why it’s called semiconductor. And it’s a lot of, you know, the process uses silicon as a base.
But there’s a lot of doping of different chemicals and so forth.
Siddhartha Ahluwalia 9:47
And that’s why Silicon Valley is called Silicon Valley, because they’re the semiconductor’s originator.
Raja Manickam 9:51
Correct, correct, yeah. So that’s where the name silicon comes, right? Because the base material is silicon.
So, but I think what happened was in the 80s and probably the 90s is where the pieces broke out, right? So you could, you know, like, for example, TI sold the equipment group out. So they started companies which were just doing equipments, right?
And then, you know, actually the TI guy, Morris Chang, is the one who started TSMC. And he also was a fab manager in TI. And he went off to Taiwan to start TSMC, which is, you know, everybody knows TSMC is the world’s largest foundry for wafers, right?
Siddhartha Ahluwalia 10:41
Taiwan Semiconductor Manufacturing Company.
Raja Manickam 10:43
Yeah, it was started by, you know, the TI guy, right? And, and so, you know, so pieces of it became available. So now, I think probably in the mid, early 90s and mid 90s, so chip startups started coming because now they could go and outsource, you know, they could get the fab done somewhere.
They could do the packaging, there was companies like ASE and Amcor. So the entire food chain, right, of semiconductor, which was vertically integrated, became more horizontal.
Siddhartha Ahluwalia 11:16
And what’s the difference between a chip and a semiconductor?
Raja Manickam 11:19
Oh, I mean, chip is a semiconductor, right? I mean, it’s…
Siddhartha Ahluwalia 11:22
All the NVIDIA chips that we talk about today are all semiconductors.
Raja Manickam 11:26
Yes, yes.
Siddhartha Ahluwalia 11:27
So today, essentially, in the AI world, NVIDIA has a lead in all the the semiconductor chips that can facilitate high compute on a single chip.
Raja Manickam 11:40
That’s correct. That’s correct. So there’s something called Moore’s Law, right?
You may have heard of Moore’s Law, which…
Siddhartha Ahluwalia 11:46
Founder of Intel.
Raja Manickam 11:46
Founder of Intel, you know, which said that, you know, you could double your capacity in terms of, you know, every year, every second year or something. So you’ve exponentially, you know, you could pack more and more transistors. The more transistors you pack, the more computing you can do.
Right? And that’s where, you know, you talk about… And the way you pack more is you reduce the process nodes, right?
So now people are talking about 2 nanometer, 3 nanometer. So when I first started, it was in not nanometer, it was in microns, right? So thousands of nanometers.
Now you’re talking about, you know, 3 nanometer, 2 nanometer. So the transistor sizes have become smaller and smaller. So you can pack more, right?
The more you can pack, the more computing power you can do. So I think, you know, so this evolved, right? So now there’s a specialization in everything, right?
So there are equipment specialists, right? And within that, right, you know, they have lithography, has a, you know, like ASML, you may have heard of. They got applied materials.
And then you have, you know, fabs, you know, TSMC, UMC and all. They’re just doing the wafer fab. And then you’ve got the back-end packaging.
Siddhartha Ahluwalia 12:59
What is fab?
Raja Manickam 13:00
So wafer fab, so once you do the design, right? Now, you know, actually make the chip.
So this chip making is actually done in the wafer fab, right?
Siddhartha Ahluwalia 13:11
Fabrication is the short form of fab, right?
Raja Manickam 13:13
Wafer fabrication, yes. So wafer is, you know, you start with the silicon base, right? We call it like a pizza, right?
It’s either an 8-inch or 12-inch pizza, right? And then you build structures on top of it. You basically build a lot of transistors on top of it, connected with metal layers, right?
And the more you, you know, like a NVIDIA chip goes through, you know, huge amount of processing, right? They keep building, building on top of it, you know, 200, 300 steps. They have to go through one wafer before they come out.
And that’s why, you know, it’s very expensive as well, right? Both because the equipment that you need for NVIDIA kind of a chip is very, very expensive. And the yields and all that.
And then you have like analog chips like TI and all that does, right? Which has been around for, since I started in my career. Which you do not need, you know, extensive processing.
So one of my goals, you know, for India was, you know, let’s start with that kind of analog and power and all that. So IVP is a power semiconductor company.
Siddhartha Ahluwalia 14:28
And your first company, Tessolve, was a services company. You worked for other companies.
Raja Manickam 14:32
Previously.
Siddhartha Ahluwalia 14:33
Yeah.
Raja Manickam 14:33
Yeah, yeah. So Tessolve was, it provides services for Intel, AMD, TI. So we were doing like, you know, they would outsource some design or they’ll outsource testing, which they were doing internally.
They would just outsource it to us, right? So Qualcomm was one of the big customers that we had.
Siddhartha Ahluwalia 14:52
One thing which I want to ask is, semiconductor was, you know, since Intel was at its peak in 80s and 90s, semiconductor was at its peak, right? But since 95, when Microsoft became big, then since 95 till, let’s say 2023, it was all ruled by software, software and internet and mobile. And now semiconductors have become mainstream again.
Why is that?
Raja Manickam 15:21
Yeah, that is a very good question, right? See, you know, it’s like, it’s like the highway, right? Once you build a highway, there’s a lot of things that happens around the highway.
Siddhartha Ahluwalia 15:35
Yeah, you’re right. Once you build a highway, then there’s an ambassador car, but people are not satisfied with the ambassador car. They want to build new designs.
They want to build Ferraris.
Raja Manickam 15:44
Yeah. So when you have a highway, right, a lot of economy comes around it, right? You know, factories come next to highways, you know, homes comes and, you know, connectivity increases and all that.
But nobody bothers about who built the highway, what technology went into the highway, right? So our story is that, right? We are the highway, right?
The hardware guys are the highway. So, you know, there is not so much recognition on, you know, on the highway itself, right? So I always think that, you know, the software guys are daylight robbers, right?
So the highway has been built by this, you know, hardware people and they just take advantage of it, right? So they build, you know, applications on top of it. So as a consumer, you know, you see the applications, right?
You don’t see the, you know, what goes on inside, right? I mean, even if you look at your Apple iPhone and so on, right? There’s so much of technology that goes into the hardware, right?
Siddhartha Ahluwalia 16:46
So I think, you know, that was a story. I think Apple is able to capture that value because it does it end to end, right? If Apple was only selling the phone and no applications on top of it, it wouldn’t have had the current market share it has today.
Raja Manickam 17:02
Yeah. So actually, you know, so what has happened is, Apple is a very classic case where actually they build, you know, the computers, right? You know, you remember the Macintosh and all that, right?
And then they went to the, you know, iPod and then, you know, the iPhone and all that, right? So they actually build the hardware and then they build, you know, the applications that are like iTunes and stuff like that, right? So it’s a very good case of the hardware driving the business, yeah.
But it is an exception, right? Like you said, right, on Microsoft and all that, they were just riding on other people’s, you know, hardware, right? Now, what has changed is the Apple model.
Now, Apple actually makes their own microprocessor.
Siddhartha Ahluwalia 17:47
Okay, wow.
Raja Manickam 17:48
Right?
Yeah. So they bought a microprocessor company. Now they have their own microprocessor, right?
Siddhartha Ahluwalia 17:53
These are called M1 and M2 chips, right?
Raja Manickam 17:55
Yes. There was a company called P.A. Semi, which was a microprocessor startup, and they bought it.
And then now they make that M1 and M3. So the microprocessor is very tuned to the Apple application, right? So that’s why, you know, they gave the best power, best performance and all that.
Now, this is catching on, right? So Microsoft is trying to build their own chip. Amazon is trying to build their own chip, right?
Because, you know, all their data centers and all that, right? They want to have their own computing so that even Tesla, for example, right? One of the big success stories for Tesla is, you know, they have their own processor.
Siddhartha Ahluwalia 18:32
So in all Tesla cars, the entire compute, the hardware that powers the compute is their own?
Raja Manickam 18:38
It used to be NVIDIA.
Siddhartha Ahluwalia 18:39
Okay.
Raja Manickam 18:40
And then now, you know, they’re slowly, you know, there may be some still NVIDIA chips, but a lot of them are their own technology, right? But remember, you know, Tesla or Microsoft or Amazon or Meta, they’re not chip companies, right? They’re system companies.
They’re software companies, system companies, right? They’re solutions companies, right? So they are now getting into the, you know, they have some IPs, you know, like Google, right?
You know, they have their own IP for a chip. But then they don’t actually build the chip themselves. They outsource it, right?
And this is a huge opportunity now.
Siddhartha Ahluwalia 19:14
And in most cases, the outsourcing partners, TSMC.
Raja Manickam 19:18
Yes, yes. If you want to go into the highest node possible, you have to, you know, TSMC is the only game in town.
Siddhartha Ahluwalia 19:24
Yeah.
Raja Manickam 19:24
Right? You know, it’s very hard to reach. But my point is this, right?
You know, I believe now we are going one full circle, right? So all that outsourcing, you know, has happened, right? Now, you know, they’re trying to become vertically integrated again, right?
So the system companies want to make their own chips, right? It was like TI, right? So TI made the calculator chips.
You know, you must have heard of the TI calculator, right? So because they made the chip, they could make a calculator, right? Now, and then, you know, all these outsourced happened.
Then TI became just pure chip company, right? So this is, you know, a good point for India, right? Now we are in 2025.
Siddhartha Ahluwalia 20:07
Yeah.
Raja Manickam 20:08
We don’t need to be a chip company, right? We want to actually be a solutions company.
Siddhartha Ahluwalia 20:14
Yeah.
Raja Manickam 20:14
Right? So IVP is one of that, right? So I think…
Siddhartha Ahluwalia 20:18
What solution do you offer?
Raja Manickam 20:19
So in the power market, right, power segment, we have all the, you know, the typical, you know, the chips that you need to, in order to build a power module, right? So, okay, coming to power, I think this is a very important topic. See, I believe in electrification, right?
When you talk about decarbonization and so on, right? If you electrify everything, right, that is a path to decarbonization and so on, right? So that’s what is happening in the EV market and, you know, this.
And then you also have distributed power generation possibilities with the solar and wind and all that, right? You don’t need to have this massive, you know, coal-based or thermal, you know, hydro-based plants anymore, right? You can do distributed, you know.
So both of these phenomena is a very recent phenomenon, last 10, 15 years, right? So I think, you know, once you do the electrification, then you have a good chance of reducing, you know, pollution and, you know, like electric cars have less pollution, right? And so on.
So IVP is, you know, in a way, it’s an electrification enabler. So the chips that we are going to produce is to create this, to enable electrification in a big way, right? So we are focused on, you know, how do you build chips to enable electrification?
The other thing is, you see, if you look at an air con, for example, right? Air conditioner, you know, this is a good warming up in every year and all that, right? Even Bangalore is becoming very warm, right?
The more you use air conditioner, you’re actually, you know, you’re putting a lot of heat as well, right? So that may not be the right solution, right? It’s okay, you know, where, you know, maybe 20 years, 30 years ago, right?
So now you have a lot of opportunities and technologies available where you could use electronics, you could use these chips now to actually reduce the power consumption by significant amount, like less than half. We already have, you know, companies in actually Chennai is a professor, he’s from IIT Madras, 67 batch, right? So he’s still going strong.
Who is, you know, creating these kind of products. So I think there’s a whole slew of products that are going to come in the next 10-20 years, which will look a lot different from what we are used to, right? Whether it’s a noise pollution or, you know, the amount of electricity used per gadget, right?
So I think these are all new opportunities for India, which is now getting into the semiconductor game.
Siddhartha Ahluwalia 23:06
But, but I’m still inconclusive about what makes semiconductor hot again now, right? Why will semiconductor industry compete for the best talent along with software industry? Why will, let’s say, IITians which till now chose 50% of the batch used to go in software, even if they are coming from mechanical.
Why will they choose semiconductor in the next 20 years or starting from now?
Raja Manickam 23:36
You know, I think this is, you know, Siddhartha, this is a very good point. When I was, you know, when I graduated, I’m talking about 1982, a lot of guys from mechanical, naval architecture, we had a group called naval architecture. They all went into software because there was a digitizing effort that was going on, right?
So everything was getting digitized. So there’s a lot of opportunities in software. A lot of my batch mates ended up in Microsoft, right?
Even though they were, you know, ECE guys, right? At that time, we didn’t have computer science, by the way. It came in 85.
So I think, you know, one of the big things that is happening is they have found out that tuning the hardware to the software makes a huge difference, right? That is why you see, right, Microsoft. So Apple is a very classic case, right?
The performance of the Apple iPhone depends on the hardware, right? So if you own that piece, and if you’re able to tune, right, you actually will have a better system, right? So even the air conditioning I talked to you about, right?
If an air conditioner maker, right, is able to understand on the hardware piece of it. So there’s a lot of work that is going on that, you know, helps in this. Now, the other thing also, whenever you talk about chips, right, you always talk about computing power, right?
But what has happened in the last 20 or 25 years is you must have heard of MEMS, sensors, which are now made in the semiconductor, using semiconductor process, right? So MEMS, right, you know, electromechanical stuff, right? Now you can make a tiny motor, right?
So, you know, the number of sensors and MEMS you have in your, in your cell phone is a significant amount, right? That is how you track your movement and, you know, all kinds of things that goes on, right? The number of MEMS in this is, you know, huge.
Now, all these things were not available before, right? Now, semiconductor is able to integrate a lot of these things together. So it’s not just computing anymore, right?
So you have sensors, you have measurements, you know, you have all kinds of things that you can now build in a semiconductor process.
Siddhartha Ahluwalia 25:59
Just to summarize, you are saying the amount of new applications that can be developed or new experiences using technology that can be developed, hardware is going to empower these experiences and technology.
Raja Manickam 26:12
Absolutely. Absolutely, you’re right.
Siddhartha Ahluwalia 26:14
And there will be so much customization in hardware, like last 30 years of software was because there was so much customization in software, right? Yes. If you don’t like X, you could buy Y and you can customize Y to Z.
Now you’re saying hardware was very rigid, like you had a chip and you had to do it on the same thing on the chip you got, let’s say. There was no difference between a Dell laptop and an IBM laptop because they all had the same underlying motherboard, right, which was a semiconductor. And now what you’re trying to say is every maker will have their own customized motherboard and within that, the amount of customizations can be done, can give thousands of lives to new applications.
Raja Manickam 26:57
Yes, you can now differentiate your products based on the application and it is not too expensive to do that now, right? So, the chips used to that was made like 40, 50 years ago, right? And they’re very rigid, like you said, right?
They give you a chip and then they say, okay, you go figure out what you can do with it. Now it’s the other way, right? So, you take a lot of the software and you actually make a chip out of it, right?
Siddhartha Ahluwalia 27:28
How do you make a chip out of software?
Raja Manickam 27:30
No, you see, because let’s say, you know, your software requires a lot of computing, right? I mean, NVIDIA is a very classic case, right? The reason why NVIDIA is successful is because they actually had this software, AI software, they call it CUDA, right?
So, now they had that hardware that runs this CUDA very, very efficiently, right? So, because they, you know, most of, you know, AI engine is now, you know, actually runs off CUDA.
Siddhartha Ahluwalia 27:57
Let’s get back to the history of NVIDIA, like Jensen Huang, right? Who’s the founder of it? It was an unknown company till 2016, 17, making gaming hardware.
Let’s say, if you buy a laptop, there will be a small chip, it’s a gaming laptop powered by NVIDIA. But what suddenly changed? Why it became one of the most valuable companies in the world, suddenly?
Raja Manickam 28:23
See, you know, this is what you call luck, okay? He actually, NVIDIA was a graphics, you know, like you said, for gaming and a large amount of computing, it was a graphics chip company, right? Intel also had a graphics chips company, but they were more focused on the microprocessor and so on, right?
AMD is also, you know, had graphics, but they again, you know, they were more on the processing versus, you know, the graphics. So graphics is a pure computing, you know, thing. So you had a microprocessor and then you had just pure computing for this kind of applications.
AI came in, right, which requires a lot of computing. So AI basically saved NVIDIA, right? In fact, you know, Jensen would say that, right, you know, he was in the right place right time, right?
If AI was not there, you know, NVIDIA would not have been able to survive in the long term, right? But I think this is where, you know, the AI saved NVIDIA and, you know, and then they had a head start because there was no company which had that kind of graphics chips. And he’d also had, you know, I don’t know whether it’s intuition or, you know, he was very smart that he had the software to do that, which was the CUDA.
Siddhartha Ahluwalia 29:41
What does CUDA do?
Raja Manickam 29:43
So CUDA is basically, enables very fast AI computing, right? So, you know, any chip, right, you need to figure out how you compute that, right?
So, you know, which steps to go to get the, you know, the best result quickly, right? So you have these algorithms which did that. Just like, you know, if you look at the Microsoft example, right?
You can’t, you know, run away from Microsoft operating system. No matter who makes this software, whether, you know, Dell or whoever it is, you still have to, you know, use Microsoft, right? Microsoft Office or something like that.
So this is what it is. CUDA is for AI. So everybody has to now depend on that.
Now, he had a head start, right? Because they are the ones who owned CUDA. So, you know, they could make the chips for efficiently to do that.
Siddhartha Ahluwalia 30:33
But I’m still not able to understand what CUDA does for an end user like me.
Raja Manickam 30:38
So, no, actually it is internally, right? There is a, you know, software, there’s multiple stacks, right? So where you see is on the application, you know, on the top, right?
But the internally, you know, how efficiently it runs, right? So it’s more in the, you know, the deeper layers.
Siddhartha Ahluwalia 30:54
For example, you’re saying chat GPT today by OpenAI is powered by CUDA.
Raja Manickam 31:01
At some level inside will be, you know, at the chip level, there will be some form of CUDA will be running.
Siddhartha Ahluwalia 31:06
And let’s say tomorrow somebody produces another NVIDIA like chip, and they don’t have CUDA. What difference does it make?
Raja Manickam 31:14
So they will have to, you know, build all that, you know, software, compilers, and all those kind of things.
Siddhartha Ahluwalia 31:19
This is the software that interacts with the end application layer and the chip. That’s what you’re saying.
Raja Manickam 31:26
It’s mostly, you know, on the chip side of things. It’s on the deeper, you know, inside. But yeah, of course, right?
You know, the next layer, it will have some interface. I mean, you could think about something like a, if you look at a microprocessor, right? So microprocessor is the same thing, right?
So you have a lot of software that runs on the chip. The application sits on top of it, right? So now, if you are the guy who’s going to be defining what the chip is going to do, right, you know, or that firmware level, you have an advantage.
Siddhartha Ahluwalia 32:01
Why is nobody able to compete with NVIDIA?
Raja Manickam 32:04
Well, I think they are. They are now, right?
As I said, right, you know, they were the first in the, you know, so they have a head start. I mean, you know, if you look at that, even on the chip side, they’re struggling, right? Because, and, you know, at some point Intel or AMD or, you know, all these guys will start to give them some competition.
And then, you know, like people like Microsoft, right? Because they come from the application side. They’re now starting to make their own chip.
Siddhartha Ahluwalia 32:31
My thing that I wonder is, why did Intel and AMD get left behind in the AI race as compared to NVIDIA?
Raja Manickam 32:42
I think, you know, see, again, same thing, you know, when you are in the microprocessor business, which was doing really well for Intel, right?
Siddhartha Ahluwalia 32:51
All three were doing microprocessor, right?
Raja Manickam 32:53
Not so much NVIDIA. NVIDIA was on the graphics.
Siddhartha Ahluwalia 32:56
Okay. So they were building graphic chips and Intel and AMD were building only micro.
Raja Manickam 33:02
They had some graphics, but NVIDIA was a graphics only chip.
So, you know, if you look at it from a business perspective, right, Intel was focused on the microprocessors, you know, the x86 thing, right?
Siddhartha Ahluwalia 33:15
The one we call RAM in the computer, right?
Raja Manickam 33:19
Well, the x86, right? You heard of x86. Or you, you know, nowadays you talk about ARM, right?
And then you talk about RISC-V, right? So they’re all like, you know, different ways of, you know, running.
Siddhartha Ahluwalia 33:32
The hardware compute power.
Raja Manickam 33:34
Yes. Yeah. So the graphics is, you know, is a specialized thing you can see.
And it was only in the gaming market, right? And gaming market was a small market. But once AI came into play, right, you know, that, you know, these guys did not expect that to take off so, so quickly.
Siddhartha Ahluwalia 33:53
And why was AI not able to run on microprocessor and why only specialize in NVIDIA chips?
Raja Manickam 33:59
No, I think, you know, you see, you will still have a microprocessor. They are processing a lot of it, but a lot of the heavy load of computing, they just throw it into the graphics chips, right? So, you know, I mean, it’s a lot more nuanced, see, because, you know, the chip size matters, right?
You know, how big the chip is, right? To do so, you need to optimize that, right? So NVIDIA being a chip company, they did pretty good on it.
But even lately, right, you see they have yield issues, you know, they’re not able to deliver enough, right? But because they’re only game in town, they can charge $20,000, $30,000 per module, NVIDIA’s largest customers. There’ll be all these guys, you know, who are playing, you know, supplying to data centers, you know, where there’s a huge amount of computing, right?
So, all these guys, the guys who are building all these servers, right, super micro and all that, but they’re like in between guys, right? So ultimately, it will go into data centers and, you know, government.
Siddhartha Ahluwalia 35:04
You’re saying ultimately Microsoft, Google, and Amazon are the largest consumers of NVIDIA. I heard, like, the last AWS, like CEOs of these companies, for example, Mark Benioff from Salesforce, he was having dinner with Jensen Huang and trying to convince him to give him some more, trying to buy some more chips for AI. Everybody’s trying to woo him.
Raja Manickam 35:31
Yeah, because, you know, like Salesforce, right? You know, they want to do a lot of computing to, you know, for the AI learning.
Siddhartha Ahluwalia 35:39
But most of the computing is done on the cloud. I don’t know what’s the relation between their hardware and…
Raja Manickam 35:43
No, because cloud ultimately has, you know, hardware sitting there. So that hardware is NVIDIA’s hardware. Broadcom is also getting into it, by the way.
Siddhartha Ahluwalia 35:52
Okay. And do you think anytime NVIDIA’s lead will get shattered because now everybody would have this technology?
Raja Manickam 36:02
I always believe that you can’t hold, like, 90% market share in anything. Right. So eventually, yes.
You know, I know how quickly it is, you know, anybody’s guess.
Siddhartha Ahluwalia 36:13
And tides change very soon. Like, for example, IBM was the NVIDIA of 1980s, like having 90% share in the computer market. And then everybody had the computer technology, so it got democratized.
And then, essentially, IBM had to pivot to being a services company.
Raja Manickam 36:35
Exactly. See, we imagine if IBM… IBM was a leader, right?
You know, they had phenomenal way for fabs and so forth.
Siddhartha Ahluwalia 36:42
I think IBM was like the desirable company to work for in your time when your batch graduated.
Raja Manickam 36:49
Well, you won’t believe this story. I had job offers from TI, Fairchild, and I’d interviewed with IBM. I was waiting for IBM’s response.
Right. And finally, I called IBM. I told them this.
I said, look, you know, I want to take the job, but I really want that. Then IBM did not give me the job. I did not get the job in IBM.
Then I chose TI. So you’re right. So IBM was the most sought out at those days.
Siddhartha Ahluwalia 37:19
Was the Google of those days.
Raja Manickam 37:20
Yes. Yes. So, I mean, even if you look at Google, right, you know, we would have thought that somebody would challenge Google, right?
But now, you know, because with AI and all that is becoming, you know, the world, right? So I’m sure that, you know, NVIDIA will have something and, you know, the whole market is growing. So I don’t think, you know, you can keep buying $20,000, $30,000 chips anymore.
Siddhartha Ahluwalia 37:45
Where do you think India has an advantage? Like India definitely had an advantage in terms of IT services, because when Narayan Murthy scaled Infosys, like what happened was 1991 liberalization and $1 suddenly costed like Rs. 20 or Rs. 40. And now when Narayan Murthy was earning in dollars, he was paying in rupees and he could train all the fresh talent coming out of colleges in Java, in C++ or any computer language for that matter. India had a huge unemployment problem in the 1990s.
So he said, why not train millions of people in software and, you know, I’ll make sure that I’ll get enough work for them. And that’s how Infosys, TCS, Wipro became the leader in IT, yes, yes. So my question to you is, where does India have an advantage in this, you know, new semiconductor and AI race?
Raja Manickam 38:49
Yeah. So I have a slogan, right? And say, you know, there is an Indian in every chip, there is an Indian in every chip, right?
And the reason why I say that is, we touch every aspect of making a chip, right? Software that, what we call EDA tools, which are used for, you know, designing a chip.
Siddhartha Ahluwalia 39:11
But Indians are not manufacturing it.
Raja Manickam 39:13
Yeah, so I’ll tell you, right? So there is an Indian, you know, who would have been in the design aspect of it, or the tools for the design, would have been in the software piece of it that goes into the chip, because the chip doesn’t work by itself, right? You know, you have to have firmware and operating system and all that.
So that’s why I said there is an Indian in every chip, except, right, the manufacturing piece was missing, right? So by and large, right, we’ve already, you know, are there in the semiconductor, right? So if you look at other countries, right, the size of, you know, one is, you know, the size, right?
Other than China, right? You take Taiwan or Korea or Malaysia or Singapore, all these countries, right, with a lot of the manufacturing is happening. They just don’t have the, the size, right?
Siddhartha Ahluwalia 40:04
The manpower.
Raja Manickam 40:04
The manpower, right, that is required. That’s one.
Second thing is the depth of talent, right? You know, I mean, think about Taiwan, right? You know, how many, you know, graduates they can come out in a year, right?
Siddhartha Ahluwalia 40:18
I don’t know.
Raja Manickam 40:19
I mean, it’ll be like Bangalore will have more than, right, you know, the entire Taiwan put together, right? Or Chennai, for example, right?
You know, they have so many engineering colleges and all that. So I think, you know, one is, you know, and also the depth of, you know, if you look at all these countries, right, Malaysia, Thailand, Philippines, and Singapore and all that, they have a lot of manufacturing and not so much on the front, but on the back end, we call packaging and so on. But the depth of, you know, engineering requirements, right, are still not there, right?
So they do, you know, they can do manufacturing, but not what I call engineering intensive manufacturing. And also the other phenomenon that’s happening is this customization is driving a lot more engineering activity, right? So it’s not that, you know, you build, you know, a system like today, right?
You can pick a car and you can pick the features in the car that you want, right? Right. My father-in-law, you know, he always had an ambassador car, right?
There’s no feature other than the grill in the front, right? You can’t, you know, pick any feature, right? So think about it, right?
So there’s so customization that goes on, which is basically engineering, right? So a lot of that activity also happens in India. We are so, you know, set up for doing that, right?
And we also have like what I call frugal engineering, right? Culturally, you know, we know how to, you know, work and bring this out. So there’s a huge advantage.
And semiconductor is a technology that the globe requires, you know, whether you’re talking about ESG or global warming or whatever it is, right? You require semiconductor has or technology has a solution. And semiconductor is a foundation for that, right?
Now, the solutions that have come so far, in my opinion, right, is just scratching the surface and the use the most benefit has gone towards the affluent people, the Western countries and affluent, right? It has not touched the 80%, you know, which are the normal, you know, lower, you know, which is what is required, right? Whether you’re talking about, you know, clean energy or, you know, clean water and all those kinds of things, technology has a solution.
So I think in the next 10-20 years, right, you will see solutions for the masses. And that solution needs to be a frugal solution, right? You know, Western country, you know, you could buy something, you know, air con for, you know, say $1,000 or so, right?
Now, we need to have, you know, cooling available for $50 or so, right? So how do you create those kinds of things, right? So I think there’s going to be a lot more of this.
The products that are going to come out of India are going to address the global issues that, you know, to solve some of this.
Siddhartha Ahluwalia 43:19
But the thing is, why doesn’t India still have a TSMC or NVIDIA or Intel or AMD till now?
Raja Manickam 43:28
Yeah, so that’s why I told you, right? See, we, one is, you know, we didn’t have good policies and all that, you know, the import export policies, you know, they were government related policies, tax related policies, and so forth, because a lot of this manufacturing requires a lot of stability, right? And, you know, so we are kind of a closed country, right, from an economic standpoint.
And then the mid-90s, you know, then they started opening up and all that kind of stuff. And, you know, people were, you know, all these companies you see were actually run by Indians, right? You know, if you go to Qualcomm, probably half of Qualcomm is Indians, right?
You go to Applied Materials, which is a big, you know, the largest semiconductor equipment maker, you saw, you know, a lot of Indians as well. But I think India as a country, right, was not suited for, you know, from whether tax rules or, you know, many of these, you know, political.
Siddhartha Ahluwalia 44:30
Has it changed? Because unless this changes, nobody will start a company in India in semiconductor.
Raja Manickam 44:35
So, certainly, I think, you know, I mean, personally, I feel that, you know, we’ve come a long way in enabling some of this, you know, still there’s a long way to go. But certainly, there’s a lot of changes that has happened, right? You know, I’ve now, you know, started my company.
I want to do the manufacturing in India. The tax laws need to be a lot more stable, right? You can’t change the law every year or retro the law, right?
I have some bad experiences in that. But I think, you know, it’s changing. Certainly, it’s changing.
And if you look at, you know, the semiconductor market is supposed to double, right? It’s about half a trillion right now, right? 500, 600 billion dollars is a semiconductor market.
It’s supposed to double, right? It took about 40 years, so, you know, to get to this level. In the next 10 years, it’s going to double, right?
So, imagine, right, you’ve got to squeeze this into a 10-year frame. India has to be a bigger player in this. Where will you get the people?
Where will you get, you know, the cost advantage?
Siddhartha Ahluwalia 45:36
But I’m not seeing Indian entrepreneurs, many people like you venturing into semiconductor space. Unless until they start their companies in 2025, India will not reap the benefits in 2035.
Raja Manickam 45:48
Yeah. So, I think, you know, one of the problems that I see, you know, some of this, you know, especially manufacturing requires capital, right? And we are so, you know, what I call ROI companies, right?
Return on investment or return on equity, you know. So, even before we start, right, you know, we look at this, whereas China never looked at that, right? They said, hey, you know, we need to do this and, you know, it doesn’t matter, we’re going to lose money, right?
So, I think government is now, you know, what we call, you know, viability gap, right? They’re funding the viability, right? So, from a manufacturing standpoint.
Siddhartha Ahluwalia 46:28
I think the mindset you are right needs to change. We rather than saying that, hey, a share market would give me 15% ROI, why should I invest in a semiconductor company, which is new, so much risk.
People need to put money in innovation.
Raja Manickam 46:44
Exactly. I mean, think about, you know, NVIDIA and all these guys, right? You know, NVIDIA would have gone through at least a few times where they could have gone bankrupt.
Right? I mean, Jensen will be the first guy to admit, right? So, all that, you know, and, you know, the other thing is, what I, you know, I’ve studied is the US, for example, the affluent people, the wealthy people actually funded this, right?
So, like Lockheed Martin was funded by the local people where they were, you know, so, you know, people who were in the steel or some other industries, right? You know, they funded some of these things, right? And I think, you know, that needs to happen as well.
Siddhartha Ahluwalia 47:25
Yeah, but that has not happened enough. That’s why the reason most of Indian venture capital funds are dependent on foreign money because even the Indian, you know, big family houses, which in the US would, you know, put their money into university endowments, right? They want to do it all themselves.
If you see today, if you talk about semiconductor defense, Reliance and Adani’s of India, they want to do it all themselves and capture the market share rather than investing in new entrepreneur, which I think is strange. Like nation building would require much more efforts from the top, not just from the government work from.
Raja Manickam 48:07
Yeah, so far, you’re absolutely right. All the initiatives that has gone in the 23-24 has been government initiatives, right? Private sector needs to step up, right?
I mean, there’s only so much government can do, right? So, I think one is, you know, private sector has to step up and I believe the, what I call challengers, right? Not really the, you know, the big companies that, you know, the brand names that we know, but there are a lot of companies which are in the, you know, 500 crores, 1000 crore kind of, you know, and the youngsters, you know, that families, you know, that kind of, you know, they’re willing to put 100 crores, 200 crores, 300 crores, you know, to try something new, right?
I see that as, you know, I think a very positive sign.
Siddhartha Ahluwalia 48:58
So, one thing is that there is announcement of India building a 100,000 crore semiconductor factory, like is some work really happening on that?
Raja Manickam 49:10
The first, yeah, you know, so like the first batch, I think was 140, 150,000 crores or something has been allocated, right? You know, between micron, Tata’s, CG power and so on, right? So, I think that has been, you know, and the work is actually going on.
Now, how soon they will come up and all that, you know, I think Tata’s will definitely make it happen, right? I think the more important thing is, you know, how do we build products out of this? How do we get Indian companies to benefit out of this?
Right? We don’t want to get into this cycle, where we’re building again for these, you know, foreign companies to use, you know, the fabs and, you know, and basically subsidized by…
Siddhartha Ahluwalia 49:53
We don’t want to become a semiconductor services nation like IT services nation.
Raja Manickam 49:57
Exactly. You’re absolutely right. And that is where, you know, I feel that, you know, we need to step up, right?
We need to create our own, you know, Qualcomm or Broadcom, right? IVP is one of them that, you know, we’re trying to do that. And there are many more, right?
No needs to come. And, you know, government, I think to some extent, right, this DLI and all that is, you know, is helping. States are also, you know, starting to step up.
The private sector really, really needs to fund. That’s why I said, look, the very easy thing to do is, when government is funding a startup, why don’t you piggyback with the fund?
Siddhartha Ahluwalia 50:34
Yeah.
Raja Manickam 50:34
Right? You know, government is getting free money. We can come in and take equity with these guys, right?
So, I think, you know, that cooperation needs to happen.
Siddhartha Ahluwalia 50:46
And what are some of the U.S. partnerships with India chip industry right now that are starting to happen?
Raja Manickam 50:53
I think, you know, see, you know, it’s very, I don’t know how to explain this, but, you know, geopolitically, right, you know, U.S. is pretty much trying to align with various countries, right? And in semiconductor, right, there is this very high-end and low-end, you know, all kinds of products, right? You know, the whole spectrum of products required, right?
Say, for example, you know, you want to make a solar inverter. You know, U.S. is also pushing for a lot of solar engine, all that, right? The solar inverter doesn’t require a state-of-the-art wafer fan, right?
You know, IVP is, you know, making some products for the solar inverter, for example. Now, that manufacturing is not going to happen in the U.S. U.S. is not going to make inverters, right? Because one is, you know, their cost will go through, you know, and, you know, they don’t even have the people to do it, right?
So, I think, you know, alternatively, right, they’re going to look at countries like India to take that kind of market, right? So, I think, you know, from U.S. standpoint, they will want to maintain the lead in technology, but anything mature and all that, they want to come from certain, you know, alignment that they have, right? So, I think India has a very good shot at it.
Siddhartha Ahluwalia 52:12
But can India really meet the global semiconductor standards and needs?
Raja Manickam 52:17
Absolutely. Absolutely. You know, you know, see, when I started Tessolve, for example, you know, there was a lot of questions, you know, oh, you know, you probably will do some low-end stuff and all that.
I’m very proud to say, I mean, Tessolve does some of the, you know, the Snapdragon, you know, Qualcomm products.
Siddhartha Ahluwalia 52:39
Design or manufacturing?
Raja Manickam 52:40
No, I mean, we do the test development for that, you know, the hardware that we use for testing, the design is done in Coimbatore, right? And these are like, 50-layer, 60-layer, you know, boards, right? Big boards.
It’s designed in Coimbatore, right? The test development is done here locally, right? So, I mean, you know, all these companies, whether it’s Intel or TI, I know they’re very dependent on India to do it.
So, absolutely, I have no doubt. Now, on the manufacturing, you know, if you want to be like a TSMC, it’s probably going to take 20 years.
Siddhartha Ahluwalia 53:13
Why is that?
Raja Manickam 53:14
See, because, again, right, you know, design and engineering and all that, we are good. But when it comes to manufacturing, right, we are not very hands-on people.
Siddhartha Ahluwalia 53:23
Why do you say that? They are the manufacturers in India.
Raja Manickam 53:26
So, no, they are, right? So, the problem is, the IT and software industry has pushed a lot of these bright people to, you know, be, you know, sit at a desk and, you know, create code, right? And they’re very good at it.
But I think there needs to be a change in our education system, where even in from high school onwards, right, people should be trained to enjoy working with their hands, right? Semiconductor requires that, right? You know, if you work in a fab, right, you actually have to work with, you know, chemistry, you have to work with mechanical, you know, all kinds of things, right?
And we are not, you know, so if you look at Europeans, for example, right, there’s a lot more hands-on activity they do, right? They repair their own bikes, they repair their own thing, right? They can fix their own car if they need to, right?
Our guys, you know, even changing engine oil, they’ll have a problem, right? They don’t know where’s the nut and bolt, right? So, I think, you know, that needs to happen in our education system and take a lot of pride in, you know, building products.
So, I think that will happen. And just culturally, right, you know, there’s a lot more precision required, right? But, definitely, I feel that, you know, we can do it.
There is no question. I am absolutely positive.
Siddhartha Ahluwalia 54:46
So, one question is that there’s a $15 billion incentive plan for semiconductor in India to become a semiconductor hub globally. Is, Is that also happening?
Raja Manickam 55:01
The first batch, you know, I mean, the government actually, you know, gave the money and, you know, the first round, right, last year, 2024. So, the second round, you know, I was told that, you know, the 2.0 is going to be announced, you know, in the next budget. And there’s a significant amount of money that is coming.
Now, they may focus on the ecosystem guys, equipment guys, you know, chip product companies and so on. May not be as much for the manufacturing.
Siddhartha Ahluwalia 55:31
Which are the Indian companies that you respect in semiconductor space?
Raja Manickam 55:38
Not many come to mind.
Siddhartha Ahluwalia 55:39
But a few. It can’t be just IVP in semiconductor.
Raja Manickam 55:44
So, I think, you know, L&T semiconductor is making a good push, right? I’m aware of what they are trying to do. L&T has put $100 billion towards that, right?
Siddhartha Ahluwalia 55:57
But L&T is not an Indian company, right?
Raja Manickam 55:59
No, L&T is an Indian company.
Siddhartha Ahluwalia 56:00
Larsen & Toubro?
Raja Manickam 56:01
Yeah, yeah, yeah. Oh, yeah, L&T is an Indian company. After the, I don’t know, after the independence or even before independence, it became an Indian company. It’s 100% Indian company.
They’re into construction.
Siddhartha Ahluwalia 56:12
Mostly construction.
Raja Manickam 56:14
But they have, you know, they just bought a year ago, they announced, they put $100 billion in creating a chip company. So, it’s called L&T semiconductor. And then the other one, you know, I would say CG Power, right?
You know, you must have heard of CG Power. You know, they’re also very successful, part of the Murugappa group. They’re also, you know, trying to, you know, bring up their own chips, right?
So, there’s a good one. Tata bought a chip company called Saankhya Labs. So, you know, they’re trying to get into making chips.
So, I think, you know, they’re all small now, but I think, you know, they put enough, you know, beat behind it, build it up, you know, they will be successful.
Siddhartha Ahluwalia 56:57
And you have said you won’t retire until semiconductor manufacturing in India really happens, right? Now, you are at age 68, and this is your second startup. Why is it important for you to make this happen?
Raja Manickam 57:12
I’m a big believer. I’m 69 this year. I’m a big believer that without manufacturing, you can’t build a chip company, right?
See, why a small country like Taiwan…
Siddhartha Ahluwalia 57:26
Is IVP in manufacturing?
Raja Manickam 57:28
No, IVP, I’m saying it will be a chip company, a fabulous chip company.
Siddhartha Ahluwalia 57:31
You will just design chips?
Raja Manickam 57:33
We will design chips and we will actually sell our chips in our brand.
Siddhartha Ahluwalia 57:36
Who will manufacture them?
Raja Manickam 57:37
We will outsource the manufacturing. So, let’s say Tata comes up with the fab, we will, you know, outsource to Tata to make the wafer. Maybe CG Power will do the packaging and, you know, this is like NVIDIA, right?
See, NVIDIA doesn’t have any manufacturing, but they’re a chip company.
Siddhartha Ahluwalia 57:52
They outsource all their manufacturing to TSMC..
Raja Manickam 57:54
Everything, 100%. They don’t have any manufacturing.
Siddhartha Ahluwalia 57:58
That’s why when NVIDIA stocks rose, TCMC stocks also rose a lot.
Raja Manickam 58:02
Goes up and, you know, like Intel comes down because Intel has its own manufacturing.
Siddhartha Ahluwalia 58:05
Okay, why is that?
Like, why did it come down?
Raja Manickam 58:09
No, because, you know, now you have to manage your assets, right? You know, guys, I told you, right, you know, you got to keep investing in manufacturing to keep up with technology. So, either you’re a chip company, you know, without any…
Siddhartha Ahluwalia 58:22
So, you’re saying you have to decouple both of them?
Raja Manickam 58:23
That seems to be the formula. If you look at Broadcom, Qualcomm, they’re all, you know, fabulous. They don’t have any manufacturing.
The guys who have manufacturing are struggling. Right?
So, manufacturing itself will be an activity by itself and a business by itself. And then you have the, you know, chip companies which will use them, right? So, I believe that, you know, in India, right, you know, if we don’t have a manufacturing capability, we will not be able to cross the, you know, remember I told you, you do design and then when you make the chip, right, the physical chip, there’s a lot of learning that happens there.
If we don’t have it in India, we are dependent on Taiwan or Japan or Korea to make that for us. They have the knowledge, we don’t. So, that is why even a small country like Taiwan has MediaTek, for example.
MediaTek is as good as Qualcomm, right? There’s a company called RichTech, which makes analog chips, which is, you know, as good as, you know, some of the big, you know, TI and OnSeminar, but we have none, right? And I believe that unless you have the manufacturing piece of it, it will be hard for us to scale to become a chip company.
Siddhartha Ahluwalia 59:41
And who are the top three companies in India in manufacturing, right now?
Raja Manickam 59:44
So, the ones that are being funded now, right? Tata’s, and then CG Power, and then Keynes.
Siddhartha Ahluwalia 59:54
And by manufacturing, you mean any chip designer can come to them and they’ll manufacture chips for them?
Raja Manickam 1:00:00
Well, yeah. So they’re building the, you know, like Tata is building a wafer fab. So, you know, I will design my chip to be fabbed, you know, in their fab, you know, we’ll have some design rules and all that, and they will, you know, fab and give me the wafers, then I’ll get it assembled and packaged it in CG power or something.
Siddhartha Ahluwalia 1:00:20
What is the assembling? Like, what is it?
Raja Manickam 1:00:22
What we call, see, when it comes in a wafer, right, you know, the circuits come in a wafer, but ultimately, you’ve seen the PC board, right, you know, you have to mount it on a PC board. So, you need to do a packaging, the interconnect happens.
Siddhartha Ahluwalia 1:00:36
And CG power is the only company that does it?
Raja Manickam 1:00:39
No, no, CG, you know, there are many, you know, Tata is also doing it. So, they have a fab and they have what we call them, you know, OSAT. OSAT has been given to three companies, which is CG power, Tata and King.
Siddhartha Ahluwalia 1:00:51
And do you require a government license to do it?
Raja Manickam 1:00:54
No, there’s no licensing, but, you know, the government is giving them money, right, 50 percent is coming from central government.
Siddhartha Ahluwalia 1:01:01
How much does it require to build a manufacturing plant for semiconductors?
Raja Manickam 1:01:05
No, I think fab, they throw numbers in billions, right, wafer fab.
Siddhartha Ahluwalia 1:01:09
It requires billion dollars to build a fab manufacturing company?
Raja Manickam 1:01:12
No, I mean, they, probably the one that Tata is building, they are talking about 4 to 5 billion.
Siddhartha Ahluwalia 1:01:19
What is the lowest, if an entrepreneur tomorrow wants to start a manufacturing company? If an entrepreneur like you or me tomorrow wants to start a manufacturing company, how much money do they require?
Raja Manickam 1:01:30
No, I don’t think, you know, an entrepreneur will be able to fund a manufacturing company. We can do a startup for chip design and, you know, using those manufacturing, right? That is why government is, you know, giving this money.
Siddhartha Ahluwalia 1:01:44
Why can’t they? Is it just because of the capital requirement or something else?
Raja Manickam 1:01:48
No, I mean, you know, let’s say I want to start a wafer fab.
Siddhartha Ahluwalia 1:01:52
Somebody, how much money do you need to raise to start a wafer?
Raja Manickam 1:01:55
So, you know, if the government is giving it, you know, 70%, I have to raise 30%, right? So, let us say it’s a 5 billion dollar.
Siddhartha Ahluwalia 1:02:03
So, you think minimum requirement is a 5 billion dollar?
Raja Manickam 1:02:06
No, I have my own views on this. I think you can start with half a billion all the way to 10 billion if you want, right? But I don’t think our country is ready for even 5 billion dollar fab, because that’s, you know, very high end, you know, loads and all.
We have to walk before we can run. And then, you know, also, right, you know, you don’t, you know, you’re not ready to make NVIDIA chip, right? As I said, there’s an all analog chips and all those kind of things that you can make.
400, 500 million dollars you can do, right? So, actually, we are telling the government, you know, focus on this kind of smaller manufacturing units.
Siddhartha Ahluwalia 1:02:45
If you have to put a number to it, what would be the India semiconductor needs in terms of billions of dollars? And how much of it is met locally? And how much of it is met through imports?
Raja Manickam 1:03:00
Look, today, almost 100% semiconductor is imported. Nothing is made here, 100%.
Siddhartha Ahluwalia 1:03:07
And if you have to put a number, how many billion it would be?
Raja Manickam 1:03:12
I mean, they are throwing a number like our import bill for semiconductor is 45 billion or something like that. 45, 50 billion.
Siddhartha Ahluwalia 1:03:23
And we have woken up just now to address that.
Raja Manickam 1:03:28
Yeah, I mean, I think a few years ago, right, you know, there was some minister made the comment that our import bill for semiconductor is higher than our import bill for oil.
Siddhartha Ahluwalia 1:03:39
Yeah.
Raja Manickam 1:03:40
Right, so we already suppressed that, right? And, and then, you know, it is also very strategic, right, for a country, a large country like India, right? You know, if, let us say, you know, you are importing from some countries and they decide, you know, they do not want to supply to India, we will be stuck.
Siddhartha Ahluwalia 1:03:57
Yeah, I believe a complex thing like every defense equipment, like an aircraft or a tank would have thousands of semiconductors.
Raja Manickam 1:04:05
Yes, absolutely.
Siddhartha Ahluwalia 1:04:06
And even if we manufacture our tanks and aircrafts in India, we would have to depend on a foreign company to give us the semiconductor.
Raja Manickam 1:04:16
Yes, yes, it is a very dangerous thought, yeah. So that is why, you know, I think, you know, from a strategy standpoint also, right, we need to have our own industry.
Siddhartha Ahluwalia 1:04:26
And does China own their own end-to-end supply chain in semiconductor?
Raja Manickam 1:04:29
So they started this about 30 years ago, right? So they have, you know, indigenized a lot of the products.
Siddhartha Ahluwalia 1:04:36
Which is their largest company in semiconductor space?
Raja Manickam 1:04:39
They have, you know, a few companies, you know, from a fab standpoint, there is a company called SMIC, S-M-I-C, which is pretty large. You know, I mean, they took actually from Taiwan and all that. And then there is Huawei, right?
You know, Huawei, Huawei makes telecom equipment. They have a silicon arm called HiSilicon.
Siddhartha Ahluwalia 1:04:59
When it is so critical to our country, why are we not discussing it enough?
Raja Manickam 1:05:04
I mean, currently we are, right, and I would say that, you know, we woke up a little bit late, for sure. But I think, you know, it’s been addressed in the last two, three years. Prime Minister talks about semiconductor in the United Nations.
So I think, you know, there is enough traction now, visibility.
Siddhartha Ahluwalia 1:05:27
And are there a lot of dependencies, as you said, to start a semiconductor company like yours, IVP?
Raja Manickam 1:05:34
Yeah, yeah, yeah. So, you know, we need a lot of experience of working outside of India, right? You know, when you send the design files, you know, it is, make it.
And then you need to understand, go to market. How do you sell into this? How do you, you know, so all that experience is, you know, it doesn’t exist within India, right?
So we are consumers, you know, so we are not makers, right? So, you know, so all that the learning has to happen.
Siddhartha Ahluwalia 1:06:07
Thank you so much, Raja. This has been a very eye-opening and insightful conversation. I wish to do a part two of it sometime soon.
But I think this is enough food for thought for our audience to chew upon and realize why semiconductor is so important to our nation.
Raja Manickam 1:06:23
Thank you. Thank you, Siddhartha. I enjoyed the conversation.
Thank you.