Anthropic & OpenAI Have Changed What Moving Fast Means | Krishna M, Elevation Capital

When frontier labs can build everything, when moving fast doesn't mean what it used to, and when what was impossible can now be done in days, what do you build and how do you build it?

Krishna M is an AI Partner at Elevation Capital. He has seen the 0-to-1 journey twice as a founder and has also experienced what large-scale building looks like at tech giants like Meta and Cohesity.

In this episode, we discuss why a founder's fearlessness and ambition will define the destiny of the company far more than technical skills. Krishna shares how Indian founders have come a long way and now build like they know no fear. We then explore how OpenAI and Anthropic have changed what it takes to build startups, and what founder traits are needed to create outlier companies in this era.

Watch this episode to understand the new definition of startup speed and what to build when frontier labs can build entire product categories themselves.

01:00 From 2x Founder to AI Investor
03:37 Why Bet on the US-India Corridor
07:07 Can Indian Founders Build Giants?
10:13 What Makes Krishna Bet on a Founder?
12:33 The Founder Traits That Matter in AI
14:18 Why Speed Is the New GTM
17:52 The New AI Founder
19:13 How Krishna Builds Conviction
23:26 Why Elevation Bet on Portkey
28:20 How Elevation Evaluates AI Startups
31:17 The AI Security Opportunity
32:04 Your PMF Can Die in a Quarter
32:57 Why Open Source Still Matters
34:11 India's Infrastructure Advantage
36:51 Where the Next AI Opportunities Are
41:35 What Matters Beyond Revenue
43:36 Why India + US Is a Powerful Combination
44:59 What Cohesity Taught Him About Scale
47:13 Selling Data to Broke Professors
50:18 Why Two Weeks in SF Won't Work
52:06 Fundraising Trap Founders Miss
56:47 Outliers Aren't Well-Rounded
58:49 "We're Wrong More Than We're Right"

-------------
India’s talent has built the world’s tech—now it’s time to lead it.
This mission goes beyond startups. It’s about shifting the center of gravity in global tech to include the brilliance rising from India.
What is Neon Fund?
We invest in seed and early-stage founders from India and the diaspora building world-class Enterprise AI companies. We bring capital, conviction, and a community that’s done it before.
Subscribe for real founder stories, investor perspectives, economist breakdowns, and a behind-the-scenes look at how we’re doing it all at Neon.
-------------
Check us out on:
Website: https://neon.fund/
Instagram:   / theneonshoww  
LinkedIn:   / beneon  
Twitter: https://x.com/TheNeonShoww

Connect with Siddhartha on:
LinkedIn:   / siddharthaahluwalia  
Twitter: https://x.com/siddharthaa7
-------------
This video is for informational purposes only. The views expressed are those of the individuals quoted and do not constitute professional advice.

Transcript

Copy Transcript Copied

SPEAKER_00 0:00

Within like 18 months of journeying with elevation, you invested in port key, and port key has one of the phenomenal things that when I heard about the port key pieces, I don't see why customers are gonna need this. And then I met with Rohit, and on the way back, I'm like, dude, this makes a ton of sense. Everybody's gonna need it. Do you see more of this happening from India? Given how fearless our founders are, they'll show up in the valley the moment they have an idea. I think that has become a superpower. With your kind of background, being a T T Wed co-founder of Catherine working in the time quantity, do Indian founders like really understand infra what the US market needs? Infrastructure founders in India are actually working on Vanguard frentier problems. And as a result, they can actually build category leading companies. And for me, I think there's nothing more exciting than a category leader. Raising capital is not the right bar. I think building a business is the right bar. Sometimes raising money too quickly can be hi, this is Siddhartha Luwalia, your host at Neon Show and managing partner at Neon Fund. A fund that has invested in some of the best enterprise AI companies between US and your corridor, like Atomic Works, Spot Draft, CloudSec, and many others. Today I have with me Krishna Mehra. Krishna, welcome to the Neon Show. Thank you for having me. I have known you for a long while, I think five, six years ago. When I started my fund one. Yes, then I've got to do that. I remember. I remember. Then I preached to you, and I think both of us have come come a long way uh since then. Right. Remember it imagine that you know I'll be sitting here in Paul Auto and raising our fund for when I preach for fund one. Fingers crossed. Yeah. Yeah, but but uh I think uh uh God has been kind. And I think uh the corridor has blossomed, the India US Corridor that we both focus on. Your journey in fact has been very inspirational. You started out of IT Karakur, started Capillary, which is today one of the public companies in India. Uh right. In India has very few pure software public companies, right? At that scale. Amagi is one of them, capillary one of them. I think the bunch of others, but not yet public. Public, yes. Rate gain is one, yes. Yeah. So so thanks for establishing that category. No, of course. Yeah. Uh most of the credit to the capillary team. I've been out of the circuit for a bit. But uh yeah, it's been a crazy. Yeah, it's starting is hard, like 0 to 10 or 0 to 50 is really hard. Yeah. Alright. And and then you worked in big tech or you started your own company again, uh right in the US. So you uh and then you worked at Meta after that. Yes. Uh right. And post Meta, you have been a partner at Elevation, one of the top tier funds in India. And you are, I think, uh focused only on AI practice at Elevation. Yeah, unfortunately, I'm a one-trick pony. I don't understand much else. And I think uh uh within like 18 months of joining Elevation, um you invested in Port Key, and Port Key has one of the phenomenal exits in the Indian ecosystem. Port Key, phenomenal founders, like congrats on on that. Tell us me about your journey between like investing in AI your learnings. What kind of founder are you looking for? What's your thesis that you are looking for? What are the some things? You know, I'm asking a long question, and maybe you can process it, take time and break it down. What are the things that you think have worked for you? And what are the things that you are learning from the mistake? Yeah. No, see, I think the first part is uh that uh, you know, as I started this investing journey, uh, one thing which was very clear is I wanted to double down on the corridor. Um the way I look at it is in the first 10 to 15 years, the corridor produced uh a handful of winners.

SPEAKER_01 3:53

Yeah.

SPEAKER_00 3:53

Good winners. Postman, postman, charge bee, you know, a bunch of companies that have done well. Uh but I mean when I look back 10 years ago when we were in the early days on capillary, uh, there was hardly any playbooks for us as founders. Yeah. Uh they were actually, I remember there weren't no no PMs in India. Like we hired MBAs from B school and said, okay, let's figure out how to do PM. Uh there were no UX designers, like our first few US design hires were like graphic designers. Uh there was no GTM talent in India. Yeah. Uh we had no idea how to do how to sell internationally. So when I just think back of that time and compare it to founders now, founders are way more ambitious, founders are way more audacious. Uh, they take a flight and they are actually like in SF in two weeks and living like a local.

SPEAKER_01 4:50

Yeah.

SPEAKER_00 4:51

Uh so that change has been incredible. The other part is every part of the tech value chain, the quality of talent in India is top-notch. And you know, not just like EPD building talent, but even on the GTM side, India has actually built one of the strongest muscles for uh both marketing, inbound, outbound, even AEs who are closing deals sitting from India, uh, and the planning and all of the orchestration around it. So I think when I look look at that level of ambition of founders, I look at the quality of talent available. Uh you know, if we produced a handful of winners, I I see no reason why we should have uh many, many more in the next 10 years. Right? So the first part of the thinking is hey, uh this corridor is really going to prosper. Many people have doubts on that corridor. Why do you think doubts? I think uh I think the the part people forget is it takes time for things to build up. Uh, you know, an avalanche doesn't, you know, it starts with a snow, a small piece of stone rolling in the mud somewhere, uh, but that's the start. I think you know, you don't see how it will look at the end of the journey when it's a massive avalanche. I think people uh kind of mistake it. I mean, nobody calls the Israel US corridor, right? Like the idea is every Israeli company is like valuable from day one, and people are jumping head over here. All USVC owns a piece of every Israeli company. Exactly. But I think India is still early in that journey. And I think there's also been obviously India macro itself is a big investment thesis. So I think people just compare historical returns uh in each of the markets, and they say, Oh, I'm just gonna stick with whatever is more proven. But like, hey, like every investment company says historical returns are no indicator for future performance. What that means is I think we might also miss it significantly to the upside. But what makes you bullish? Like, why like why you know I'll play a devil's advocate though I invest in this order, but for audience, like we both can't be on the same side. Absolutely, right? Uh why can India produce $10 to $50 billion outcomes in AI space? Like that's the outcome that people matter in the US, right? Nobody talks about a unicorn anymore in in Bay Area. Fair, fair. Uh, and see, I think it's hard to say uh, you know, what the outcome, a $50 billion outcome doesn't look like a $50 billion at seed. Yeah, it probably looks like a half a billion dollar outcome. Yeah. And then you'll find that, oh my god, this is continuing to gather momentum and it is much bigger than it is. I mean, a lot of people also passed Anthropic at seed and series A and Series B and Series C. For them, anthropic was very expensive. 100 million, they were raising at 600 million values. Correct. In hindsight, I think it was very cheap. Very cheap. But you know, I mean, I think the idea is it was also very, very hard for us to predict, both to the upside as well as to the downside. And I don't think outcomes start looking like a $50 billion when you're raising the first round. Uh so from that perspective, I mean, I remember, for instance, Innovator, I remember clearly because I was involved. Uh, most people passed them at like seed and series. They were like, you guys are sitting out of India and going to build a healthcare company in the US. Initially started a services company, they started with doing some bunch of data aggregation, selling data packs to universities, like they did all kinds of random stuff. But I think the the level of hustle I've seen with those founders is incredible. Yeah. And I think that was truly the spike. Like it was very clear to me that these guys are about to do something in life. So, from that perspective, like it's easy to pass just on thesis. I think it's important to think through what this can become. And oftentimes people are missing that. But but as a fund, we discuss offline. You you invest if the thesis doesn't matter, then you don't invest, right? Because so it on the it depends on the stage. I mean, you're right. We uh we are by nature a very selective fund. So what that means is uh oftentimes we look for both the found the team's pedigree as well as the thesis and our conviction around all of it uh to truly drive conviction. Uh far more at obviously series A, where you have more proof points, lesser at seed. And you know, I think we need to look a lot less at that thesis by itself, at pre-seed, where you know you're you are taking up your founderback. But I think given the fact that we are fairly selective as a fund, I think we are much more indexed on actually having conviction across both those axes. Uh, but you know, I think the idea is we can always catch them given we are multi-stage. Oftentimes the thinking is hey, how is the risk reward looking now versus later? Can we catch them later? Where should we catch them all versus later? I think we have a lot of those discussions. Yeah. So for how many like in your last 18 months, uh, how many checks that you have done are pure founder bet right now when when you know you were not very clear on the thesis. Um I'm trying to think of it. I mean, I know one for sure, which was just a pure founder bet. Because I didn't need it. Uh yeah, so uh a company called Cynthio Labs, uh, they sell into pharma commercial and you know, started with some voice AI, but actually have a full suite of products for the pharma commercial use case. And this was, I think, a lot more of a founder bet than uh anything else because they didn't have revenue when we had invested. But I mean, I had closely followed the founder's journey for at least four or five months. What did you like about that journey? I think number one, so two things. Um, very deep expertise in pharma background. Uh one of the founders was earlier at McKinsey, had worked with pharma clients for almost 10 years. The other two uh have been you know kick ass engineers from Amazon. Uh, but I think that combination of tech plus pharma experience was very evident. And I would say we see teams with that pharma expertise, but oftentimes they're very stuck in their ideas. I think this team was just incredible at also their uh, you know, mental flexibility. So over that four years, I actually saw them pivoting and trying out three to four different ideas. Um, and you know, really breaking into large pharma. A lot of those conversations started happening. Um, they hadn't landed anybody as a real customer as yet, but we could see that momentum building up. So I think a lot of it just broke down to that experience with pharma, that domain expertise, but also the shipping velocity and ability to move very fast, which is not very common with these vertical teams. Yeah. So I think it was this combination of both of things, which was, you know, so it ended up being almost like a pure team back. Understood. So uh coming back to uh the corridor, right? And you said the $50 billion outcome doesn't look $50 billion at seed. Correct. Yeah. So what are the founder traits that make you move at pre-seed or seed stay? That this founder will probably create a large outcome. Yeah, so I think uh see, there is uh uh obviously all of this has been beaten to death by other people. So I will not spend as much time, but you know, uh technical skills, commercial skills, uh, you know, ability to build a team, um, ability to actually sell to not just customers, but to talent, to investors, all of that. So I think all of those things matter. I mean, we are all evaluating those things all the time. But I think one thing which has become even more important in AI as compared to before, exactly, is is actually just pure velocity. And what do you mean by pure velocity? Uh I mean, how quickly are they shipping? Are they shipping almost every day or not? Uh, what is the mindset of the founders? Uh, how much flexibility of mental models do they have? I think that has become far more important because in AI, speed is everything. Everything is moving so fast that if you actually sit idle, um, you will basically be uh you will lose your PMF within like a quarter, right? Um, so I do feel that that velocity and that sense of urgency is one of the most important traits in AI. As compared to, let's say, if I you know move the clock back five, seven, ten years, and we looked at SaaS founders. Um I think a lot of focus people gave on uh ability to have done GTM before, you know, things like that. I think people focus has moved a lot more to just pure velocity. And what what do you think has made this focus happen by the shift from GTM to to velocity? So there my general thinking is that when I when we look back at SaaS, right? Uh see again in the corridor, when we look at when new SaaS companies were started, the earliest SaaS companies in the US were almost in the early to mid-2000s. That's when they started. Of course, the floodgates opened late 2000s with the emergence of cloud and AWS and things like that. But like Salesforce was started 2002, 3004, somewhere in that time frame. Um, so you know, it was already being baked. There were enough players in every category. Um, so when we looked at that first crop of SaaS companies that came out of India, the ability to differentiate yourself just through pure GTM was far more important. So the early success came through the inside sales, outbound motion, uh, which people perfected and rightly so. Um the what has happened with AI FE Fast Forward is the field got reset. So suddenly there was a lot more white space and especially product white space for founders to go after, which is where I think instead of just pure GTM edge, uh ability to move fast and technical younger founders are actually seeing a lot more success. So a lot of it is because of where we are on the cycle. I think as corridor people, we hadn't even seen the early part of the last cycle. We are seeing the early part of this cycle. But I think that's what gives me excitement that even in the early part of this cycle, we are uh sure Silicon Valley is still ahead, you could argue. But if in the world of SaaS, Silicon Valley was five years, ten years ahead, well, it's probably three to six months ahead. And right now, what I'm seeing, uh, one of the best things 23-year-old IITN uh graduated one year ago, they're moving to Bay Area. Yeah, they are like, they'll get on a plane. I'm like when I I remember I moved to the US. Uh I had a hard time. Uh I had to hire a rental car, I couldn't get around without it. Uh, phones were not very common, so data plants were not common, so trying to find my way around was hard. Uh, I remember in the beginning I would ask people to drop me from one place to the other because there were no tabs, there was no Uber. Right? Uh so there was that we were very scared when you know we were making that journey. But when I meet founders today, oh my god, they're like so much ambition and zero fear. Yeah, they become SF native very quickly. Yeah, within within a few weeks, they're like, I'm I'm staying in dog patch, and you know, I'm I'm taking the Caltrain and Bart every day, and like, dude, like so I I do feel that the level of ambition has gone up significantly. Uh the fearlessness has gone up significantly. And I think that's what makes me very excited about the corridor. Because ultimately uh the companies that become very valuable, it is purely because of the founder's ambition and fearlessness. Um, a lot of other things come to play. PMF has to play, technical jobs have to play. But as the comp like the difference between a company with a half a billion outcome and a 10 billion outcome is just purely the hunger, ambition, and fearlessness of the founder. So on that, I feel Indian founders have come like way, way higher notched. Got it. And uh uh right now, let's say when when you were starting Capillary, like you, as you mentioned, GTM and everything that you had to learn from day zero, what are the tools which the founder can tap on day zero today that that were not there 10 years ago? No, so I mean, I don't think we have to teach GTM to founders anymore. I remember uh at that time we would do all these uh sessions where founders would get together and do playbooks on how to do GTM and things like that. I mean, younger founders I meet, I think the challenge was there wasn't any talent you could go and ask. There weren't advisors, there weren't people who had done it before in the ecosystem. Now it is way richer. I don't think founders are like going into a uh group session talking about playbooks, how to get to the US. They just take a plane, they come here, and they just go figure it out. So it is like whether a kid is moving from uh uh you know New Jersey or or uh Bangia or Bangalore, I mean they have the same skill set and they are coming here with the same network. So from that perspective, I don't think uh founders are at a disadvantage of any kind. Yeah, whereas I would think a kid from Bangalore, 23-year-old, is more hungry. Way more hungry. Way more hungry. In the last 18 months at television, how many companies would you have seen? Oh, I've probably seen thousands of companies. So uh I think uh and honestly, the big takeaway from me after doing all of that is how much higher the level of ambition and hunger is. And you have only bagged five, four or five, right? I mean, we are in the business of I mean, VC is a business where you say no most of the time. Yeah. And you want to be very selective. I think there's a lot more we are tracking. Hopefully, we get to partner down the line, but uh and we earn that right to partner down the line. But no, I think uh And out of these thousands, how many of them would you would you have taken second, third, fourth meeting? Whether you can share some stats. I'm trying to think. See, we don't generally manage stats like this, but I think I would say uh if I look at the funnel, if I'm meeting a hundred companies, I think we probably spend more time on somewhere between 10 to 20. We will do more deeper evaluations. And of let's say if we spend more time deep deeper evaluations on 20, we will end up writing a check on say uh one to two. So that's generally would be a broadly a ratio, but sure, some of it will vary a little bit. And and what would it take? Like you and Puri are the AI team at television. What would it take to get a check from you today? So again, I think it depends on we are very conviction-driven. So the question is what is driving that conviction? And I think uh we need to have answers on our like obviously, we're not looking for all the answers. I think the most important part is we are looking for clear signals which truly drive our conviction on why this can be a large business. I think the second part is even through that journey of meeting, let's say a founder will meet at least a few times, four or five times between that first meeting and by the time we are writing a check. Um, the one thing which makes a real difference is if we see that founder growing even through that journey, right? Because that really shows us hey, we can reference people, we can do a bunch of thesis work. Uh but at the end of the day, what really drives that conviction is that direct one-to-one interaction with the founders. And if we see that founder truly growing in every meeting, and you know, every time you meet them, they have a new set of answers, and you feel that this founder is actually figuring out more and more about the world around them in every meeting. I think that truly drives conviction. So, of course, there'll be some soft factors of various kinds, but I think uh just how quickly, what is the slope of the founder's learning journey itself is far more important. So So you would be any time, any point of time tracking around twenty, thirty companies very closely. Probably more. Probably more. Uh we've actually now built a full AI-driven suite to do a lot of our uh euro tracking and things like that. But uh hopefully we're tracking a lot more. I think there's a lot more problems. You built it. The team has built it and like I've also contributed by Friyan. Got it. But it's it's an amazing time to be a builder. Wasn't you tempted again to become a builder again? Um, so see, I think part of the thinking was when I mean I had done two companies before. Uh, then one of the reasons to go to big tech was I realized I've only done zero to one, I haven't seen any scale. So I wanted to see scale, and that's how I landed in big tech. Uh, got to see that scale, got to got that learning. Uh, and then I think again, it's that itch of like I want to do something new and I want to be learning all the time, which I think drove me to okay, I want to do again something different now. And I think venture was both exciting from that perspective because one, I'd been actively angel investing, so it was always a lot of fun working with other founders. So, from that perspective, plus I also get to learn a lot. I mean, I'm starting from scratch, uh, so it's again a zero-to-one journey for me. So it's been it's been fascinating. Right. And and tell us about what the thesis behind Portkey, how long you were in touch with Rohir, and what made you give you the conviction to cut a large check? Like probably it's one of the largest uh valuations that elevation came in. Yeah, yeah. So uh see, uh what's very fascinating is I remember when uh I'd heard about the Portkey thesis uh the first time, uh this was while we were driving to their office in the office when uh this is uh probably January of last, 2025. Okay, right? And I was like, I don't see why people are Kathnies are gonna need this. I don't get it at all. So I was very like, you know, uh while driving there, I was very dismissive of that idea in general. And then I met with Rohit, and on the way back, I'm like, dude, this makes a ton of sense. Everybody's gonna need it. I think the the clarity of thought and strategic thinking truly stood out even from that first meeting. Yes. Right? You could see that, hey, here are a set of founders who are actually seeing five steps ahead where the next right turn is, and then the next left turn, and then the next next right turn. You could clearly see that prescience from those founders. And I think that truly stood out. So, in fact, we uh you know we have always liked them as a team. I mean, I I I even I know even before I'd met, we had met them before, and we had liked them quite a bit at Seed as well, but couldn't get there at the time. So we in fact proactively invested time and actually understanding the space much better. Um and one of the things which was very clear is uh portkeys in the space of AI gateways. And there was a past concept called API gateways. And I think most people were confused about what's the difference between AI gateway and API gateway, tomato tomato. Uh I think we had actually did work to figure out what the delta is. We could clearly see daylight between why AI gateway is different, why it can be a much larger outcome, uh, where the security game can be played down the line. So all of that was actually very clear to us. Um, and then I think uh again, as we saw them execute and we saw them continue to build, um, I think the kind of customers they were winning, uh, they got some of the most discerning customers on the planet, right? So people like ToeDash, Epic Games, Albertsons, Community Bank of Australia, PG ⁇ E. I mean, a really good set of customers across the spectrum, right? Uh but I think feedback from all of them was also stellar. So it was three things that came together, right? Like one, we clearly understood the thesis area, why AI Gateway was important, why everybody's going to need it. So it's a big, large emerging category. What was also very clear was Portkey was a clear market leader on the enterprise segment. Again, the the if you look from outside, the space was very crowded. The people like Open Router, the people like Light LLM, the people a bunch of players even on the enterprise side. But I think we could truly differentiate this and figure out what the market map kind of looks like and see that hey, Portkey is actually the leader on the enterprise side. And then the third part was just how far ahead the founders were thinking. I think there was no reason to have any hesitation in backing them. And and when did you close the term sheet last year? Which month? Um, I don't remember the specific date, but it was somewhere in that September, October, 2010. And I remember between the term sheet and the time that we closed the transaction, I think the Putkey uh business had grown almost 50%.

SPEAKER_01 27:29

Wow.

SPEAKER_00 27:30

So significant amount of growth even between that time and you know, we continued to stay very, very excited. God, it was tracking a company over six, seven months. We were tracking a company for yeah, well even longer.

SPEAKER_01 27:45

Wow.

SPEAKER_00 27:46

And we kept meeting them every three to four months, understanding where the business is going. So I think uh part of it was also just kind of continuously building that conviction. So so again, coming back to the point of uh the velocity the founder shows. The velocity, every time is the founder learning something new, are there bigger and bigger unlocks they are finding in their journey? So let's see me ask, you know, how many people sit in an elevation investment committee? So uh in general, uh you know, elevation has a uh, you know, we we you know we keep in believe in including a large group. So we have a fairly large number of people. Yeah, I can't also get into it. Maybe we cut this part out a little bit. But maybe if you can share the process at elevation or making sure, sure. So the process uh at elevation is obviously uh, you know, we are organized by sector. So there's a team that is doing all of the B2B AI, enterprise AI investments, the team which is doing all the fintech work, there's a team which is doing all the health tech work. So we also become much more um experts at each of those spaces, right? Uh and we are getting better and better calibrated every um every day. So when people uh get introduced to elevation, they'll first be handed to the right team. And then the team will spend time on the or with the founders, your thesis. A lot of the focus is for the deal team, the team which is actually working on that deal, to actually get to conviction. We are very, very conviction-driven. We don't do any spray and pray checks. Um, so we do a fair bit of work uh to get to conviction. So just to think back, we're talking about portkey. I think we must have talked to at least 30 to 35 customers of Port Key and their competitors to truly understand the space, right? So we do a fair bit of work to get to conviction. Um, and then once we the deal team is at conviction, typically this process can take uh from that first meeting to getting to conviction, this process can take between maybe a week to a few weeks. Um, at seed stage, it's shorter, at series A stages it is longer. Uh, but we'll do a fair bit of work and we'll also meet the founders multiple times during the process, right? So every time we are sharing our work, we are seeing what the founder also has learned, how they react to it, etc. So it's a little bit a lot more back and forth, a lot more uh you know collaborative process with the founders. And once we get to conviction is when we'll take it to IC. Uh typically, I would say if the deal team has true conviction, I think you know, elevation is again our thinking is very conviction driven. So if DL team has conviction, very high chance likelihood of it's not consensus-based, yes. It is not consensus-based, it is based conviction-based. Right. Um, so yeah, the process, it's also amorphous because each company is unique. So very often, depending on a company, depending on the category, we'll figure out what questions do we want to be able to answer uh to get to conviction. God. So and what excited you about the space in case of Portkey? So, in the case of Portkey, uh, I would say the part which was very clear to us, this AI gateway space is rapidly growing. Uh, I spoke to a bunch of people we spoke to both in India as well as here, and we could clearly understand this gateway space is very fast growing. And people who didn't have a gateway said, hey, we actually need something like this. So, what was clear is this is going to be pervasive. Everybody's gonna need it. I think the second part, which was also very clear, is that uh even though it is more focused on map model management, model routing as a starting point, the security and governance use cases were actually become very clear to us. So if I go back and look at the CP, the the investment note that we had written, I mean, we talked about how this can be a security play down the line, and clearly that he is validated. Because they saw the same thing. They saw the same thing. So I think uh in our conversations with customers, in our conversations of with people at the academy, we could see that there are so many, you know, um security and governance optionalities in this business. Understood. And and maybe you can also talk about uh uh, you know, uh about the founders, right? Uh you mentioned a couple of things, like clarity of thought, strategic thinking. What did you see special uh in terms of execution, team building? So, one, I think the team was extremely scrappy. Yeah, honestly, to a fault. Um, I think like we I remember having conversations with Rowith, like, hey, you need to hire more people because you guys just don't have bandwidth to do as much as you want to, right? Uh, but they were scrappy to a fault. But that also meant that the product market fit was real, right? It was a team of whatever, 12, 13 people that had grown to it, grown Potke's business to almost like a couple million in Russia. By that time, other companies in India grew to like 50. So I think that was incredible to see. Uh, the other part around the team was also uh, you know, very strong set of generalists. Uh Rushank was the GTM guy, did you know across the board, lots of things. Their engineering team was top-notch. Uh, in fact, you know, we heard from their customers how good the engineers at Portkey were, both in terms of the core tech, but also in terms of providing support and things like that as they needed. So, overall, I think obviously a stellar team, but also extremely scrappy. And how do you evaluate product market fit now? Has it changed from the SaaS wave? So, evaluation of the product market fit hasn't changed. I do think that what has changed in the AI way is a wave is PMF is a lot more ephemeral. So, even when let's say we look at Port Key, they initially started with model routing, but they quickly expanded into observability, into um, you know, guardrails management, into various other things. But they expanded significantly into FinOps and managing budgeting, cost control, key distribution, things like that. Uh, and they had launched this MCP gateway because MCP was going, agent gateway because people wanted one place to govern all of their agents as well. So you could see how quickly the space is moving, and the market leader has to be willing to move with it. Uh, the other part I think again was um I think the articulation of the founders on how they're gonna build the business was also very clear. So, uh, for instance, most of the revenue came from their managed offering. But I think uh Rahit and Ayush were very clear that they actually want to open source even larger portions. Like if you look at a traditional business, you would say, okay, let's try to extract value. But these guys were like, we want to actually open source everything here, which we believe the value will be captured more at that governance and security level. Understood. Yeah, I remember in a podcast, Rohit mentioned a conversation that somewhere they were able to get into 0.5% of the AI traffic globally. Correct. Then North Star was also very clear that let's capture as much traffic as possible, we'll find more and more monetization levers down there.

unknown 35:35

Yeah.

SPEAKER_00 35:36

Yeah, which is very incredible. You don't see that in founders. Founders usually optimize for revenue and not their fault. The ecosystem pushes them to push them towards uh revenue to making your North Star, which is more uh input driven than outcome driven. Absolutely. And are you seeing more infra because this is an incredible acquisition, right? Completely from India, first new income to be acquired by Paul Auto Networks, right? An infra company. Do you see more of this happening from India? Like what advantages Indian infra founders have? Yeah. So see, my general thinking is if I look back, look at the India product engineering, design talent. I mean, we are working at the frontier, right? Like there is no AI product that people in India are not using. In fact, all the big model labs are hanging out in India for DAO and Mao Karmin. Right? The point is people are seeing the problems at the frontier, and they are often trying to solve the right problems. Let's say, as you compare to a vertical thesis, the challenge in verticals is unless you are here close to the customer, how can you build for verticals? But the person sitting in India actually has a very clear visibility of the latest and greatest problem that people need to solve to get the next unlock. So I think that is number one. Number two, the ecosystem around you also gives you the right feedback on is this working, is it not working, is this useful for me? I mean, all of that early feedback to shape the product is actually right there in their vicinity itself. Their friends are actually using this, building this, everybody's gonna have a bunch of founder fans who are using again a bunch of infrastructure tools, things like that. So I think the quality of feedback is also very high. So it helps shape the best in class product in some sense. Of course, monetization, you still have to be globally focused from monetization perspective. But I think being able to shape work on the right ideas and be able to shape the right product is extremely valuable. And then given how fearless our founders are, they'll show up in the valley the moment they have a right idea to come and monetize. I feel that has become the superpower. Versus, say earlier it was in the SaaS era, it was like, okay, there's a workflow, uh, somebody else is doing it, I'm going to do it better, cheaper, faster, and do an inside sales motion to get leverage on my competition. I think, at least in infrastructure, we feel that middleware, infrastructure, dev tools, all of those categories, founders in India are actually working on Vanguard frontier problems. Uh, and as a result, they can actually build category leading companies. And for me, I think there's nothing more exciting than a category leader. Uh, in Portkey's reason case, that was very exciting. That hey, clear chance to be a winner in the category. Uh, what are the five to seven themes that you are really excited about? So uh we talked about infrastructure. Yeah. I think um obviously that is a big area of focus. We keep meeting a lot of infrastructure companies. And infra, if you can specify what what all things you are seeing in infra. So, see, we are seeing a lot of stuff across the board.

SPEAKER_01 38:57

Yeah.

SPEAKER_00 38:57

Right? The parts which I think I would say we are less excited about is generally things around the SDLC right now. Because the challenges in SDLC, you're in the way of anthropic and open AI. Yeah. So hard to underwrite anything on what direction they'll take. Yeah. And honestly, my feeling is they'll do everything. Why would they leave any for other people? I do feel that opportunities will emerge in SDLC, but we need to let things settle down a bit. But other than that, I think generally, you know, anything to do with core infrastructure, we're also seeing stuff a few things on the data center side now. We're seeing things in data infrastructure. So tax in area, we've been excited about. Um, so across the board. Right? Um, I would say the other reason why infrastructure is also more compelling is at least what I learned at CoECT is uh getting infrastructure right is not easy. The kind of the complexity and number of configurations and systems that exist in the wild are far more varied. Yeah. And getting all of that perfectly tuned is a very, very hard problem. Um, so it's not something that somebody can just go and cloud code over a weekend. So, from that perspective, also feels more defensible. There's IP, things like that. Um, the second area is application layer. On the application layer, again, I would say we were if I move the clock pack six months, I think we were more excited. Uh, it's been a bit more confusing lately, given what can the modules do, what will the models leave is unclear. Um, but I think the reason for excitement there is that, especially at the application layer, either you're working on that final mile of delivery where you're taking the headache, taking the delivery, doing the work of the FDE, which people can't clot code overnight, or you're using truly uh bottoms-up uh you know distribution. I think both of those are exciting uh for different reasons, like either the stickiness or the speed of growth. So on the application layer, I think those two, both of those areas are exciting. Uh generally, we've been interested in verticals, though. In verticals, I would say we look for more cross-border teams versus uh so we want at least some founders who are very deeply seated in the in the domain, and then other founders who are building the thing, the tech shop, so to say. So we have to find more diverse teams on the vertical side. And then the area which also we've we've kind of stayed grounded and and keep looking for opportunities is the whole AI services thesis. Um has been harder to find high-quality teams there as we would have liked. So if anybody's building there would love to chat, we um uh, but there is so much happening, you know, on that front. Um, for instance, the whole migration of SAP from ECC to S4 HANA is a trend we've been you know watching closely. The in the Salesforce ecosystem and the ServiceNow ecosystem, there's so much development happening. Those guys are trying to go more horizontal, so there's more opportunity for new players to emerge. Uh, there's also opportunity for new players to emerge in the wake of new large platforms that are being built. So, for instance, a Sierra, um, Crestar, Decagon, OpenAI, Anthropic, uh doing the last mile work around this to let uh you know enterprises get value out of it. I think that itself is still hard. So, I mean, I think we are very excited about that too, but still looking for very good teams to and I believe with your kind of background, being a CTO and co-founder of Caplary and working in meta and cohesive, you naturally have a right to win in infra space. Yeah, I think infra I understand hopefully a little bit as well. So that helps for sure, uh, being technical. But I think, yeah, so by also excited that there's a lot of action there. Yeah. And and how how many companies that you meet, let's say if you meet 100 companies in a month, how many of them would be building in infra? If I meet 100 companies in a month, I mean I would say there's at least 30, 35 million. No, there's so many inside of need other than that. Yeah, yeah. Across across Bay Area in India, there's a bunch of action in infra. But it's also harder to succeed in infra because people will not use 50 different infra. Correct. Correct. So you have to be differentiated, you need to have a distribution advantage, um, maybe some early strong open source traction, all of those things are very helpful. Okay. And uh because monetization happens very late, let's say. Uh uh in in case let's say you would have evaluated port key at seed stage at earlier stages, right? Monetization very, very late and hard. So, what are the other signals that you are looking at infrastructure? So I think um I would say that uh we like obviously monetization is going to be very early. We want to understand willingness to pay, but not as much actual monetization, but more from a buyer's perspective. What is the willingness to pay? How critical is this piece of infrastructure? Can they is it a nice to have or is a must-have? Can they live without it? So, all those questions are we what we want to answer? Uh, how will the pricing scale? So we try to answer those questions, but don't need real proof points. I think we'll also do a lot of work around it. Hopefully, also help the founders. Uh, but I would say the important part is hey, what is that unique IP we are building? Why is it differentiated? Um, what does the rest of the ecosystem look like? Um, what are the hard problems we are solving? So I think having a good understanding of all of those things is important.

unknown 45:05

Okay.

SPEAKER_00 45:05

So even if you look at like Portkey, I think when as a team we had looked at at seed, at that time, I think they weren't quite building what they were building and the open source distribution ad kicked in. But I think by the time the open source distribution act only kicked in. Got and uh do Indian founders like really understand infra uh what the US market needs? I think on the infrastructure side, probably the closest. Okay. Right? If I look at, say, uh uh, you know, let's take the case of financial services. I mean, India financial services and US financial services are very different, right? If you look at uh horizontal software on what accountants need, what uh you know, even the average person needs in an office, it might look at cultural understanding is very, very different. Very different, right? But when it comes to engineering, when it comes to core infrastructure on the developer side, it's actually very similar. Um, the it's it's like a 1920 difference between the two. Uh, it's not the most actual platform engineering teams, even for US companies, are probably sitting in India. So they are the ones who are dealing with a lot of this infrastructure. Uh, when you come to things like data engineering, if there are new challenges in data engineering, how do you deal with that? If there are new challenges with uh managing pipelines, if there are new challenges with uh, say managing production stack of GPUs, founders in India are actually selling, you know, working on very frontier problems. So I think that is the reason why you know we are more excited. And you worked in Cohesity at a very large scale, which was a startup not so long ago. Right. What were your learnings there working at Cohesity? No, I I mean I think Cohesity was an incredible learning experience. Of course, Mohit has built a remarkable enduring company. So the uh the most fascinating part is you know, I got to see true enterprise scale. Uh, our largest customer was a 50 million a year account. So, how do you actually deal with it? How do you solve those problems? The diversity of use cases, the diversity of infrastructure they also have, and all of the core problems KohCity was solving when you so Koh City had this web scale file system that was being used for secondary storage, so a lot of backup data protection, things like that, but then built a lot of security and other things on top of it. But when you look at it, um, you know, let's say the company started with focus on VMware, SQL Server, the even the number of configurations of VMware are so many. Like how many do you actually do? How what do you leave out? How do you prioritize? Those are truly hard problems. And in fact, I mean, uh I would say what was very interesting was uh even like six or seven years into building the product, we had, for instance, not managed to cover all the use cases even for a VMware. So, how do you then prioritize based on customer needs? In some cases, whether the customer is willing to pay for it or not, um, how do you provide support at that scale? Because something with some customer or the other is breaking all the time. And how do you build a process to actually handle upset customers? I think uh just got to see a lot of how to deal all of it with scale. And and Mohit is building now on another company. He's he's I mean, hey, one unicorn is hard, one decacorn is hard, one decacorn is hard, two decacons are impossible, and Mohit is building his third decaco. So another beast, like very uh very different, difficult to work on. Like Mohit is a true beast, like it's hard to compare anything with anybody with him. What has been you're learning investing in innovation from very early days? And how did the company pivot? Yeah, so see, I think the even from the very earliest days, just the level of hustle they had, um, willingness to move fast and really go all the way for the customer was like super evident. Healthcare has always been a very difficult market in the US. They have it's been uh so so it's interesting because when uh I had first met them, they were sitting in Noida and they were actually doing this uh, you know, data packaging and data sales to universities, and they had scaled to like you know uh almost a million ARR just doing that. And I was like, oh man, if you guys manage to get university professors who have no money to go spend a million ARR, you must be doing something right. And I think that level of hustle was very clear, but that also gave them the entry into healthcare because what's your right to enter into healthcare? By doing that, actually, what happened is one of their professors, I think from MIT or however one of those places was an advisor to a health plan whereas value-based care became important, those guys were looking for help, they went to that advisor, and that guy pointed them to Innovator. So that's how Innovator got into the problem. And kudos to uh, you know, Abhinav, Kanav and Sandy. Uh they, the moment they got into healthcare and realized this is such a large problem, they were actually willing to shut down their old business and just focus on it. Right? The so one is having that mental flexibility of go from one problem to another and going where the opportunity is. The other part which I think is underappreciated, is just healthcare is incredibly complex. The number of EHR integrations, getting them right is very messy. Uh so for the longest time, Innovasur was actually uh you know seen as a services company before Palantir was cool. But all this while they've been doing an FDE model, right? Uh I think So Universal was also doing FDE model before it became popular. Before it was called FDE, right? So, because the thing is, when they got a customer, then very often you would actually need to go build a bunch of corner cases, things needed around it. The data that comes in is different for each EHR, but also each version of the EHR. So Epic version 2.1 and Epic version 2.2 and Epic version 2.3 looks different. So then you definitely need to do almost service-y kind of work in the beginning to onboard a new customer because even the EHR integration is very different. Correct. So, you know, they had an army of people doing it, and you know, a lot of people scoff at it, but I think that turned out to be their big mote and advantage. But even now, I think they are integrated with almost like 100 health systems in the US, which is no mean feat. So, part of that willingness to actually go focus on what the customer needs versus what investors are telling them, I think was also a superpower which we often you know discount. And what's your observation been on some of the Indian founders that moved to SF? Their process of discovering big problems which are still unsolved. How do they do it? Yeah, so see, I think the uh there's something magical about obviously being in the center of all the action, right? Like you get, I mean, very often you'll actually find that uh a problem that you think is very small sitting somewhere else, and you know uh Bangalore or India is not just one of those places. I mean, you could be sitting in Paris and not have any idea of where the problems lie. So, I mean, at least the Bangalore ecosystem is far deeper and richer than Paris, I presume. Uh but the way I look at it is uh that that appreciation for problems that are going to become much larger in the future, I think is much higher. Uh, nowhere else could you say, hey, I want to actually build a new lab, a new lab to do research and raise like half a billion dollars, right? Uh so the risk-taking appetite of the ecosystem is also massive. People have made money on by taking these kind of risky like anthropic, both investing as well as working there, right? So that I think makes a big difference, and people are willing to uh make many more leap of faith than than people are elsewhere. So I think that's a real advantage. And some of of course, a lot of those ideas die. Obviously, we don't get to hear about a graveyard. There's a graveyard of companies, but then of course, I think once you find the problem that really sticks, I think that becomes a much larger problem in the future. So I do feel that you get to see the latest and greatest, but also begin to dream and think a lot bigger. And what would be your advice be, you know, to first-time founders, 22, 23 year old looking to move to SF? How to do that? So, see, uh the as compared to again 10 to 15 years ago to now, I think the level of support system, the level of uh you know, help you will get is a lot more. All you need to do is ask. So there is a ton of communities, even in SF and you know, community spanning FSM and Bangalore, where you'll find help, how to find an apartment, where to stay, where to live, what to eat. I'm vegetarian, what should I do? All of those questions, you know, much easier to find answers today than it was 12 to 15 years ago. So I think the first part is I think probably less to do with SF, but more with introspection that, hey, what problems am I excited about? Because again, uh entrepreneurship is a long journey, often very, very rocky and very painful. So you might as well pick a problem that you're passionate about, number one. Uh no point in you know faking it and finding that, oh, I'm stuck with a problem I'd never actually cared about. The second part is once you have that and you get even some signal about whether it is directionally correct or not, uh, definitely worth spending extended time in the bay. I I don't think like short three, four days, two-week trips help because you need to let serendipity happen. Yeah. You need to come spend some extended time, go to a bunch of events. Many of those events will be low quality, but you'll find some that are high quality and really make a difference.

SPEAKER_01 55:29

Yeah.

SPEAKER_00 55:31

So I do feel that excess spending some extended time is very helpful. And the other part is I feel for that, unlike earlier, there's a lot of uh you know, investors who take that very early back. Like uh Nyon obviously is one which has done an incredible job there, but there are others too. So I think there are enough supporters that people will find if they're actually willing to hustle and actually do the hard work.

SPEAKER_01 55:55

Yeah.

SPEAKER_00 55:57

And what is your learning on not do's because now you have you are a through valuation investor in many companies. I've previously been an angel across 45 plus companies. Yeah. I mean, I think for founders, um you know it's it's hard to say because everything changes every few years. But I still believe that, hey, don't pick up a problem that you're not, you don't think you'll solve for 10 years. Uh, because you're gonna have to give 10 years to an idea. Uh so if you're half-assing something, I don't think it's worth it. Uh, especially if you get some early success in terms of getting funding and things like that. Sometimes that is actually much worse than the counterfactual. Because if I let's say I start with an idea I'm not passionate about, yeah. And I don't think it's gonna work, but maybe like a big name VC comes and invests in that, then there are two problems. One, I start believing in that idea, which I originally didn't believe as much in.

SPEAKER_01 57:05

Yeah.

SPEAKER_00 57:06

Uh, and I spent years working on that idea, which was essentially a waste of time.

SPEAKER_01 57:11

Yeah.

SPEAKER_00 57:11

Right. So I do feel that sometimes raising money too quickly can be counterproductive, not for every founder, but certain archetypes. Uh, right? Some like Vijaya of Atomic Work, I think incredible founder obviously knew what he wanted to do right out of the gate. So it makes sense. But there are many others where people will actually be able to raise some capital. So raising capital is not the right bar. I think building a business is the right bar. Uh and I think staying truly focused on that is important. Because if customers uh follow you, then the investors will definitely follow sooner or later. They will follow. But the challenge is sometimes you'll get uh investor validation, but customers don't buy it at all, and then you end up wasting a lot of time and energy. And now, let's say if you've raised, you know, just let's make it this up five million dollars, your expectation for the next round is also a lot higher. Yeah. So I think you're also raising that expectation. And if at the end of that five million you're not at 2 billion ARR, then nobody's going to write your CJ check. So sometimes I think being more scrappy very early on, and actually finding an idea that you care about and the customer cares about is more valuable than trying to get that early validation of funding. Agree. So Krishna, across the you have seen many large outcomes now. Obviously, Capital was a great outcome, innovation, fantastic outcome. So, what differentiates a good outcome from outliers? Like Port Kia would call still an outlier. So I think again, I would put all of these as hopefully outliers in my own. Oh, it's an outlier. Yeah, yeah. So I feel the the uh the uh uh it's it's hard to define because it's not the same answer in each of them. Uh but I think the by the time you get to that outlier outcome, uh and maybe it's hindsight is 2020, but I think the very clear spike and the strength of the founders is very clear. Yeah, I mean, these outliers are created not because of the founder checked all the boxes, but because the founder was just 10x better at certain things. Uh-huh. Like I think capillary has been like a true strength of resilience. Yeah. Um, and you know, I mean I think Anish has had to has lived through most of it. Yeah. But you know, most people had written off capillary several times, and you know, it has been, you know, incredibly successful now. But it took was a long time in the making. Um, I just incredible velocity and hustle. Yeah. There's the level of hustle has been incredible. And willingness to learn like a domain that they never knew about, right? So I think that is also commendable. In Portkey's case, at least I the way I look at it is that that product and strategic thinking was a clear spike that we could see right from the beginning. So I do feel that that that spike is always very clear. And ultimately, that is what we are drawn to as investors. So you are you're saying the clarity from a founder. Clarity, what is the true strength that is truly showing me shining through is what we also have to look for from the earliest of stages. Yeah. Thank you so much, Krishna. This has been an amazing conversation with you. No, thank you for having me. Thank you for being so candid. All right. And I learned from you what what you saw at those earlier stages, you know, which is now obvious to everyone. I mean, I think we are wrong more often than we are right. But it feels good it feels good to be validated once in a while.

SPEAKER_01 1:00:55

Thank you. Awesome.