Transcript
Speaker: You're in a situation of war and the cameras are being used as backdoor to watch what you're doing or to hear what you're talking about. No more your drone, it's a Chinese drone. Today it's changing. So what's the trigger? For years, India built the plastic case and somebody else built the brain inside. A handful of engineers finally decided that was not good enough.
Speaker: Manu Iyer, co-founder and general partner of Blue Hill Capital, is the man writing them their first cheque. On this episode of Founder Thesis with Akshay Tad, Manu breaks down his playbook for Frontier Tech.
Speaker: China is now probably at par with the US. I will disagree slightly with that assessment. They've been able to copy the chips. They've not been able to copy the machines. Is anyone trying to build the volunteer of India? Google was untouchable until it was.
Speaker: Manu Ayer, you are the founder and GP at Blue Hill VC. Welcome to the Founder Thesis podcast. ah What is Blue Hill VC? Give me like a headline understanding of it.
Speaker: Sure. ah First of all, great to be here, Akshay. um Blue Hill VC is a frontier tech fund. What we're trying to invest in is foundational tech ideas grounded in intellectual property in either engineering or science.
Speaker: So we're trying to back foundational ideas being built in India around semiconductors, defense, space, material science, manufacturing, biotech, and other such areas.
Speaker: I noticed you say frontier tech and not deep tech. um Isn't this the same thing? I mean, what's the difference? I think that's by design, Akshay. So though the recent trend has been deep tech is everything, right? ah Everyone is a deep tech startup and everyone is a deep tech investor.
Speaker: And, you know, I'm trying to ah differentiate what we're doing vis-a-vis ah what is happening in the market. ah Generally, the bar has been set fairly low. There was this phase when everything was AI, right? Everyone would just get a.AI domain and would expect like a 20%, 30% markup in valuation just because they had the domain.
Speaker: And I think we're we're broadly in that you know bubble phase in deep tech as well. um every Every startup is somewhere or the other deep tech. Software startup is deep tech. ah you know ah Hardware startup is deep tech. Guys copying stuff from you know that already exists somewhere else is deep tech.
Speaker: right Guys adding 10-15% to the status quo is deep tech. So ah somewhere I'm trying to break that logic or break that barrier and put a clear case to what we're doing ah differently.
Speaker: And that's why I call it frontier tech. What ah like, ah how would you define frontier tech? What is truly frontier tech and what is not? Right.
Speaker: So ah like I started off saying, the idea of frontier tech for me is tech that is grounded in intellectual property being built ah can be soft, can be hard.
Speaker: But people trying to change the way something is being done in a way that has not been done before. globally. So I'm not using India as you know the scale or the a comparison point or the reference point. I'm using a global scale ah in trying to compare the tech that is being built because clearly we have the capability. We have some of the best engineers in the world and you know best scientists in the world.
Speaker: ah the The gap has always been around the commercialization. And today that is clearly happening. And that's what we are trying to back. So somebody building, let's say, beauty products which are science-backed, would they qualify as deep tech or frontier tech?
Speaker: Yes or no. Yes and no. In the sense that ah if there is a clear science moat around it, sure. Right. If ah you're doing something on the science front ah that, you know, changes the paradigm of what has been done already, for sure.
Speaker: Got it. Okay. How do you... ah do the due diligence on somebody who's pitching to you to really understand that, uh, is, is this truly frontier tech or not?
Speaker: Uh, I think that's, that's a great question. Again, see the, the thought process with us as a fund investing in some of these spaces is we want to have people in our team who are fundamentally engineers who come from that background, uh, I think that's like the base level for us.
Speaker: ah We need to have people of you know that ah you know knowledge, that understanding need not be specifically from that very field, but at least they have a deeper understanding and you know ability to pull apart the strings within an engineering use case or within a science use case. So for us, that's the first step.
Speaker: The entire investing team is just engineers. And this I would compare to other funds ah Generally, VC is you know people commerce background people, largely.
Speaker: ah Specifically, we have our own interest spaces within this. I myself am an electrical engineer. So EVs, energy is a strong area of interest.
Speaker: Semiconductors is also broadly connected to that space. ah The ecosystem of space is something that I track personally, have been tracking personally for the last 10-12 years and have a strong understanding of you know what is happening there.
Speaker: ah We have a fantastic team ah internally. My partner, Shridhar, also an engineer, but you know more from a ah addresses or looks at this from a more finance and structuring angle. And then we have a team of ah ah you know investment managers and analysts.
Speaker: ah Above that, what we have is an advisory board. ah Two people ah the advisory board comprises of. ah One is Professor Ashok Junjanwala, very well known in you know in the deep tech space. He is known as the father of ah the deep tech ecosystem in India and someone who is supported ah you know that that momentum of deep tech happening in India right at the IIT Madras Research Park.
Speaker: ah He was the guy who said, you know, India, we do great R&D, but there's never been the right commercial outcome to it. People do R&D from the lens of going and, you know, attending conferences and presenting a paper, but not really focusing on the commercial outcomes. And, you know, 15 years back when, you know, the whole research park ecosystem at IIT Madras was set up and, you know, the momentum that has generated, I think they did a valuation of companies last year from that ecosystem. They put that number at 55,000 crores.
Speaker: Now, that is real, you know, commercial impact at the end of the day. And, ah you know, someone like ah Professor Ashok, who was very closely involved with the gestation of multiple companies ah that came through that ecosystem, the most recent one being Ather, which IPO'd and the market really seems to like, you know it's hitting new highs. ah Companies like Medibuddy, Hasura, Exponent Energy, Detect, ah multiple companies that have, you know, created real value.
Speaker: have come through that ecosystem. ah The other person we have on our advisory board is Vinod Dam. Now Vinod again is you know known as the father of the Pentium chip. 30 years back, he created the Pentium chip at Intel and has been very closely involved with the growth of the semiconductor ecosystem. Today he sits on the board of ah the India Semiconductor Mission, the advisory board of India Semiconductor Mission, which is this ambitious program to develop the semiconductor ecosystem in India.
Speaker: And he sits on the design linked incentives committee where they can give grants of, you know, to two, three million dollars today to fabulous semiconductor startups. And so between the two of them, we have fantastic reach. Their own knowledge and understanding is great. So...
Speaker: From what i understand, one way of evaluation of deals and due diligence of deals is the people that you have with you, either as co-founders or in your team or as ah advisory board members. um Is this the primary filter with which you decide whether to invest or not? Do you look at things like, say, TAM, what is the size of the opportunity or the profile of the founders? or yeah Are there other things in your checklist that these at least three out of five, if they are yes, then we invest? you know i mean like I'm just trying to understand how how ah a company goes through the process and gets invested in at Blue Hill.
Speaker: It makes sense, Akshay. So what ends up happening is it's not just the people, but it's also the lens. um So ah we we are looking at ah you know, the problem statements, you know, how unique they are, how foundational they are. The biggest challenge, I wrote about this recently on LinkedIn, there are in India, you know, there is now a ton of money that is being ah provided to startups and to funds that are, you know, addressing the deep tech ecosystem.
Speaker: In fact, I think earlier today or yesterday, the second 10,000 crores from, ah you know, which has been allocated to SIDBI, you know, there is a clear focus on development of that capital to deep tech funds.
Speaker: And similarly, there is this RDI pool of capital, 1 lakh crores, you know, that has been announced. And then, you know, they've asked for ah applications for second level fund managers ah like myself. And then there is, you know, startups also, which would be addressed through the RDI capital. And so there are now There is now a lot of money that is available in the ecosystem. But the biggest challenge is to find the right people, you know, who are solving real problems vis-a-vis the people who are integrators, people who are sourcing stuff from, you know, across the border, ah you know, from the US, from China, integrating it in India and reselling it.
Speaker: right and that's become the biggest challenge you know for people who are you know administrating some of these pools of capital to identify who the real builders are and you know to identify who the you know quote-unquote pretenders are so that is a lens that you know we need to use uh you know fairly strongly a lot of that has to do with the founders uh What is you know the problem that they're addressing? Why are these guys the right people to address that problem? and Why is it a pain that you know they decided they should be the ones embarking on this journey to solve that? And you know that's apparent that will become apparent very quickly when you interact with some of these founders. And I say this as an investor, the most difficult thing that a founder would end up doing is to take money from an investor, right especially an institutionalized investor.
Speaker: Because at the end of the day, ah it's not my money that I'm investing. I'm getting that money from someone else. And there is a short clock running the minute I give them money. Actually, not even the minute I give them money. The minute I collect that money, there is a short clock running. I'm responsible to return you know what is called IRR.
Speaker: And so that's always going to be something that I'm going to you know ask founders about. And so you know I would you know even go go so far to say if you don't need to raise money, you probably shouldn't. right It's not for everybody. It's it's not ah not everyone likes someone looking over their shoulder, asking questions, asking you know why you're doing things in a particular way.
Speaker: ah The other thing for us is see we're looking for people who... you know And it's it's very hard to judge in over you know three, four, five, six meetings. But you're looking for people who are you know who have that ability to persevere.
Speaker: And generally, we take our investment decisions over a period of time. you know Even though mentally we would have conviction you know from one to meetings, we will still take multiple more meetings to add to that kind you know ah that conviction. Right. I might get to 90% very quickly, but to go from 90 to 100 will take maybe five, six, seven more meetings.
Speaker: Right. And, you know, you just keep adding that two, three percent conviction every time you have that conversation. And, you know, that is where you get a feel for the quality of the founder. You know, are they really going to persevere through this? and So we are trying to evaluate ah for all of that. But the primary lens would be tech. The secondary lens would be the founder. Right.
Speaker: And then everything else kind of, ah you know, helps in putting, ah you know, logic around the first two. The time and, you know, I mean, it it it is it is definitely something that we would think about, but we are very early in the journey. Generally, Akshay, we're probably the first institutional investor, if not, you know, among the first few that come in. And, you know, we are open to ideas where we are the founder themselves are looking to create new markets.
Speaker: The first company that we invested out of this fund is building a reusable 25 ton payload, completely reusable launch vehicle. I believe you you spoke to the founder as well, right? Yeah. Manun Ayer, he was here on your podcast, Etherelex. Now you have to be fairly crazy to decide to you know embark on solving that problem statement.
Speaker: ah The easiest, I mean, the most common question, you know, you you would have, you know, any founder would have come across, you know, let's say in software was a, you know, why can't Google do it? Or why can't Microsoft do it? Or why can't, you know, 50 other, you know, large software companies do it, right?
Speaker: And... It's because they can't. they don't They don't see the problem statement. It's not as important to them today or they're just so caught up in what is printing money for them that they don't see this as a problem that needs to be addressed. And you know it's always you know a new wave of startups, a new wave of companies that come and change what looked like a completely you know ah unsurmountable business before this. right Google was untouchable until it was.
Speaker: right Google was ah printing money through AdWords and you know whatever else, you know ah all the different ah ah sources of revenue, YouTube was throwing money. And then you know people suddenly yeah you know started using ChatGPT and Claude and your results appear within you know your conversation with these guys that you don't even open, you don't click on links.
Speaker: yeah Google now starts throwing you, it gives you a summary, you ask a question, it gives you a Gemini summary right there, such that you don't even open the first three links that show up. And so they are now having to change their business model.
Speaker: And that's something that ah you know is going to happen all the time. And that's what you know that's the thing that we're looking at. We're looking at founders that are super passionate about what they what problem statement they are looking to solve and are crazy enough to look at such problems and you know go and attack it.
Speaker: And you'll you'll get a flavor for that. you know you know once we even yeah Even with my own portfolio, ah you know when we talk about the type of problem statements that you know we've invested in, you'll you'll start seeing a pattern with some of those.
Speaker: Is there never a situation where you genuinely feel that there is no time for this? The founder is very passionate about the problem and he has the tech talks. Yes. Okay. A hundred percent. A hundred percent. What's an example of it? nice it's It's going to be very hard to give you an example because it's a live case and you know the the founder might probably end up watching this. So, see, there is there is ah there is a you know there are founders which are looking at very niche problem statements, within their own ecosystem, they're just so close to the ground. you know They feel it to be such a you know ah you know like earth shattering problem that you know within their own ecosystem, they see it as a big problem statement. But as somebody who is not so closely looking at it, you know, with a magnifying glass, you know, we're looking and saying it's too niche, you know, for us as a fund to invest in. You know, it might be revolutionary tech, but, you know, there is still that tech risk that we have to underwrite. And then you go into the market and then you realize that that market is really small. And that's not something, you know, even if it's a 30, 40, 50 percent jump, right? It's that 30, 40, 50 percent jump that you're doing over probably the next three, four, five years.
Speaker: And by that time, the existing tech will also keep improving. right It's not that the existing landscape is going to stay exactly the same. They will keep adding that 10-15% improvement, refinement that is happening.
Speaker: And so some of these guys are closely wedded to ah you know their own tech. And what you know when when we end up liking the founders and you know we have a good feeling that these guys are good founders, I think it's in our best interest to ah sit down with them and you know try to open up the time for them. right Say, hey, you know while we really like what you're doing, maybe this should be one part of your broader story. It shouldn't be the whole story. How about we look at, ah you know, something, you know, instead of a thousand feet, how about we look at it at a 5000 feet? You know, can you address a broader ah problem statement there?
Speaker: And, you know, this the one that you really like would be one of maybe five verticals within that. right And that's that's how you know we're looking at it. Definitely, I can think of you know two, three founders that you know in the last few months I've you know had conversations, ah super passionate, super excited. ah the The great thing, Akshay, nowadays we're seeing oh fantastic crop of founders coming from you know our own, let's say, defense agencies, the DRDOs, the ISROs, the HALs, scientists D, scientists E type of profiles who are stepping out confidently and saying, hey, you know,
Speaker: We think we can build outside of that ecosystem. We think we can raise money. you know We've been working on you know this gas turbine for the last 15 years. Or you know we've been working on this you know missile system for the last 10 years. Or you know we've been working on Gaganyan for you know five years. And I think there is this use case that I can now step out and start building in private.
Speaker: And the fantastic part today is that There is a a use case where these guys can step out. An organization like InSpace will support you, you know, with your build out, right? ah Space as a startup ecosystem was largely non-existent seven years back, right? And today we have like 350 startups in this ecosystem today, ah you know, under InSpace. Guys building from, you know, launch.
Speaker: What is InSpace? InSpace is this umbrella organization associated with ISRO. ah So they help in ah connecting startups to ISRO, to commercial use cases, to other space agencies, ah to the government, to help you you know get across.
Speaker: Like in an industry body, like a FICCI, NASCOM. It's not an industry body in that sense. It's it's connected to ISRO directly. So it's not like a public-private body in that sense. Semi-incubator kind of a... Correct. Correct. it that would That would probably be the best way to describe it. And today there are like 350 companies under that. Right. And they also help you with giving you grants. You know, they give you matching grants. If you raise money outside, maybe you raise a couple of crores.
Speaker: um There are different programs in which they will give you couple more crows as a grant. And that can help you ah continue your build-out process. right And today, very diverse problem statements within the space ecosystem, like I said, launch vehicles to you know earth observation, to radio communication, to light communication, to propulsion, to thrusters... So, you know, um what changed? Because I remember when I would interview deep tech founders way back, say, 21, 22, the funding ah amounts, funding rounds for them used to be significantly smaller.
Speaker: and today it's changing. So what's the trigger? what What has changed over the last couple of years? This has, I think this is mostly like last two, three years where this has suddenly changed. What happened?
Speaker: So multiple things happened, I think, at some point in time, just tour de force, you know, there's just a tidal wave crashing in and, you know, it ended up moving ah moving things along. See, I've been investing in this ecosystem, this deep tech ecosystem from like, what, 2016, right? I've been writing checks in the IIT Madras ecosystem. I've done seven investments in IIT Madras ah starting 2016. The amazing thing that has changed over the last few years is that the government has woken up to the problem. you know and When I say government, I mean, you know, the states, the center, you know, everyone across the ecosystem has come to the realization that we need to build sovereignty in technology. So this like a triggered by the China factor?
Speaker: To a certain extent, I would say ah not specifically China, I think there is, you know, geopolitical issues, our dependencies and, you know, defense imports and, ah you know, cameras and semiconductors and, you know, space, right? Everything is just so heavily dependent on, you know, outside, right, quote unquote, whatever that outside might be, wherever that outside might be, right, that is China, and that is the US. And the problem actually is that, you know, the friend of yesterday is today's foe and the friend of today is tomorrow's foe, right? It's, ah you know, the the geopolitical lines seem to change as we are talking, right? Every every hour, every two hours, there is something new in the Strait of Hormuz, you know, there is something new in...
Speaker: ah Ukraine and Russia, ah the the the paradigm of ah geopolitical relationships has changed significantly, you know, as as we know it. And I think it's ah it's a good place to be. i think, ah you know, it's important that we build that sovereignty out.
Speaker: ah the the I think there there is realization. I'll give you one example. and ah All our cameras, right, CCTV cameras have historically just come from China. And, you know, ah we might build the housing in in India, maybe the plastics and, you know, the cover, you know, all of that might be built here.
Speaker: The bar will come from, you know, one of our factories. Right. But at the end of the day, the fundamental camera is still coming from China. Right. And the chip is, you know, being manufactured, let's say, in Taiwan or Korea or something like that. Right.
Speaker: Now the government realizes and you know, there was this very stark meme that was going around where DRDOK under there was a you know camera which said Hikvision and someone you know someone had taken a photo of one of the generals with a Hikvision camera in the background. And then there was this question that, hey you know, that's ah a huge backdoor. Hikvision is a Chinese company.
Speaker: Hikvision is a Chinese. I mean, all of them are, right? Dawa, Hikvision, most of these guys are. Even the ones that are, you know, today, quote unquote, Indian are not completely Indian in that, you know we're not building the chips for it. They're assemblers. we're assemblers, right? And so ah there is now this realization that, you know, we can't be dependent on that. I don't know if you ah remember seeing this ah on on maybe Reddit or something. So there was this guy, you know this, ah your smartwatch, right? Tracking your runs. and There was this run that was tracked on Brava, right? Out in the middle of the ocean. There was like, you know, circles like, you know, out in the middle of the ocean. And then it just turned out that there was this Marine on a ship that was supposed to be hidden, who did his runs and then posted his run to Brava.
Speaker: And then now this location of this warship was just visible to everybody just because of, you know, the small loophole. Right. So we, You don't know where these gaps are, where these loopholes are. you know if youre If you're wearing one of these watches or you know you're wearing a smart device, everything is communicating to yeah the cloud, to the internet. Everything is uploading data. Everything is sending information. And the last thing you want is to be in a place where you're in you're in a situation of war and the cameras are being used as backdoor to watch what you're doing or to hear what you're talking about. Similarly, you know, there's a very stark use case in the drone ecosystem, right? A majority of our drones, are you know, see, we say some in some cases we're saying that drones are built in India, but largely they are being assembled in India.
Speaker: Hmm. The flight controller largely comes from outside of India, from China in some cases, ah comes from other places as well. And the flight controller is the brain of the drone.
Speaker: The flight controller can communicate, right? It can speak to, you know, the GSM networks, GPS, satellites, right? It can connect. And so while the drone might be Indian in name, when it comes to war and, you know, if it you know came from a Chinese vendor and you know there is a backdoor to it, at you know when push comes to shove, it's no more your drone, it's a Chinese drone.
Speaker: And that's something that the government and the defense agencies are waking up to the reality of And they're saying, hey, you know, we need, ah you know, the flight controllers to be Indian. We need to see that. We need to see the source code. We need to see the chips. We need to understand that. In fact, after Operation Sindhur, there was an emergency procurement of some 852 drones, which was issued fairly quickly. And then it was revoked back fairly quickly again because they realized that the flight controllers were not being built in India.
Speaker: And so there is that realization that is happening. There is and it's happening in waves across different industries. Very clearly in the semiconductor industry, ah the government understands that, you know, we can't be importing every single part of our semiconductor ecosystem, right? Anything to do with semiconductors for India is imported. right right we don't have the We don't have the ability to build a chip. right The foundries don't exist. The first one ah hopefully going to go live in early 2027 is the Tata foundry.
Speaker: um you know Going to come online fairly quickly. lot of money that has gone into it, lot of government money that has gone into it and you know the Tatas also. But that is the only foundry that India is going to have. right The other one is ah you know a government foundry at SEL Mohali.
Speaker: right And that's a very old one, I think building at 180 and 130 nanometer nodes, which are dinosaur age nodes. no one No one is really building meaningful chips at those nodes.
Speaker: um The Tata semiconductors, you know they're saying ah they'll start at like a 28 and all and they will you know eventually come to a 12 nanometer. But you know that's a real meaningful step in the right direction.
Speaker: And then to support it, we have you know the OSAT, Outsource Semiconductor Assembly and Testing Ecosystem. ah that is also coming up. ah There are, I think, six, seven fairly advanced projects across the country, ah you know, that are happening.
Speaker: There is the Keynes one. the This OSAT is a service business like they serve as fabulous So what happens is the chip is designed in India by a startup, let's say the chip design is sent to TSMC in Taiwan.
Speaker: The TSMC will send you the wafer, right, which is like a, i mean, I did that, but it's not that big, it's usually eight or 12 inches. And, you know, it has to be cut, you know, the the wafer has to be cut.
Speaker: And that that is the work that, you know, people like Keynes and, you know, CG Semi and HCL and, you know, multiple others are now saying, hey, we will get involved in that.
Speaker: and Today also, you know we end up having to send, even for that process, you end up having to send it into different countries, Korea, Japan, ah you end up sending that there and then you get back you know the boards themselves into India. right And ah today there is that ecosystem that is building up, which is saying, hey, we will do some of this here.
Speaker: And that's a much smaller investment. like I think the Tata foundry, they're saying $9-10 billion dollars in outlay. The OSAT will probably be 200, 300, 400 million dollars in outlay. And which is why, you know, lot of the people who initially showed a lot of enthusiasm for saying, hey, I'll also do a foundry ended up not embarking on that journey because just a lot of capital to be put together.
Speaker: you know, put it into the ground. And I think people are saying, hey, let's first explore with those sad piece and then see how that goes. And so even there, right? So um today, every single chip, your chips in your cameras, your chips that go into your battery management systems for your vehicles, your weighing scales, pressure sensors, simple ones. These are not complicated chips.
Speaker: The chips are still coming. You know, the design is coming from China. It's being printed out in, you know, probably, you know, hundreds, hundreds of thousands, yeah you know And then you know it's shipped to India as part of a weighing scale or a camera or whatever. And today there is that hopefully you know we start building some of these chips ourselves. There are multiple fabulous companies that are saying, hey you know I can build out that basic chip that is needed for a camera use case or ah or for a BMS use case. I can build that out ah locally.
Speaker: I have the design for it and then I can you know tape it out with TSMC or PSMC or Global Foundries or you know one of the big guys and you know we can build that out. So there is that layer of startups coming through the ecosystem. Then there is this other layer which is saying, hey, I will address more complex problems, more fundamental problems, more foundational problems where we're trying to beat, let's say, use cases in hyperscalers or high precision computing. Right. Can i now build something fundamental in that space? And that is also happening. Interestingly, actually, there are guys, you know, X Intel, you know, X Qualcomm, X ST Micro, guys who have been in the semiconductor ecosystem within one of these large, you know, global semiconductor players. you know, been there, led teams, you know, built chips, AMD, for example, ah you know, built built within that ecosystem. And now, you know, having the confidence to step out and say, you know, I think the ecosystem is mature enough for me to build a startup in India.
Speaker: you know we have, you know, a couple of lakh engineers working in the semiconductor ecosystem alone, you know, in in India, largely around Bangalore and a little bit, I think, in Hyderabad. And, you know, these are now guys who are willing to step out from cushy jobs in those spaces. And, you know, the the programs like the India Semiconductor Mission, the Design Link Incentives Program, right, where they're giving you money to pay.
Speaker: ah you know There is a tax that you pay when you're a semiconductor start startup, right? Like a Google tax, you know, if you're you're playing this where you're paying this app store tax, right? that or you know Play Store tax that people launching apps have to pay. No matter what, any revenue that comes, you know money goes to Apple or to Google. Similarly, ah in the in the semiconductor fabless space, right you have to pay your tax to use these EDA tools from Synopsys or Cadence. ah These are tools that you know are ubiquitous because only when you design there does a TSMC understand what you have designed. right they They are comfortable if you use their tools and the format of their tools. You submit the design, it's easier for them to process. So you end up having to work with one of those and it it can be really expensive. right you you know You're down $200,000, $300,000, $500,000 in just...
Speaker: you know, ah cost, you know, licensing cost to one of these guys. And today the government is saying, hey, I will pay pay for those. I will give you, i will reimburse you if you end up, you know, paying a cadence or a synopsis. These are software, this EDA that you're talking about. EDA is just design tools, right? It's just, ah you know you know, a simplified way of thinking about it is if you know AutoCAD, this is probably, you know, an AutoCAD for semiconductor design, right? Got it.
Speaker: um You know, and these are these are tools that are indispensable. ah Interestingly, there are startups which are trying to build tools to disrupt these spaces also. Very interestingly, three, four companies in India which are saying, hey, I will take on the EDA a sector itself.
Speaker: And you know that is that is one interesting space that seems to be opening up ah as well. But yeah, so the the the, I mean, just rewinding back to the the original question is that there is now significant support from you know the center, from states giving you land, you know states saying, you know if you put in 100 rupees in capex, I will reimburse you 35, 40 rupees of that. The state saying, okay, know if you're actually doing manufacturing, I will ah you know take away GST. I mean, I will take away the need to pay GST on it.
Speaker: And so a lot of support that is coming in, um giving these guys confidence to build here. See, what ends up happening, though most of the global GCCs are somewhere in India. right The biggest you know capability centers, you know what they call, are located somewhere in Bangalore, Chennai, Hyderabad, Pune. So if you take like a Rolls Royce or a Walmart or a Qualcomm or a Broadcom,
Speaker: or you know Target. right All of these guys have some sort of capability center in one of these cities. And people have been using their tech you know bent of mind to be solving problems for these guys sitting out of Bangalore anyway, because it's cheaper to hire an engineer for a Rolls Royce.
Speaker: to to be you know hiring a guy who understands an injector design for a gas turbine. He's sitting in Bangalore, right? The largest Rolls Royce GCC is in Bangalore. GE sits out of Bangalore. Collins Aerospace sits out of Bangalore. So today, a lot of these guys anyway, as a large organization are utilizing that arbitrage. Think about it, you know, a startup founder can anyway, he's anyway from here, you know, he understands the ecosystem, has relationships out here. He can hire some of those same talent to build a startup.
Speaker: ah So they get to build it at much lower cost than they would, you know, if they were building it out in the West. And the fact that there is now government support is helping. The other part that has shifted is investor interest. right In twenty sixteen seventeen it used to be the alumni of IT Madras largely who were investing back into the ecosystem. um you know Professor Ashok was obviously well connected into you know the government and you know a lot of those ecosystems. But initial investor interest was from the alumni. Today, given how that ecosystem has panned out, you know like I said, 55,000 crores of value, a lot of investors are seeing that, hey, you know we need to be present in this ecosystem or we're missing out.
Speaker: And multiple such ah spaces have opened up at IIT Bombay, IIT Delhi, Kanpur, Kharagpur, other institutions, ISE Bangalore, for example, IIM Bangalore, all of these guys now have incubators and support and different programs to support startups. And so that is you know something that is definitely helping the build out.
Speaker: And finally, I think the fact that people have seen outcomes from investing in these spaces, right? See, investing and paper profits are nice. When you actually have an exit is when people say, OK, I can actually make money from investing in this space. What are the outcomes that have happened in deep tech?
Speaker: I would say the most recent one would be Aether. I think that's been a fairly large one. IdeaForge was in that ecosystem ah you know fairly early. But today there are now these larger, ah let's say Series B, Series C type of capital that is coming in.
Speaker: Atomberg is a big, you know fairly large outcome that is you know happening you know in the broader ecosystem. you know They started off, IIT Bombay guys started off building these motors, more efficient motors for fans. right Now, I mean, just think about how crazy that is. Right. They took on something as mundane as a fan. Yeah. ah And said, hey, you know, we will take on the existing, you know, Cromptons and Usha's of the world where nobody would have thought, you know, You're you're ever going to be able to change how people perceive you know their their purchase of a fan and you know the efficiency of it. They were charging a premium to sell to to sell fans. If you if you recall, right a you know your Usha fan will be $2,500. These guys will be charging $3,000 or $4,000 to buy a fan and they're still sold out.
Speaker: And I remember having a conversation with the founders, Manoj and you know, the they were like, see, I can reduce it, but I don't have the capacity if I reduce the price and you know, I get that much more demand. I don't have the capacity to, you know, build as many fans as you know, I might end up getting orders for and I'm just happy to sell, you know, in that ecosystem today. And today they are branching out into so many new areas. right They're now selling you know water purifiers and they just announced one Atomburg innovation company where they're hiving off you know specific innovations within the company as a separate a unit.
Speaker: And there's a company called Sedimac, which had an outcome as well. So there are now clear visibility. There's now clear visibility around an investment in some of these spaces giving a large outcome that then gives confidence to investors who are always looking for comp comparables. right They're saying, OK, if I invest in this, it could you know end up like an atom bug. Or if I invest in this, it could end up like an ether. Where did you raise from? I believe your fund size is about 150 crores. That is correct. So it's a 350 crore fund with a 50 crore green shoe.
Speaker: ah We're almost done with the 350 crore part. ah My largest investor is Sidbi. We have a 70 crore commit from them. Additionally, we have ah the Kerala government, which has given us a commitment as well. We have the UP government, which is giving us a commitment. um And the rest of it has largely been, you know, a very democratized race, ah been raised from, you know, my connections, my partner's connections, people who have co-invested with us in, you know, the previous journey I mentioned.
Speaker: Like, ethanies or institutions? Institutions. It's no. So the institutions are only the ones I mentioned, you know, the SIDBI and, you know, some of the state government funds. ah The rest of it is HNI's professionals, interestingly. And we have very interesting professionals, the India head of Sengobain, the India head of Daimler-Benz, people who understand, you know, the technology, who understand, you know, what it takes to build startups in this space is backing us.
Speaker: Saying that the opportunity in the next 10 years, 15 years, 20 years is going to be huge in these spaces. And we definitely need to underwrite, ah you know, people like you who are trying to ah find outcomes here.
Speaker: And so we have some very good professionals who have invested with us. There are H&Is, you know, there are family officers which have invested. And it's it's ah this given that this is our first fund, it's been a very democratized raise from across the ecosystem. i have investors, largely Chennai, Bangalore, Hyderabad and Bombay, ah but a smattering of a few other cities as well.
Speaker: ah But yeah, has been the institutional piece has been probably yeah about little less than one third of it. Okay. ah I want to kind of understand... ah What China did right and if India can replicate that? Because you know in terms of frontier tech, deep tech, whatever you may call it, China is now probably at par with the US.
Speaker: So what did they do right to have such a strong ecosystem? Like you said, cameras, they dominate. Electric vehicles, they dominate. um What are they doing right? So I will disagree slightly with that assessment, Akshay. I don't think they're at the cutting edge of that tech, right? They've just been good at copying a lot of the tech. ah One very interesting, you know, like observation for a lot of people in the ecosystem is they've been able to copy the chips. They've not been able to copy the machine that makes the chips, right? There's a machine called a lithography machine made by this large Dutch company called ASML.
Speaker: And, you know, they there's there's this lore that goes around saying, you know, yeah it's it's like a machine that probably spans like two, three rooms, right? And, you know, you take apart one screw and the entire thing falls apart and, you know, they're not able to reverse reverse engineer it. I think they've been very good at, ah you know, reverse engineering a lot of these ideas from the West, you know, copying it because, know,
Speaker: by virtue of them being the you know factory of the world, they get the blueprints for everything and you know weak IP protection and you know they end up in a place where they are able to copy it you know put a different label on it or you know marginally change the design or structure of it and you know go to market. simultaneously. If you look at some of the BYD cars, you couldn't tell some of them apart from a Tesla car. It looks almost exactly the same. The engineering of it is exactly the same. I think what has been done very well is the state support. government government support in that ecosystem has been absolutely fantastic. right They are willing to underwrite you know losses in the short term, understanding that that creates a moat in the long term. And know that just created insurmountable moats in a lot of these spaces today. Like you said, today, if you want to buy an EV cell,
Speaker: you're almost, you know a lot of your procurement will be focused on China. and There is Japan, there is Korea, there is maybe Leclanche in Switzerland, but otherwise, you know, it's very China-centric if you're looking at, you know, procuring cheap sales, you're looking to, you know, go to market fairly quickly. ah The chip design, I would say again is, see, these are very basic chip designs. You know, a camera chip is not a very revolutionary ah piece of tech or, you know, let's say a chip that controls your weighing scale, right? It's basically doing, you know, some basic calculations internally. ah You know, it's not not complex tech. It's just that they've, you know, been able to punch it out a million times over at a very low cost. right And I think it's slowly happening in India that that that piece is happening where the government is getting involved. Like said, even with the Tata semiconductors, you know, if 40-50% of that outlay is coming from the government, it's a fantastic thing.
Speaker: And that's happening across the foundry ecosystem. It's happening in the Fabless ecosystem. It's happening in the OSAT ecosystem as well. So the government is waking up and doing, i think, its part.
Speaker: ah The government is... you know devolving money to funds like me and you know saying, hey, you know you invest 2x of what I give you back into the startup ecosystem. right And that just ah has like a compounding effect downstream you know when more and more money gets gets into some of these startups. And so I think that that is going to be a big part of how this ecosystem really you know compounds. The biggest challenge, I think, Akshay, is that the money has to find its way into the right hands.
Speaker: i I talk about this on and on a recent LinkedIn post. you know is ah These guys who are integrating versus you know the guys who are actually you know inventing.
Speaker: The guys who are you know but doing the actual engineering versus you know guys who are bringing stuff across the border, like I said. And how do you ensure that the money goes to the guys who are actually inventing and not the guys who are integrating?
Speaker: um There's a lot of money that is being allocated to this ecosystem. I'll give you one example. right ah there was this There was, I would i say now, there's a subsidy called the FAME subsidy, ah you know which was being given to electric vehicle companies, ah guys who are building EVs, OEMs.
Speaker: ah Based on the battery size, you could claim you know a certain amount of subsidy. And there was a requirement that there has to be a localization component over a period of time. that, you know, the ecosystem needs to be localized. ah What you end up building for in that vehicle has to be more and more procured in India or built in India.
Speaker: So, Multiple people procured that you know ah got that subsidy. Some of the big names in that, Hero Electric was one of them, Okinawa was another. And you know for five years on, and I think the government had given about 1500 crores of subsidy. government woke up one day, realized that ah neither of these two companies and quite a few others were actually building in India and said, I think you you know you've been hoodwinking me, you know give me back that money. and you know the government the company ran into problems, you know they had to shut shop or you know they had to return that money.
Speaker: And while it's a good thing that you know they ah eventually woke up to the problem, the youth the the the thing I always point to is you know the opportunity cost around, you know let's say the actual inventor, let's say the Aether not having got that support, they could have been much, much further ahead in the ecosystem today. They're already you know built what is the best electric vehicle today. The market realizes it. the And when I say market, I i mean, you know both the the share market and the actual user market both realize the quality of an Ather.
Speaker: And today, you know if they had access to that 1,500-clo subsidy, it would be far ahead of the ecosystem. Somewhere along the way, these quote unquote integrators broke the trust of some of the users. Right.
Speaker: ah People then suddenly said, OK, if these startups are not doing a good job, maybe I need to go back to, you know, these erstwhile OEMs, the TVSs and Bajajs of the world. And purely by virtue of that, you know, there is a market for you know, TVS and Bajaj in the EV, OEM space by virtue of the fact that they exist in the ICE space as well.
Speaker: Right. And that trust deficit, you know, once broken, it's very hard to gain back there. And I think that's something that, ah you know, we need to control for, you know, with all of these pools of capital that are coming in, we need to ensure that the money is doled out very carefully, very judiciously and controlled outcomes, controlled environments, there should be very clear, defined protocol to you know devolving that money and ensuring that it teaches the right people.
Speaker: What kind of risks are you comfortable taking? For example, in deep tech investing, there could be a technology risk, there could be a commercialization risk. So what kind of risks are you okay taking? What kind risks are you not okay taking?
Speaker: See, ah I think at the stage we are coming in, we are underwriting pretty much all the risks, and you know, that are possible, right? VC is that game where I'm, you know, underwriting very heavy technology risk. but You come in at 3 Series A?
Speaker: Like, what stage do you Yes, I would come in between seed and pre-series A, usually between TRL 3 and TRL 4. And just can you elaborate more like what would be a TRL 3, what would be a TRL 4?
Speaker: Yeah, so for us, what we are looking for is something that is beyond, ah you know, paper, you know, ideation. We're looking for some sort of ah proof of concept, some sort of testing that has happened in a lab, very early you know interest from a commercial user, very early interest from people in that industry, maybe who have ah you know seen the status quo in that industry over 20 years and seeing ah hey you know how this disrupts what already exists, what has been working and you know potentially changes the paradigm there.
Speaker: So we are looking for something that is beyond the paper, right beyond you know diagrams on a board. We're looking for something a little beyond that. ah Generally would not invest just purely on an idea because we are coming in where we are putting in at least a million dollars on an average, could go up to probably $2 million dollars in a first check.
Speaker: And, you know, at a seed stage, pre-series A stage, you know, we're coming in where we might end up owning 15-20% of that company for that investment. And so, commensurate to that there has to be you know the the risk has to be commensurate to the potential you know reward that you know we are underwriting there um what we what would be non-negotiable is like I said earlier you know the founder you know ah quality, the the pain that that founder has seen that has you know put him on this path, ah you know the ability of the founder to tell that story also becomes fairly important because the the challenge in this ecosystem is not everybody understands the nuances of the tech. right
Speaker: If someone comes and says you know like in in you know your own earlier conversation, with Manu Nair, there are not too many people who would understand the depth of the problem statement that ethereal exploration is solving.
Speaker: A of people will just come to you and say, you know I thought Israel already does that. I thought, you know why couldn't SpaceX do that? And you know it goes back to my old you know Microsoft and Google statement, you know the guy with the you know, trillion dollar valuation could probably do this. Why does it you know why will a tiny startup from, you know, a country that doesn't have very serious pedigree in this space? I mean, we have decent pedigree, but, you know, nothing compared to a SpaceX, for example. Why would they be able to solve it? And I'll just give you some perspective on, ah you know, the the gap that has been created, right? And ISRO on an average has launched, you know, probably six times every year over the last 10, 15 years, right? On an average, six times, right? ah
Speaker: SpaceX has, ah do you want to take a guess, Akshay, how many times SpaceX launched last year? More than 50. I'm assuming at least one a week. Yeah, they're doing one every three days, give or take. Okay.
Speaker: And they're they're pushing to, you know, that to be even higher, one every two days ish type of thing. And that is possible because of the fact that they are reusing.
Speaker: Right. They are reusing the stage one, that piece that cuts off at about 70, 80 kilometers falls back into water. is now coming back and landing and then is being refurbished and reflown.
Speaker: ah They now have ah cohorts of stage ones. They have a stage one that has flown 35 times, another stage one that has flown 30 times, another stage one that has flown 25 times. so they're just They're running out of parking space for these stage ones. right And the question is, hey, you know, if I'm able to recover stage two, then, you know, the story becomes, it's like, you know, aeroplane flight, right? It's like our, you know, Airbus and Boeing, you know, I just get to reuse that vehicle every single time, you know, hundreds of times. And then I get ah unit economics becomes very attractive at that point in time. And that's what these guys are trying to achieve. Now, the the question is,
Speaker: this guy, the founder needs to be able to go and articulate that story, needs to go out and be able to sell the story to an investor, sell it to, see it's one thing to sell it to me, you know, let's say, you know, I'm a little bit more in the weeds, ah you know, let's say in this specific case, had a very strong understanding of what these guys were trying to build.
Speaker: But how do you then go and you have to, you know, at some point in time, sell to people who are not technical, who are, you know, people with checkbooks, but do not understand the tech. Right. And how do you sell to some guys saying, hey, you know, I'm going to create this whole new market. It's SpaceX today, ah you know, does, ah you know, what is has brought down the cost of launch to you know, a few thousand dollars per kg, right? I think five, six thousand dollars per kg.
Speaker: I'm going to bring that down one tenth. Right. And that's a pretty crazy thing to say. And so there is a fine line there. There is, you know, ah guys who are only selling the story and not really solving the problem. And then there are guys who are too technical, who don't have that ability to you know, go out and ah sell the story and, you know, ah the opportunity set. And I think there is, you know, you have to find that balance between, you know, this x extreme and that extreme where you're, you know, building ah the tech as well and you're able to sell that story. So one of the things I think we find fairly critical is to have a
Speaker: a little bit of a diverse team. We prefer not to underwrite single founders, you know, guys who are just one guy, where the guy is also doing, you know, let's say the the rocket design, and then also paying his Eitel bill, right at the same time, right. And I think that's something that we are very cognizant of at the time we we're underwriting some of this risk.
Speaker: I'd like to understand some of the ah like the most prominent names within, say, semiconductor startups. like like Who are really startups to keep an eye on or in defense tech or in space tech if if Maybe for these three sectors, if we can just go one by one, then tell me what what are the names to look out for?
Speaker: Sure. I'm obviously going to plug my own companies, but i will I will try to keep it a little fairer than that. So we've done now we will we'll end up doing four companies in semiconductors. um We've done two already.
Speaker: We did one company called Sofrozyne and both Sofrozyne and the other one is called OptoML. Both are in Bangalore. um One of them is doing ah ah semiconductor. i mean, they're doing chips for wearables, right? So now today, like if you if you take a, mine is a Garmin, if you have an Apple watch, right? These watches are fairly complex in that they can sense multiple body, you know, parameters indicators, right? They can take in a lot of parameters, right? Your skin temperature, your heart rate, your EEG, you know, a host of other things. Hello.
Speaker: what ends up happening is the ecosystem is such that it's designed you know like a market. right a Garmin wants to give you everything at very high precision.
Speaker: right So there's they they're going to pick you know the top of the line chips to do your SPO2. There's a separate chip for that, to do your skin temperature, there's separate chip for that, to do your pulse as a separate chip for that. And the reason the ecosystem exists like that is because the lower end guys will want to give you probably only the skin temperature and your heart rate. They will not want to get into some of the more nuanced ones. And so it gives them that option to pick and choose. And what these guys are saying is, say you know, and given my understanding and background in, you know, the bio med space and, you know, my ability to understand some of these signals, you know, in a very noisy environment, see, there's movement, there is sweat, you know, there is hair in between, right? How do I understand the signal?
Speaker: These guys are saying I can do one single chip to give you all seven, eight parameters ah at a cost that would be probably one tenth if you had to add these costs up individually, right? Initially, right, you would have one from analog devices, one from TI, and then you have to integrate all of this, supply power to this, you have to have a software layered on top of all of this these guys are saying hey you know I can just do ah one chip to you know rule them all yeah exactly as I was going there but perfect one ring to rule them all um but yeah
Speaker: So again, like fantastic guys, guys who've, you know, been in this industry, they've already built, interestingly, built some of these chips before ah much lower complexity ones, guys with backgrounds, PhD backgrounds, you know, have worked in the industry, have worked in the biomed space and, you know, decided to ah go out and, you know, attack this problem statement. um The second one is a company called OptoML.
Speaker: Now this the other extreme. So, Sofrozyne is building at 28 nanometer with TSMC. OptoML is building at 12 nanometers with TSMC and they've already built out. So the design has happened, they'll have the chip early in May.
Speaker: and I think and you know, ah maybe this will be debated, but as far as I know, they're the first startup in India to build at 12 nanometers. And I say startup, right? It's a 14 member team and you know, they've built out at 12 nanometers with TSMC, fairly big deal. For what use case? i'm I'm coming to that. So these guys are doing something called in-memory compute.
Speaker: and doing that in analog, right? So the use case primarily will be for inferencing hyperscaler chips, right? So for these AI use cases, but for inferencing, not the training. See, training requires very high precision. So generally you can't do training in analog. You can, you know, do it with digital.
Speaker: ah They're doing it in analog. ah But what specifically is the the the big deal here is something called in-memory compute. ah Again, one of the things I've written about and you know on LinkedIn fairly recently is that see in a lot of these AI chip use cases, 70-80% of the energy consumed by the chip is not going towards compute, but is actually going towards moving the data back and forth. So you had two very separate environments, you know storage and compute.
Speaker: And then you had a bus where you would transfer data from here, compute transfer it back. What OptML is trying to do is to bring some of that compute into memory itself and do ah what would be called a pre-compute and you know essentially save on having to transfer this data back and forth a little bit. And through that, they are trying to break certain performance benchmarks, performance per watt, performance per dollar.
Speaker: right And they are attempting to do that. Interestingly, the bus itself also they're trying to change, they're trying to use light ah to replace the existing conventional buses and that's a big deal. ah you know there is and There's a whole different use cases, whole set of use cases around light on semiconductors. In fact, ah from you can people are working on light in satellite communication as well. right and know Light is ah is probably something that you could you know call infinite bandwidth. You're not limited by you know the actual ah bus that you're transferring data on.
Speaker: And that is that is what these guys are attempting to do. We've already taped out with the TSMC. Like I said, it's fairly big deal to tape out at 12 nanometers with them, with TSMC, with fairly active and involvement of of the team there as well. Like I said, we were we're hoping to have the chips from them in early early May next month. Another interesting company, again, not my portfolio, but an interesting shout out I'm going to give. I really like the team and really like the ah use case is a company called Caligo out of Bangalore again.
Speaker: Caligo Technologies, guys building a very interesting, it's a change idea, it's changing the number system. They're using something called Posit to allow for high performance compute, HPC. ah This is more precision compute.
Speaker: And, you know, very experienced founding team ah already have built out, you know, already have taped out a chip at a a lower node and, you know, they're working towards you know, taping out at a higher node now and have proved out some of the use cases at ISE Bangalore.
Speaker: ah Interesting, interesting team. um All of these have been in the fabulous space. Interestingly, there's a company called Maker Microsystems, which is in the, interestingly, in the metrology for fab space.
Speaker: Now, metrology is a fancy word, ah broadly means measurement. um matt So, see the fab process, right, if you go into the foundry, right, the foundry processes today are becoming extremely complex, right. Eventually, if I don't measure that, and then I add, let's say, seven more processes on top, if the first one was wrong, all of, you know, you're just wasting, you know, on time in the in the foundry. So, you have to measure.
Speaker: So, there are two three ways to do this. You know, you can do a 2D measurement from the top of the surface or you can break the chip, break the wafer and, you know, look at it from a profile. so that's a destructive process. So, you have a scanning electron microscope. you have something called a transition electron microscope. You have an atomic force microscope, EFM.
Speaker: What these guys are able to do is something called a Acoustic atomic force microscope. So they are adding um ah the, so the AFM is actually your, you have like a gramophone record cut tip, right? And they are getting that tip to resonate and, you know, touching the the surface and they are able to read and then they are able to turn the tip.
Speaker: And they're able to give you a 3D view of the processes that you have done, wherein you're able to measure, ah you know, like an ultrasound, let's say you're able to measure, right? Today, if you do an ultrasound, they're able to measure the baby's development, the growth, you know, certain body parts, you know, hey, you know, is everything, you know, going as per plan?
Speaker: Similarly, in this case, you know, they're able to measure, you know, the processes as it happens within the foundry, they're able to measure and give you a very good reading of what is happening. And hopefully, the idea will be that they will it will improve yield of the foundry, right today, let's say maybe one of these lower processes yield is 80%, you know, can they push that yield up significantly? And, you know, they can improve the throughput as well. So this is again, a very interesting company in this space. um We're doing one in verification, I spoke about these, you know, EDA tools and all that, right? And the complex complexity around that, we're doing one that is building in that space as well. ah But you know, that's that's still cooking as we speak.
Speaker: um I'm going to switch gears into space. um So the the first company, i mean, that's the only company we've done out of this fund. And this fund is a year and a half old. Ethereum X is building, you know, reusable launch vehicles. Now, obviously, there is a podcast where Akshay has spoken to the founder of Ethereum X, Manu Nair. So that's there's and that's an interesting conversation. But like I said, and like I alluded to,
Speaker: a little earlier, these guys are building reusability of the second stage of the launch vehicle. And that's a really hard problem to do because by the time you are with in in the second stage, ah you don't have access to your air breathing engine, which you jettisoned and just come back and probably landed. um So I can come back from, you know,
Speaker: let's say a 500 kilometer orbit to a 100 kilometer, which is where, you know, the atmosphere technically starts, it's called the Karman line. I can come back from 500 to 100. But if I need to enter 100, I need access to my air breathing engine, which I don't have. And what these guys are attempting to do is to, you know, reuse, ah you know, the reentry heat, right, or the fictional heat that that exists, they're trying to capture some of that heat and use that as propulsion and know fantastic founding founders. founders ah They've actually now meaningfully built, ah you know, engine engines in that ecosystem. 80 kilonewton second stage engine, 1200 kilonewton first stage engine. Both have been built out. um
Speaker: the the The use case is for a 25 ton payload launch vehicle. Right. And that is large. Right. The Falcon 9 and, you know, file between the Falcon 9 and Falcon 9 Heavy, we're talking 16 to probably 22 tons of payload.
Speaker: ah We're talking... 25 ton payload, which is larger than that. And that's meaningfully ah you know large. I think the broader launch market ecosystem in India has generally been focused on the small launch launchers, right? 500, 700 kg type of use cases. ah But our use case, which is you know the 25 ton use case,
Speaker: becomes a lot more meaningful for launching constellations. right And the demand for space is largely coming from constellations. A Starlink today has 10,000 satellites in space. They want to make it 40,000.
Speaker: Kuiper of Amazon, you know they want to do 4,000, 5,000 satellites. They are trying to do another project called TerraWave. And there are probably now 15 odd constellation companies, you know, UTEL, SAT, OneWeb, SES, you know, host of others. And everyone, you know, constellation is not going to be one satellite in one inclination or, you know, one orbital plane. You'll probably put 25 in each of these orbital planes. So now if you need to have good coverage, especially in a communication or like an Internet um you know type of use case, you're going to put 2000, 3000 satellites. And that's the demand. And these are satellites today. The paradigm has shifted such that, ah you know, you're not building satellites to last you forever.
Speaker: 15 years. You're building satellites to last you 3-4 years, maybe 5 because it's become so much more cheaper to send a satellite. The cost of the ride used to be the most prohibitive cost, not the cost of the satellite itself.
Speaker: So you ended up having to build satellites that would last 15 years. And one of the big reasons would be because you had to carry power to last. you know You'd carry these big battery packs and then you'd put like a big solar panel to feed those battery packs. And then you'd put you had to do something called radiation hardening because these electronics were exposed directly to radiation because they are outside of the atmosphere and you know ah you know the atmosphere's protections to the electronic circuits and you know to us as human beings is not available to the electronics that are floating above. and they would get fried. So you had to do this process called radiation hardening, which was also expensive. So all of these things, the minute the cost came down, right, and SpaceX started, you know, like like i said, 170, 180 times last year, the cost of sending something up became so much cheaper that they're like, I don't care, you know, if the satellite conks off in four years, I'll just send 25 more.
Speaker: And that paradigm has changed how space is being thought of today. And hopefully we are, we're going to be, ah you know, one of those cogs in that journey where, you know, we're bringing down that cost to, you know, $500 per kg by a recovering, you know, stage two.
Speaker: um The other interesting use case, and I'm not going to name the company because something that we're trying to incubate, um you know, Elon Musk spoke a lot about ah data centers in space. right um And you know I had the privilege of meeting him in in November of last year and you know sitting down for an interesting chat.
Speaker: ended up spending, I think, two and a half hours with us ah at his facility in in Austin, at the Starlink facility. And you know very, very interesting, very open conversation ah and One of the interesting things that I like to say, so in that period, and i'm I'm going off on a slight tangent, but I will come back. In that period, when I was in the US, I met both. the I mean, when I say I met, the delegation met. It wasn't like a one-on-one, you know, ah Chai Pe Chatsya or anything, but just just me, you know, me and him shooting the breeze. No. ah So it was a delegation of about 15 people and the group met Elon, the group met Sam Altman at the same time. okay And for me, it was just so stark how differently you know i had percy i felt about both of those conversations. Elon was thinking just so big, such you know big use cases, such big problem statements, thinking about how to move humanity forward. is a guy who's not thinking about personal wealth and you know personal you know benefits and you know moving himself. He's put every dollar that he earned at risk. you know
Speaker: ah SpaceX, when they were launching, right the first 12 times it failed. right And the last time was literally going to be the last time. The time they succeeded, would if yeah if they had not succeeded, they would have probably gone bankrupt. And from there to a whatever it is, right? Trillion and a half journey is a journey that is to be marveled, right? And this guy is sitting and running seven of these, you know, huge problem statements, right? At the same time, we're all sitting and thinking, hey, you know, I'm not able to multiplex from, you know, problem A to problem B within my universe, right? And then he's like saying, okay, I'll think about SpaceX. And then, oh, by the way, I'll have time to tweet, you know, ah from the loo, you know, And then I'll think about Tesla and how to improve the supercharging ecosystem. And okay, by the way, let's also think about Neuralink.
Speaker: Let's change how you know we can communicate ah to the brain. And oh, by the way, you know let's think about transportation. let's you know We visited the boring company. Again, mind-blowing, right? ah He's just saying we can 100th or one fifty at the cost of you know how much it to...
Speaker: tunnel board today right just because people have not thought about fundamentally what are the problems in that ecosystem and hey you know how do I change that right everything he's ah addressing is very first principles ish and I think that's something that you know of all the things that I've taken away is like okay how do you you know filter everything down to first principles and you know it's It's slightly controversial, but you know I met Sam Altman nowhere way close to being as impressive, very you know very narrow in you know thought process, clearly not ah the the big thinker thinker that you know maybe some of these... we give them they give him a lot of credit or they you know because of ChatGPT or whatever it is. ah you know
Speaker: the payin know power previous backgrounds in startup or whatever, right? But somehow, to me, very clear, you know, who is who's the real builder builder in the use case, and who's probably not. And in fact, I think Elon had a part to play OpenAI as well, right? Significant part to play, you know, in a $38 million dollars check. And I think that's being arbitrated in court as we as we speak. Yeah.
Speaker: Yeah, so ah the the what I mean, yeah, there's a long way to come to orbital data centers, but sorry for that tangent, but you know, ah just an interesting tidbit there. So one of the biggest things see as as we look at data centers today, right? Though the biggest challenge for a data center is what? power, right? And, you know, cooling, right? And so cooling means one of the ways to cool is water, which is the easy access to it. And those are resources that you cannot scale as quickly as you can scale your, you know, chips and, you know, your your computing and everything else, right? Power cannot, you cannot build power like that, right? You have to plan for powers, right? Like yeah ah power projects are five years if, you know, on on like a good
Speaker: period of time, you know, everything clicks into place, land allocation and, you know, the netas and babus and everybody comes together. And I say this netas and babus, but it's the case globally, right? I mean, it's just you change what you call these people. But, you know, none of these things are happening.
Speaker: You know, maybe China is that exception there where, you know, as as the the way the country's political system is, they can power through a lot of these things. But pretty much everywhere else, you know, India, the broader West,
Speaker: This is not happening, you know, in a year or two, right? You're going to get stuck in, you know, in in acquiring land and, you know, building permits and whatever else not, right? And so how are you going to, you know, for all the the hype that we give on, you know, data centers, how are you going to scale the power that those data centers need, you know? And how are you going to scale for permissions and let's say political, you know, machinations that exist, right? Which is, hey, you know, oh data centers are bad, data centers are horrible, you know, let's go protest, right? We had that a fair bit in, you know, a lot of our states in West Bengal, in Tamil Nadu, you know, Sterlite and, you know, the Tata plant and whatnot, right? And there's political machinery that is, you know, sitting behind some of this. So his thesis was, you know, we should just build in space where I'm not going to face any of these regulations, I have access to unlimited power. right For him, the the thought process has always been, you know, solar is the is going to be the ah be all and end all of energy, which which also makes sense. Actually, the the method of solar, I think, you know will have to evolve a lot, lot more. But it is this
Speaker: infinite giant ball of fire that is you know throwing energy you know all across the world and why do you look for other ways of doing it? right And you know I don't know if you've heard of this thing called the Kardashev scale, ah which talks about humanity's progress as an ability to harness the energy available to it. right So you have level one, level two, level three, where you know you harness all the energy that lands on Earth. right And then you're able to harness all the energy of your star through a Dyson sphere. to Dyson spear through. So the the idea is, you know, that's how you move your move, the move humanity forward, right? If you're able to capture all the energy of your star, that gives you so much more capabilities to think about different problem statements. And so I think Elon is somebody who trying to address, you know, ah humanity on the Kardashev scale, right?
Speaker: And so the the big thing for him is, hey, you know, how about I just throw satellites up there, right? And, you know, they have access to solar power. we give them, you know, the the solar panels and, you know, the ability to harvest that. We'll put, you know, ah NVIDIA Jetson on that satellite and let's do some compute there, right? Rather than having to...
Speaker: ah you know, do it on Earth where I have issues on energy and I have issues on cooling as well. right And there you have the ability to in space, you have the ability to transfer heat out into the, you know, into space itself, into the void, that giant s sink that is at, you know, ah very low temperatures. So that is his broad thought process. And so somewhere along the way, we also think there use case for that. The biggest problem bottleneck today, Akshay, in going into space and it's going to become much more stark is the ability to downlink data from space to Earth. I spoke about this briefly, right? I said light, right?
Speaker: And people are saying, hey you know, ah radio waves are full, right? The radio spectrum is full. And then you have like, especially in countries like India, where The defense spectrum is like big chunks of spectrum that are blocked out for defense. right And all the other spectrum that you use, there's like some satellite or the other trying to downlink.
Speaker: To give you some perspective, right at low Earth orbit, let's say 500 odd kilometer orbit, um a satellite from horizon to horizon you will see for less than 7 minutes. It's traveling 25,000 odd kilometers per hour. So in that 7 minutes, today you're capturing so much of data. right You're talking about hyperspectral imagery like a company like Pixel is capturing imagery in so many different ah this things and it has to now downlink that data back to Earth. right And this is heavy data. right It's not, you know, your, ah you know, small like a selfie or something that you take. This like heavy, you know, high resolution data that has to be brought down and the ability to downlink that is limited to line of sight. So you need to be able to see the satellite to be able to downlink from a ground station. Right. And so the minute you it goes past that horizon and you don't have visibility, it now has to either transfer to another ground station.
Speaker: And that's going to be additional cost. Let's say, you know, if I don't have access to that, ah you know, i now need to pay to connect to that ground station to downlink from there and then transfer further and further on.
Speaker: Right. So slightly impractical, more and more going to be problematic. So there are two, three ways to address this. One of them is, you know, light based downlinking. Right. So instead of using radio, I use light.
Speaker: Now, light obviously has its own challenges. You need a you need line of sight and you need a clear day. ah So cloud is going to cause problems for you. And the the challenge which I think a lot of people are addressing is so you need to track the exact emitter so that that that piece that is emitting the light, you need to be able to track it at precision from the earth and I need to follow it all the along all all along the way. right And this is something that people are solving for. um i think they're doing it at you know lower distances, but you know they'll eventually get that type of thing. And the minute you can do that, it just opens up the ability to downlink huge amounts of data.
Speaker: The other approach is to say, hey, you know what am I downlinking the data for? right um Let's say I'm capturing an image of you know ah where I am today in Chennai, let's say around my office. And the question comes up, you know I want to track the green cover in this area, which is Noongambakam. I want to track how much green cover existed today versus maybe, let's say, last year or two years ago. right Now, the reason anyone, right, like a Maxar or any of these companies ah is capturing that image is because somebody asked that question.
Speaker: Right. right And as a response, these guys go and say, as a service provider, I will take that photo, send it to you, you do the calculation. I don't care what you end up doing with it. Now, there is this thought process that, hey, you know, rather than thinking about downlinking the data, can I answer that question, do that compute in space as well and give give you the answer 14%, right?
Speaker: Rather than complicating it by sending you the imagery, then you having to run that process locally, can I just do that in space? And this is an idea that, you know, we are trying to ah you know ah incubate slash work on.
Speaker: So there's a company called Planet Lab. ah Again, like Maxar, a company that's you know global, they have a satellite constellation which captures imagery and you know gives you imagery on a per image basis. You know you you can pay them $5,000, $7,000, they will capture you know imagery and send it down to you.
Speaker: This guy is ex-employee of Planet Labs and saying, hey you know, I think I can build something in this ecosystem. so interesting idea. We're trying to see how to, you know, support and help. Not yet ready for, you know, our like a full-fledged investment, but something that we find interesting as an idea. And, you know, we're trying to see how to ah work on that.
Speaker: ah Like I mentioned, the the other big use case, Akshay, And again, hard to name companies here, but see, though we have this, you know, GPS, right? yeah So GPS is, ah is a, is a, okay, so let let me see how to explain it. GPS is to GNSS, right?
Speaker: What Xerox is to photocopy. that make sense? okay right GPS is the American version of GNSS. okay um It's just so colloquial that we use GPS everywhere, but that's not actually the, that's just, we are using American satellites, which are GNSS American satellites, which is called GPS. Now, and there is a Russian version of it called GLONASS, which is Russia's GNSS satellites, a European version of it called Galileo,
Speaker: um There is a Chinese constellation called Baidu. Now India started off very excitingly said, hey, you know we will also put our own constellation of this. It is called IRNSS, Indian Remote Sensing and Navigation Satellite Systems, I think. Later, we christened NAVIC.
Speaker: Sadly though, program has been a bit of a disaster. ah There is something called an atomic clock that these satellites need to have. The atomic clocks have failed. Some of these satellites have just, you know, run out of, you know, just become too old to be operational. know, I think the last are two launches, one of the launches, two launches failed, right? Our ah last two launches failed.
Speaker: And I think in that process took down one of the satellites that was supposed to go and replace So our constellation ecosystem today does not exist. right And I'll tell you why it's important.
Speaker: ah Positioning is important. I mean, for you and I today, indispensable. right If you need to book an Uber, you know you need to know you know what restaurants are around you, whatever. right We use it for such mundane use cases. yeah But the important use cases are, let's say, military and defense and whatnot.
Speaker: in ah When Pokhran happened, right ah when we did our first a nuclear test, the europe in retir not euness a US in retaliation turned off access to GPS for India. right So over over that entire plane, they just shut off GPS. right And so what ended up happening was all our so all our ah missile systems,
Speaker: were connected to GPS based calculating and you know had to understand their own positioning using GPS. And so suddenly we were you know left without a system. And that's when the government said, you know hey, you know we need to build ah you know our own capabilities here.
Speaker: And unfortunately, it has not happened. right I mean, it or it has not happened well. ah we We've made an attempt at it, but it's not done well. so Normally to do this positioning system, you need to be in a very, in an orbit called geostationary. Generally, the orbits that are like 30,000, 35,000 kilometers up.
Speaker: And these are hard orbits to get to harder. I wouldn't say very hard. we we We have the ability to get there and you need to put satellites in in that ecosystem. And today there are startups saying, hey, you know, I can do this in LEO.
Speaker: Right. If I do it in Leo is lower, geosage butt lower earth orbit. So which is, you know, anything from like 300, 400, 500, you know, to, you know, a few thousand kilometers also is broadly Leo, lower earth orbit.
Speaker: um But geostationary, for example, 30,000, 36,000 kilometers. Right. And there's startups saying, you know, how would I build out a constellation that does positioning, but in Leo?
Speaker: right But if I put more of them, I have the ability to control for that. And so there are startups saying, hey, you know how about I solve for ah this you know positioning you know GNSS problem statement, but I solve it in low Earth orbit and not in geostationary. And there are ah two, three very interesting startups that are building in this space.
Speaker: ah Guys who are like X is row, you know decent space pedigree who has stepped out and starting to build in this space. So that's another very interesting use case that you know I think has good commercial ah you know ah visibility over the next you know decade or so. So These are some of the space ideas. Maybe moving on to defense.
Speaker: Defense is again a space that there's a lot of activity happening. I mentioned that the flight controllers are like the basic bottleneck. Historically, you know we've been sourcing flight controllers from outside. Today, there is this ecosystem that's saying, hey, let's build these flight controllers in India.
Speaker: That's a very decent use case to do. and The broader drone ecosystem as well has largely been dependent on you know sourcing from outside of India, things as simple as you know propellers, right? These quadcopter propellers are basically consumables, right?
Speaker: Every time you fly for X hours, you have to change the propellers, chuck them out because the the blades get blunted, right? It's literally hitting the air, right? It's cutting through the air and the edge of the blade will get blunt and it will start becoming inefficient.
Speaker: So you have to replace. So even things like that are imported, right? the the The propellers will come from China, will come from Czech Republic, not built in India. Today, there are startups that are building that.
Speaker: I'll talk about our own company. We are investors in a company called Zebu, based out of Hyderabad, and they are building counter drone defense systems. Now,
Speaker: So, every year, there are probably, you know, a few thousand ah drones that are sighted crossing the India-Pakistan border. ah You know, small quadcopters that are used for...
Speaker: um let's say, like military missions. So they will come surveil, you know, something in India, they will come drop off arms and, you know, drugs and, you know, money to people ah hiding out in the hills to, you know, terrorist camps hiding out in the hills or sleeper cells in the hills.
Speaker: And usually these drones are shot down the minute we see them on radar or just on vision, we shoot these drones down. Now the problem is that that is not a sustainable use case. I can't keep shooting them out because these drones are so cheap that you know you just keep sending more and more across the border. right And so at some point the army said, hey, you know, I want the ability to go get that specific drone out of the sky such that I'm able to do forensics on this drone.
Speaker: So that is the first problem statement that our company Zebu built for is, you know, you see ah a rogue drone, we launch our own counter drone, which goes and locks itself on top of, you know, this rogue drone.
Speaker: at a 10 meter altitude using computer vision and radar and then it will capture this drone by launching a pyrotechnic explosive net and this net will get stuck in the propeller and then you know the drone can be captured.
Speaker: And so this then gives the army the ability to do forensics, right? Understand who is supplying the flight controller, what payload was being carried, what photography, you know, was probably probably captured, you know, what band was the communication happening to the controller, um you know, ah who who supplied some of the other parts of the drone. So,
Speaker: essentially forensics such that you know they can now understand better how to control for you know some of these problem statements rather than just shooting it down every single time. Very cool. Then they built a whole system around the autonomy use case in this defense piece. where they now have a host of other products, including a Coast Guard drone, which is a drone that can, you know, if it falls into the water, it can stay afloat and it can, you know, it's essentially be working. Only the motors need to be replaced out. ah So each of our Coast Guard ships in India, for example, are equipped with a drone and a drone operator.
Speaker: Now, the challenge is that these are manually operated. So the ship is stationary, let's say, and then the ship, the the the captain decides, that you know, there is some activity. Let's go explore. So the pilot will fly his drone out, right? Let's say that's seven kilometers, six kilometers out. And then it finds a piracy act activity happening.
Speaker: Today, the protocol is that drone has to be flown back to ship, landed safely, and then only the ship will take off in pursuit. OK, now that is protocol because they don't want to land that drone when it's moving, lest it hit the radar or guns or, you know, what is very expensive equipment.
Speaker: So but what ends up happening is you lose, you know, 10, 15 minutes in the process to wait for this drone to come back. Today, the system that the Coast Guard has asked for is the minute that reconnaissance thing is done, the guy just hits a return to home button and then the ship takes off.
Speaker: The drone now has to figure out how to land safely. um Yeah, it needs to understand the speed of the ship, the direction, you know the pitch and roll of the waves, and then land on a 3x3 mat that's designated to it. And that's completely auto autonomous. So they've built a host of such solutions. Very very interesting company. Interestingly, we invested in them pre-operation Sindur.
Speaker: And so there's been a lot of interest in them post that. Very good. So these guys have done done well. They have a tethered drone system which can so you know stay stay in the air for like eight hours at a time. They have sold it to you know the Patankot Air Force Base and Ambala Air Force Base. And these are drones that take the power lines up you know into...
Speaker: ah into, ah you know, flight and therefore they're not dependent on the battery pack, you know, to power the system. And eventually it is the fatigue of the plastics and the propellers that, you know you need to come back down and replace with a new drone. It gives you that ability to keep continuous surveillance around around you for...
Speaker: ah There are very interesting companies in this you know autonomy, defense autonomy, defense intelligence space, ah electro-optical systems, right? You know, the lenses, ah the data capture systems, all of that is again something that quite a few companies are indigenizing on today.
Speaker: um people building ah you know the the camera systems ah you know and they're able to capture some of this data, not just the actual lenses, but you know the entire system that operates behind it, ah you know ability to capture in ah multiple parts of the spectrum.
Speaker: and overlay some of these imageries such that you have ah you know data from across a spectrum and you can see through certain obstacles and see through the night for example and you know overlay information. A lot of these things are being built today in-house. Is anyone trying to build the Palantir of India?
Speaker: There are quite a few that are eventually trying to be that. I think we're very early in that use case, Akshay. I think that's that's the right question that all of these guys should be asking is, you know, how do you become that? How do you become that layer of intelligence for all these, you know, defense problem statements? I don't know if Palantir themselves are the Palantir of US, you know, i think ah Claude suddenly woke up and yeah there was this big...
Speaker: Hala about you know ah the the US Defense ah Administration using lot to take decisions and all that. right So somewhere along the way, Palantir's role, I mean, obviously they're so well integrated, but suddenly I was surprised when I saw you know there is that integration also happening.
Speaker: So there is, i think there are multiple layers to that. And I think there are people trying to build in each one of these layers. that auto autonomy use case. But I think eventually we will need to own that entire value chain. This is something that is so sensitive that you cannot afford to have any lapses in security, in backdoors, in firmware, software, everything has to be you know owned end to end in house. And it looks like our defense ecosystem understands that fairly clearly. There is now people who are heading AI for the army, for example, right? There is a general who's now heading that. And so there is now some you know visibility on somebody thinking, hey, you know across all our procurement, how do we integrate some of this? And some interesting thoughts around that.
Speaker: So yeah, I think lot of interesting ideas. There are now, again, 300 plus startups building in defense today, Akshay. And these guys are competing. I don't know if you've heard of this program called IDEX,
Speaker: And ah that's that's a program that's again ah brought brought brought by the ah Ministry of Defense ah that now aggregates all the problem statements. So it brings all the defense participants, right the big guys, the Air Force, Navy, ah you know Army, Coast Guard. And then all the other second tier, third tier players, the shipbuilding, the ports, Musgaon Docks, Vishaka Pattam, Armored Vehicles, Nigam, Midani, BHEL, right? Anyone who supports the defense ecosystem, they're saying, hey, okay, if you want a problem statement to be put out, right, you have some but particular problem statement that you see important and you want people to build it,
Speaker: or build for it, you can come declare this problem on the IDEX portal. um And then people can choose you know A startup can say, Zabu can say, hey, I think I have a solution for that.
Speaker: Or a Sagar defense can say, I have a solution for that. So, kind of levels the playing field, ah you know takes away the need for like you know these huge EMDs and performance guarantees and you know track record. right If somebody were to give a company a 100 crore order, they'd be like, hey, show me two orders you did for 50 crores at least right before I give you a 100 crore order. And in the startup use case, you're never going to have that.
Speaker: And that's something that this IDEX program is helping to take away. They're giving you grants starting from, I think, about a crore and a half to as much as 25 crores, actually. right And saying, you know, it'll be a matching grant. You guys go and raise an equivalent amount of money. We will give you 25 crores. So you can then start building for particular problem statements. I think this again goes back to our you know previous parts of the conversation, the support that is coming into building sovereign tech.
Speaker: And I think that's going to be the catalyzing factor that, you know, moves the ecosystem forward. Amazing. Thank you so much for your time, Manu. This was a super intense conversation. we covered so much area. This is truly a resource, like a resource with super long shelf life. Thank you so much for your time.
Speaker: Well, this was fantastic, Akshay. Thank you for having me. a Lovely conversation. And, you know, glad to have, you know, had this chat with you.
