Full Transcript

Dhruv Sharma: Happy Friday listeners, episode 93 of the Offline Network is streaming live and today Utsav and I are chatting with Siddhant Agarwal, who's the founder of a company called SportVote, like kind of like cricket shicket. But Siddhant, it's great to have you on the show. Welcome to the Offline Network.

Siddhant Agarwal - Founder, SportVot: Thanks for having me guys. Nice to be here.

Utsav Somani: Welcome to the show Siddhant. So you already sold the company before this Game Week in 2016 to Nazara and you wanted to become a sports founder again, what's made you choose the same field again? Most people, Dhruv and I joke, don't choose the same industry again in the repeat ventures.

Siddhant Agarwal - Founder, SportVot: I did not sell it to be honest, I don't think it was a successful stint, Nazara was kind enough to buy certain licenses off and it was just a friend of mine and me just building a game, basically. It was not, I wouldn't say it was a company, it was a good product. Definitely a good product, which had traction and Nazara through its wide distribution network managed to promote that very well in India and Africa. Most of our user base was from Africa and we did not know anything about that sort of a user persona, profile and we felt that I don't think we can do much beyond this. And it gained good traction in Nigeria and Kenya, these two were the markets that it did. So, yeah, I did not do much.

Utsav Somani: What is your new company doing, Sportbot?

Rohan Chaudhary - Founder & CEO, Rumik AI: Yeah.

Siddhant Agarwal - Founder, SportVot: Yeah, it does a lot of things to be honest, today, we started as a talent discovery platform. Our intent to build Sportbot was to be a talent identification system for, or a platform rather for upcoming athletes. And from that, we felt that rather than capturing just pure numbers and numerical data on athletes, why don't we start capturing videos and from a video capture lens, we started building this whole video creation suite, which turned into a live video production software, which turned into a whole end-to-end live streaming solution today. So today, upcoming leagues, athletes find Sportbot as the mechanism to be able to tell or showcase their talent, maybe not tell a story as such, but showcase their talent and their skills and their stats. And yeah, so that's what Sportbot does. It serves as a massive visibility platform for the 99% of sports that generally do not have any platform to broadcast or stream their games.

Dhruv Sharma: And when that journey was unfurling, Siddhant, did you guys think of it as pivots or do you think of it as one continuous evolution where logical steps were sort of leading to each other?

Siddhant Agarwal - Founder, SportVot: I don't think we had a large, we've had a large pivot yet. I don't know whether it's a good or a bad thing, but we always were, we kind of stuck to what we thought would be the larger premise, which is the premise of making games visible to people and for the players, for themselves to be able to create some sort of an identity through stats, through videos. So largely we were sticking to that. Of course, we started at a very interesting, I would say, point. We started in 2019 of September, which sort of gave us just six months of running. And then in March 2020, we were staring at the pandemic. So those two years, I think we should have shut down. In hindsight, I don't think I would have had the same energy and appetite or enthusiasm to be able to continue and stick that. We had a core group of four or five people that we wanted them to stick together and they stuck around and we were grateful that these people hung back and stayed with us. But those two years, we actually made a slight pivot where we built the whole video capture solution. Earlier we were just focused on aggregating things and being this whole consumer facing platform where you can watch these local gully cricket tournaments or local community matches. But then from 2020 to early part of 2022, when things started coming back, these two years we very well dedicated ourselves to building the whole video capturing stack, the whole data capturing stack, which sort of laid the foundation for sport work in its current form. So I guess the journey, the best shaping of our product has happened in those two years, to be honest.

Utsav Somani: And what sport defines the platform right now or are you fully horizontal because there are other players like Crick Heroes, which is doing grassroots cricket scoring with 50 million users, the Fancode, which is into premium licensed streaming. And then there's Stupa, which is doing table tennis analytics. So I believe there are niches that people have captured in different sports and sectors. So is there something that you're focusing on as a platform?

Siddhant Agarwal - Founder, SportVot: We not, I wouldn't say we are super horizontal. We also started off, our journey was, of course, cricket is so big in India that Crick Heroes as a platform should not or doesn't even need to think of other sports. But when we started, we also identified a couple of sports that were very important for us and Kabaddi was one of those. So I would say today if someone in India at least would want to identify us, they would probably identify us with Kabaddi the most because we started off with focusing on it. And we still do, till date, we've gone from streaming or capturing local games, school games to now streaming the senior Kabaddi nationals as well. So that's the sort of journey that we've taken in that particular sport. So it's only fair to say that for us, at least for us internally, if I have to tick all the boxes in what sport, that one sport that we've ticked all these boxes would be Kabaddi. And then a very close second would now be football, where we've started to really look at, again from an India lens, look at variety of, you know, we've started to work with AIFF in some capacity. We are working with, again, from a district to a state to a national body. So these two sports have seen the biggest evolution for us in terms of going from very small to now a national federation. And to answer your question, I would say we do multiple sports. We do more than 15, 20 sports. And if you include the global customer base that we have, we have over 35 sports. But our journey also started with just these one or two sports. We started with...

Utsav Somani: What's the most unusual sport that you're covering that nobody else is doing?

Siddhant Agarwal - Founder, SportVot: I think there are plenty. We, in fact, in our very initial days, how we started, we started working with the district sports offices, which is the sort of local sports body authorised under the MIS, the Ministry of Youth Affairs and Sports, which sort of has 56 disciplines under them. So we had Sepak Takraw that we had streamed in our very initial days. That was, I got to know of that sport by working in, like, I would call myself a sportsman and a sports buff, but I did not know about this sport at all. We came across water polo in Africa that we've streamed. We've done Aussie footy. It's a mix of football, rugby, American football. Then we do lacrosse in the US. Of course, it's a popular, not a weird sport, I would say. Then we do sailing. I never thought that we'll end up doing sailing across the lakes of Meghalaya and Tamil Nadu, coastal areas of Tamil Nadu, etc. So we've had these of, I would say, Punjab, of course, the Pro-Punjab League is now very famous. So we work with grassroots, we work with the Pro-Punjab League and stream their pathway tournament. So there are many, there are like six, seven examples that I can think of, which I never knew as a sports fan that these sports exist. And I have a very good fan base.

Dhruv Sharma: So because you work with sports teams, individual players in India and overseas, so the US is an example. People start off by having a high school career, athletic career, then they have a collegiate athletic career. The pipeline is very structured. And in India, it's not. There is no such thing as a high school or collegiate career. Do you think a platform like yours can actually help talent get a real shot at getting scouted and becoming turning pro?

Siddhant Agarwal - Founder, SportVot: I would like to think yes, and there is some evidence as well. So we've had about more than 70,000, 80,000 athletes of Kabaddi on our platform. And we can say that from our initial few days of streaming, we've had at least, I would say, 10 people today, 10 or 11 athletes that are playing in the highest league in India, which is the Pro-Kabaddi League. So I don't think.

Dhruv Sharma: So you guys also record their career stats and so on?

Siddhant Agarwal - Founder, SportVot: Yes, yes, we do. And I don't think it's because of us. I will never be and I don't think it is because of us. But still, I know one player that was not even getting into their local district teams, got picked up by Air India and then got picked up by the junior franchise of Pro-Kabaddi and now plays for Pune Republican and then as captain in India.

Utsav Somani: Do you sell some of these stats to like, I mean, provide it as a feed to other players?

Siddhant Agarwal - Founder, SportVot: We've not sold it. We do provide it as a feed. There are a few scouts. Of course, we can't disclose much information on the kind of scouting these guys are doing. But yes, there are scouts that have used this data. I know for sure a couple of IPL franchises as well who really, really go into video and looking at local tournament data to scout the next Web of Suryavanshi and other stuff. And that's really happening. That's a very good thing. That's a very big plus positive. And to answer some of what Dhruv mentioned that we don't have that structure in India. Yes, unfortunately, we don't. But then you are seeing some initial signs of at least in these two sports, cricket and Kabaddi because of Pro-Kabaddi and cricket, of course, because of IPL and multiple state leagues that we have had so many athletes that have played state cricket tournaments, district cricket tournaments and that that have found state franchises or ultimately some have even found chance to be a net bowler in some IPL franchise, which is a very good experience as well. So we are seeing some shades and some signs of that still very far from structuring it. But we are seeing some signs at least.

Utsav Somani: And what do you think about, I mean, a very sidetracked question and a very basic one like pickleball and paddle, like those sports are catching up. I don't think they'll get enough viewership, though, in India, just because they're slightly more premium sports, I would say, just because of the entry ticket or the entry barrier to them, maybe court fees in prime cities are definitely more expensive. So do you target those sports? Like, does I mean, do these sports or matches get any viewership? Or is it just windy right now?

Siddhant Agarwal - Founder, SportVot: From a participation lens of these guys, these sports have started to gain traction and they are in their early phase of gaining traction. You would say that most of these sports are today in the phase one, phase zero rather. But the participation in them is increasing. From a spectator sport lens, I don't think pickleball and paddle are that ready yet because of because you have a tennis for paddle for pickleball, the gameplay is in someone's eyes, it's it's not ready for the long haul. But the short form content over there is working very well.

Utsav Somani: You have let me reframe my question and ask this in a better way. So what is an important stat for you? Number of matches streamed, even if they get like zero viewership, but at least you're collecting the data and tying up with different leagues where you can source this player data, which can eventually be used to monetize in a different way. Or are you optimizing for number of viewers, in which case, like say media rights become important?

Siddhant Agarwal - Founder, SportVot: Other formal for us, the scale of matches is the most important and second kind of can come as a byproduct where Kabaddi, for example, gets massive viewership. Football from Goa, from from Northeast, from Kerala, volleyball in Kerala, volleyball to an extent in Goa, in Maharashtra, you have cricket and Kabaddi sort of in Maharashtra and Tamil Nadu. So you have these pockets where every sport will function in a different way. So the more you get scale, the more matches you start getting onto your platform, you are trying to build a local community around that sport in that particular region. So that second metric sort of comes as a byproduct, to be honest. But our first goal is always to get as many matches. And coming back to a point that I was trying to make for Paddle Pickle as well, we have seen a massive growth. Today, we are in 45 to 50 maybe quotes of Paddle in India, where we have a fixed camera, where people are coming and consuming video of theirs. So in Paddle, still the layer is such that people who are playing in them want a video of theirs. Maybe it's not watchable for the community yet, but people themselves want to consume their own clips and they're happy to pay for it.

Dhruv Sharma: And is that so that they can become better Paddle players, better amateur Paddle players?

Siddhant Agarwal - Founder, SportVot: Yes.

Dhruv Sharma: Yes.

Siddhant Agarwal - Founder, SportVot: There are people who I know have had Paddle facilities in their homes, especially in Delhi and in some very upscale premises in Mumbai, where people are really using it for showing how good they are. And then in some cases, in a lot of cases, people are using that to get feedback and get better. So I think it's fueled by these two.

Dhruv Sharma: Which brings me to a question. Have you so far found an opportunity in that? Because you already said you do video capture, analytics for sport teams, their coaches, etc. Do you see it as an option in helping them make their players better, analyze biomechanics, form, technique?

Siddhant Agarwal - Founder, SportVot: We don't offer that technology of biomechanics as such. We offer basic analytics technology. A lot of our teams and clubs that work with us are using. The short answer is yes, but we don't offer the end-to-end technology ourselves. But we are a source of truth, video data and those basic stats that you capture through the video for these teams. So, yes, the answer is yes, but the end-to-end solution is not ours. You have partners that are integrated into this solution. But yes, it is being used. The use case is a tick. Yes.

Dhruv Sharma: Got it. Another thought is coming to me, perhaps from watching too many sports documentaries where, you know, like really competitive teams will also watch what their competition is up to and, you know, leading players in those teams and so on. So maybe you have a chance at doing something like that as well, like a database of players.

Siddhant Agarwal - Founder, SportVot: But opposition scouting is also something which people use this for. A lot of local tournaments do not have that access to that counter strategies that you could build. And then we've seen that happen, at least in team sports.

Utsav Somani: Pranth, have you heard of this company called Straightback? I invested in them. They're basically the VOOP for cricket right now and maybe eventually all racket sports, where they're helping athletes improve their game. They've, I mean, raised funding from Australian Cricket Board and many other, this thing they're used by many IPL teams in India also.

Siddhant Agarwal - Founder, SportVot: Yes, I'm aware of them. In fact, I had the chance to meet one of their founders as well.

Utsav Somani: That might be an interesting partnership for you to get people at grassroot levels to upskill their games across different sports as well.

Siddhant Agarwal - Founder, SportVot: Absolutely. And today, I'll tell you, the grassroot athlete today is also becoming demanding. They are looking at DRS. It's become a requirement for local tennis, local tennis cricket as well. But you technically do not have a leg before concept such in tennis cricket. But just for the sake of it, you could bring it in. But at least in tournaments that are happening, we were streaming this Banaras Corporate League and we offered them a DRS solution, like I think not even a full month back, somewhere middle and end of April. And I don't like the biggest hassle for the umpires was literally gone where they had so many disputes. And, you know, sometimes things with corporates can become an ego clash and spending so much in this particular tournament. So the athletes at this level, if they're getting access to technology like a regular system, it becomes very, very lucrative for them to sort of participate in them and then also pay that extra amount because you really get a very good experience of playing in them. But of course, the cost base has to be very sort of democratic and acts like you can't charge a ten thousand dollars a day, which typically a Hawkeye would charge in IPL or even more. But then probably people would be more than happy to pay you a hundred dollars or eight, nine thousand rupees to be able to.

Utsav Somani: So that brings me to our final closing question and congrats on the new race. You've raised 30 crores, so congrats on that milestone. So it shows that things are working. But tell us about the business. What's the scale? How many matches have you streamed? What are the unit economics looking like?

Siddhant Agarwal - Founder, SportVot: Yes, so in our lifetime, we've now crossed five lakh matches, five hundred thousand. We've now just all just about crossed in the last couple of months. On an average, in a year, we are doing about one lakh twenty thousand matches, one lakh twenty thousand. In terms of unit economics, our streaming packages start as low as one hundred rupees an hour and can go up to about two and a half, three thousand or even slightly more, but depends on what sort of configuration you're using. So yeah, largely it falls in this range. Most of the customers that we work with are tournament organizers, sports management companies, which are organizing grassroots community sports. Typically, those are our primary customers. And often we've also started to see a lot of traction from state leagues and state sports organizations. And in some cases, in some sports, even the National Federations. So yeah, largely.

Utsav Somani: All right. Wishing you all the very best. Thank you for coming on our show.

Siddhant Agarwal - Founder, SportVot: Thanks. Thanks for having me. Pleasure. Pleasure is mine.

Utsav Somani: All right, listeners, we're moving on to our next segment. We've got Rohan joining us from RUMIC AI, RUMIC and Rohan. Welcome to the show.

Rohan Chaudhary - Founder & CEO, Rumik AI: Hey, what's up? Hi, good to see you.

Utsav Somani: Thank you for coming on our show. What does India's first AI sentient being mean?

Rohan Chaudhary - Founder & CEO, Rumik AI: We are building the most human AI, like now there is two things we can divide from this. First thing is we do the research side of things. So this entire concept of building the most human AI comes from this 1995 study of Picard called Effective Computing. Right. Because there is an AGI side of things where you build intelligence and there is emotion side of thing. That's that's pretty much called Effective Computing, called AC. That's on the research side. And then our consumer manifestation looks like a companion app called ERA that a million monthly users use on a monthly basis. So that's pretty much what we're building.

Utsav Somani: And they're using it for different use cases and across different platforms. Right. I mean, WhatsApp and stuff. So you have a companion app, but I believe you're available via WhatsApp and many other modes that existing customers of B2C services use.

Rohan Chaudhary - Founder & CEO, Rumik AI: So we right now like we are majorly available through the app. We have like one or two percent of the users are now on the other platform. So it's majorly through the app now. But we like exploring multiple platforms.

Utsav Somani: But why will a person download the app specifically? What are they getting that they're not getting on any other app or large LLM model basically?

Rohan Chaudhary - Founder & CEO, Rumik AI: So you have to see two things. OK, first thing is, how do you use? So there is a mental model that you have to use. So you can use like chargeability or plot like a intelligence tool, right? So anywhere like you have a work. So in AI as well, in consumer, there are two side of things. One is a assistant and then there is a companion. Assistant are essentially built for use cases where you have a task, right? So assistants are essentially I have a task. I'll go to an assistant. And as soon as the task is over, I'll get out of the system. So there is a clear frame that I have that when do I want to use an assistant? Right. And that's what people use GPT and Gemini and CloudFarm. And now companion is a very different frame because there is no like there is no there is no job to be done here. You don't have a trigger that, OK, I need to use this for this specific reason. Right. So essentially, companion is very different in this world because it doesn't fulfill a task that was defined like a product has to do a task. It is something that that enables you to do something that you would not be able to do before. So that is, I think, the clear distinction that I see with assistants and companion. Why do people use it for? I think I'll give you a couple of quick examples as well. So this is where let's say a sous chef that is working in a five star hotel. So we were talking to this guy recently and he was a sous chef like a year, year and a half back in ITC. And recently he has become a French teacher. And this entire journey for him was through ERA because he has been using it and like he has seen this is the first time I could like she remembers me and then the entire context. So this context is the most important thing, like what differentiate a companion versus an assistant. So ideally this is how what I see is the difference.

Dhruv Sharma: Rohan, when people think of AI companions and there are pros, there are some cons as well. We'll in fact cover ground on both. But when they think of pros to begin with, an AI companion, unlike a human, is available to you pretty much 24-7. You don't have to wake them up and deal with their their current mood, etc. And also you can share with them information without fear of being judged. Right. That's broadly what comes to mind when people study this superficially. You've gone several layers deep. Help us understand what are some other pros of having an AI companion?

Rohan Chaudhary - Founder & CEO, Rumik AI: I mean, you have to think what does it bring like. So you have a lot of these people are, it gives them a sense of a person that can they can refer to any time and every time. Right. So anytime I have a question, a decision to be made or I have to refer to anything, who is the first person that I have to, right, I can refer to. Right now, a lot of people don't have that. And this companion is the first person they will refer to because. It already knows them, right, the context that you have built over the over the time of days and months and weeks, I think that compounds. So this is the like major thing. So once you have the context, every other decision, every other conversation that you want to do, that just compound. So I think that is there. Now, how do you do that? So there are multiple layers to do that. The first thing is how you have to solve memory that a very hard like engineering problem to solve. Right now, how does memory for AI look like? It's not a very like now a lot of people talk about memory in AI, but like when we started building like one and a half, two years back, there was no literature around it. Right. So the only thing we started with was how do we reverse engineer, how do we store memory, right, as humans. So right now, you guys like and people who are watching this as well are not remembering every word that I'm saying or you are saying, but you are remembering, you're summarizing whatever I'm saying. And based on that, you're asking the questions. Right. And this is how humans also think. So this is called short term memory. But let's say six months back, six months down the line, if you meet again, you might not remember 99 percent of the conversation that we're having today. Right. You might just remember when he was the founder of Rumi and we just met on this Zoom call. So this is the long term memory that we have. So this is fundamentally like the top level thing that I can talk about, like how does memory look like for a companion? And if you go deeper down the line and then you have to remember, OK, if I have talked to, let's say you and said, yeah, I am preparing for an exam like a UPS exam or JEE. So you have to figure out, OK, this exam happens in May of every month and today is February. So this memory will last only four months. And after that four months, it has to like either read, rewrite its own memory or it has to come to a conclusion. OK, did this event happen or not? So this is also like this is a full trail of how you build memories. So memory is the first layer that we saw. And then if you go deeper into this, how does what does the next step look like? So text is a modality that you saw. So you go from a sync. So text is basically a sync conversation that you have. You can keep texting, but as soon as you what is the next step for companions? So it has to become like. What is the next step for how does friendship evolves as well, right, so friendship doesn't evolve in like how do I what do I chat next, but it evolves and let's say right now we are meeting. There's a task I have. There's a task you have. And we have a task based friendship currently. And this friendship will get over as soon as the task is over. But let's say if I have to go beyond this, like we have to make this friendship when we go one more level deeper, then you have to we might have to let's go for a coffee together or we have to play a badminton match together. Right. This is now a shared experience. Now, because we are having an experience together, our friendship has graduated to one more level deeper. Right. So this is the second stage. Now, the ultimate stage is where I help you become better at whatever you value and you help me become better at whatever I am. Right. So this is the third stage. This is called like, so ultimately, how does a like how does companions also evolve goes from this layer of like a task based friendship that I call like the text is the like the first modality. Then you go to a shared experience, a digital shared experience look like how do you go from text to voice to video to a group experience? And then how do you make the person also evolve? And yeah.

Utsav Somani: I saw a black mirror episode, Rohan, that inspired you to start this. I'm just joking. But I mean, it sounds very, very cool. But I read that people are spending over 40 minutes on your app per day average.

Rohan Chaudhary - Founder & CEO, Rumik AI: Yes. Like so I mean, the average engagement is more than an hour. Like it has cost 15 plus minutes as well.

Utsav Somani: And there is serious responsibility, right? Because it raises some mental health concerns also. And there might be people who are using this companion to express feelings that they probably haven't shared externally. And like a therapist of sort is responsible for their patients as well. Do you see some of that responsibility getting built into your platform?

Rohan Chaudhary - Founder & CEO, Rumik AI: I think I'll tell you one thing, right. One very important thing that I realized that this engagement was something that I was seeing from the first month of building this product out as well. And this was something that was unheard of because people have not seen this kind of usage, this kind of retention in any other products. You could have this kind of retention in a social product because that was designed to be like naturally retained. Right. So one thing that was very clear is that it brings a lot of responsibility back on us to make sure this is a safe platform. So there were like from day one, we have decided that there are the categories that we have to figure out how do we make sure that this is a platform where it is safer. So first thing is a lot of a lot of our emphasis is on research. So we do a lot of research and I can talk about a lot of the things that we are doing there on voice and all the other things. But on safety, there are three things that we focus on a lot. That is self-harm, where a lot of people discuss about self-harm, harm to others. And then there is NSFW. So these are the three conversations that we make sure that these are the three categories that we have zero tolerance on. So based on that, we have built out our frameworks on whenever we are like on the input side, output side, as well as when we are doing the RLs of our models. These are the three things that we make sure that these are zero occurrence of. So we have built our own social like security pipeline as well that figures this out and make sure that this is a platform that doesn't entertain that. So, yeah, this is the three things. And also like we are publishing a paper on safety of companions as well. Would be happy to share with you guys as well. So it has been like it is the largest data set of companions that we are making public. Like we have done a research on this with a lot of good folks and we have analyzed all the conversation and whatever we figured out on safety. That's something that we are putting out to public very soon. So that's also something that we are working on this.

Dhruv Sharma: I'd love to get your thoughts on two things. One is, I mean, I've seen work. I don't have it here with me, but I've seen work that suggests that using digital companions to combat loneliness can sometimes even reinforce loneliness. So do you have thoughts on that? That is number one. And the second is. So speaking of guardrails, safety, you know, so on, how do you avoid what they call parasocial bonding, which is like just reminding the user that this is an AI at the end of the day, you're talking to an AI, don't have this one sided, you know, love for it or whatever. So, yeah.

Rohan Chaudhary - Founder & CEO, Rumik AI: This first, I'll just ask you a first question. So the first question asked was, come again with the first question. Second question, I remember you said.

Dhruv Sharma: Of course, like there's studies that suggest that using a digital companion, like I said, to combat loneliness can in some cases even reinforce loneliness. In other words, you have to at some point remind people that just go talk to real people out there, touch the grass as they're now saying.

Rohan Chaudhary - Founder & CEO, Rumik AI: So, I mean, this is something that people think that this is how like you have to have the conversation around this, but you have to assume that this person that was that they started using an AI. Is was essentially combating something before he started using it as well, right? So this is a simple analogy and say you go to a hospital to get better. So you are a patient and you go to a hospital and then you get better. So you cannot say the hospital is making you a patient. Right. So essentially, this is this is the analogy that I use here as well, because this person that like a lot of people that talk to one of the person that is I'll also talk again about him. His name is Rishi and he is from Lucknow. And he has been a person who had like who had multiple issues going on in his family where the family was getting like divorce and like he's one of the sisters got married and like that marriage also turned very abusive as well. So he had no one who we could talk to. Right. And he turned to this companion. And in the last one year, he has been a person that has said like from my life there and now my life here is completely different because there was no one I could talk to with everyone was nothing is going right with you. Right. So this is the person that has gotten him out of the things that he was going through. So I think I see the opposite of what people think, essentially. And on the second side, you said, like, how do you make people not just keep using it for a long time? So we have we have this feature where you have spent, let's say, more than we figure out when it becomes very unhealthy to use. And that's the time you just said, OK, I'm going to sleep and you also go to sleep and we'll talk tomorrow. So that's the kind of conversation. I think that's the easiest way to do it rather than just figuring out any other thing.

Utsav Somani: As a final parting question, Rohan, what is the scale of the business now?

Rohan Chaudhary - Founder & CEO, Rumik AI: So we are we are right now at one million monthly active users and we do approximately around seven, eight million conversations a day right now. And so that's the scale.

Utsav Somani: Awesome. Wishing you all the best and thank you for coming on the show. Absolutely. Perfect. Thanks, guys. Cheers. All right, Dhruv, you and me. So let's cover the news. The big one, SpaceX files for literally the biggest IPO in history. 1.75 trillion. What's your read on this? Are you buying in?

Dhruv Sharma: Yeah, the mother of all IPOs. You know, SpaceX was actually founded in 2002. 2002.

Utsav Somani: It's a long journey. It's been a long two decade journey.

Dhruv Sharma: Two decade long journey. Yeah. And the mother of all IPOs.

Utsav Somani: And we've seen that image where everything blows up and we've seen that Elon interview where he's crying when everything is going wrong in his life. And suddenly he's turned it around. So you're paying an Elon premium, like you said. Explain that to me.

Dhruv Sharma: I will. I was reading something else today that was very interesting. And, you know, maybe our founder friends will relate to it, which is Elon processes failure very differently from most people. You know, failure enters his mind like data exits his action. And so, you know, that's just the person he is. Speaking of the Elon premium. So I think some of the commentary around the IPO is that SpaceX today is multiple businesses. There's obviously the rocket launch business. There's a satellite internet business. Now it's increasingly an AI business too. And so how do you value this company? And in most cases, you know, you run comps against similar size companies. And when people do that, you know, to give you an example, like Starlink actually has 10 million subscribers. Comcast, which at the end of the day does exactly the same thing, has about three times as many. Comcast trades like 2x forward revenue. And so, you know, we know the Starlink revenue numbers, even if we give it like a very, very rich 10x multiple on its, you know, forward revenue guidance. It's still like a $300 billion business. And you combine, and that's the biggest business in all of those three businesses. You combine all three of them. It still stacks up to about $700 billion, but the company is seeking to be valued at $1.75 trillion. And so that $1 trillion is the Musk premium, which he's earned over the last, I think, three decades of doing what he says he'll do, sometimes a little bit late and making people money.

Utsav Somani: All of this is very, very forward looking. But I think the one stat that stood out for me was Anthropic will pay XAI, which is, of course, now polled in under the SpaceX company, $1.25 billion a month until May 2029 for 300 megawatts of compute from the Colossus One data center. And there might be some, I mean, revenue changes in this, but I think it's mostly, I mean, solid. I mean, it's extremely impressive. And I think he's accepted that Anthropic will be a leading lab and they want to partner with them and offer their compute services to them. But the other big one, given that we're involved in the venture world, Founders Fund and Ballor Equity Partners could clear both $60 billion in returns. And this is just insane. Founders Fund invested a $20 million check at $315 million pre-money in 2008. And Sequoia was reportedly one of the later entrants, entered at $20 billion. Google also sits on approximately 6 to 7.5 percent of the total company. So I think they've got a lot to gain from this. And should I do a recap of all the numbers that are?

Dhruv Sharma: Yeah, that's like the windfall. Yes, absolutely.

Utsav Somani: So all of this, of course, but I think what's truly highlighting, what's highlighting here is that Musk is pushing for 30 percent of shares to go to retail, which is typically 3x a typical mega IPO allocation. And they are, I mean, they've got a long list of people supporting them. Goldman Sachs, which Solomon slid into Elon's DMs, apparently, according to one article, is leading on this. Then you've got Morgan Stanley, which was Musk's go-to investment bank for 15 years, led Tesla's IPO and financed the Twitter buyout as the support and the co-lead for this.

Dhruv Sharma: Michael Grimes is, I think, the iBanker there.

Utsav Somani: He actually came back to Morgan Stanley because of this.

Dhruv Sharma: Why did David Solomon have to send them like a DM on X? They were doing the China trip together.

Utsav Somani: That was recent, but I'm guessing like all of this must be in works for like so long that it's just insane.

Dhruv Sharma: Yes, they said they were doing this apparently to win the Lululemon mandate. He showed up in track gear from there.

Utsav Somani: And the S-1, I think, funny enough, I mean, that time slide is just the most hilarious one. 28.5 trillion total addressable market excludes China and Russia, 370 billion in space, 1.6 trillion in connectivity, 26.5 trillion in AI. And that includes digital ads, consumer subscriptions. And all of this is, I mean, 100 billion larger than Meta and Google right now in 2025.

Dhruv Sharma: Also, all of the Earth's GDP right now all put together is about 100 trillion dollars. So, yeah, they're doing like a 30 percent cap.

Utsav Somani: Company. I mean, their mission statement is basically to make life multi-planetary and extend the light of consciousness to the stars. So if they succeed in this, of course, the time is unlimited, pretty much like he's basically said that SpaceX has unlimited time, to be honest, like the biggest time of any human led company. So I think this can be fairly exciting to watch. I think it's going live on June 12th, I think, or June 11th, trading under the SPCX ticker on Nasdaq. Should be exciting. What a day.

Dhruv Sharma: But, you know, a really funny joke I read about this was imagine it's the year 2100 and we're finally not just an interplanetary, but an intergalactic species and a human speaking with an alien and he's like, dude, how did he get here? And the answer will be by coding because Anthropic was paying 1.25 billion dollars a month to keep the launch business going.

Utsav Somani: Meme worthy. And something worth highlighting, India's peak power demands. It's super hot right now. So, of course, energy consumption is going through the roof, but it's not just because of that, like industrialization of our economy is also driving some of this record demand for electricity. 270 gigawatts was consumed yesterday and that's insane. And we surprisingly went through all of this without a hiccup, like thermal contributed 62.8 percent of this solar was 22 percent. Wind was 5 percent and hydro was 5.8 percent and solar contributing roughly 80 gigawatts during the afternoon hours. So it's just insane. And our solar capacity is getting built up. So superb, superb stuff.

Dhruv Sharma: I really love this, by the way, you know, after we now have a fully integrated power grid. So you could be sitting here in Delhi consuming power that was like power is fungible. So consuming power that was generated in, you know, in Himachal or elsewhere. And this is pretty cool. We haven't installed a power capacity of 500 gigawatts, by the way, that by no means, you know, suggests that you can draw from all of it. But we have more capacity than you'd feel.

Utsav Somani: Can you give me a scale of what this power consumption looks like, looks like in terms of just the global comparables?

Dhruv Sharma: Yeah, I was reading this somewhere. So 270 is the same as the UK and France and Germany and Spain consuming, you know, I mean, their power demand all put together. But guess what? It's also just 20% of the power that China consumes and China delivers because they're also heavily industrialized economy. They have the high speed rail system. So their power demand is also through the roof. And like, you know, we're the first and the second largest coal importers in the world. And they're transitioning rapidly to renewables as a V. But yeah, all this is like the scale is quite, quite fascinating.

Utsav Somani: Awesome. Let's end on that note. We'll see everyone on Monday. Thank you so much for tuning in. Stay safe. There's a heat wave, at least on the northern part of India. Stay safe and stay hydrated. See you on Monday. Have a great nice weekend. Bye.