Amar Varma on AI’s Power Law, OpenAI Alumni Networks, and Building Mantle Against Incumbent Carta

Amar Varma, CEO and co-founder of Mantle, joins Mark MacLeod again, but this time to discuss the dramatic concentration of venture capital, with 41% of 2025’s funding going to just 10 companies. They explore how OpenAI alumni are creating a new ecosystem of startups, why traditional VC pattern recognition is breaking down, and how AI is enabling unprecedented capital efficiency through “seed strapping.”

Amar also shares his strategy for competing in the cap table management space against established incumbent Carta, explaining why moats now depend on human decisions rather than features. This wide-ranging conversation covers everything from nuclear plants powering data centers to James Bond villains becoming reality, offering practical insights for founders building in the age of AI.

Let’s get into it…

Mark MacLeod:

In this episode, I sit down with my good friend Amar Varma. If you’re a fan of this podcast, you’ll know that I have Amar on periodically for these Tim Ferriss-inspired random episodes. Amar and I always have these really interesting conversations, typically on a tennis court. And every once in a while, we’re like, man, we just got to bring this shit to the podcast.

So we covered lots of ground today, talking about the power law of venture capital. That’s always been a thing, but it’s more so than ever. 41% of venture dollars so far this year went into just ten companies. So we talk about the implications of that.

He had some pretty fun observations about how all these James Bond villains in different industries are becoming true in real life now. We talk about VC pattern recognition and whether that’s still a thing. Lots of chat about AI, and you may be running a company that has an incumbent, like, in a category with an incumbent. Amar is doing that. And so, the last thing we did was talk about the implications of that.

So lots of really good stuff here. I really hope you enjoyed the episode as much as I did.

Amar, welcome to the Startup CEO Show. It’s a true pleasure to have you here. How are you doing today?

Amar Varma:

I’m doing great. Always a pleasure to be here, Mark. Thank you for having me.

Mark MacLeod:

So, for folks following the podcast, you’ll know that Amar is by far our most frequent guest. He’s a good friend and a country neighbor for a good chunk of the year. And we end up having these very interesting conversations, often on the tennis court. And then we’re just periodically like, hey, man, this shit should be on the podcast. And so welcome back.

And maybe to kick things off, you and I were just kind of talking about a recent observation that Jason Lemkin from SaaStr made, which is like, so far this year, 41% of all VC dollars have gone to just 10 companies. And so venture has always been this, like, the power law of returns, right?

So, for folks who don’t know, most of the returns in a venture fund are concentrated in a very small number of their investments, and those pay for all the other investments. So returns have always been unevenly distributed, and capital has always flowed to the winners. But now Jason’s saying it’s more polarized than ever.

And so just to put some numbers on this, and then I want to turn it over to you for your thoughts. So far this year, through I think this was July, $197 billion in venture funding, and 81 of that went to like 10 names. The names we know, like anthropic, OpenAI, et cetera. So yeah, what are your thoughts on that?

Amar Varma:

It’s a new game. It has changed and, in its purest form, has become the search for the Power Law. And as time is tracked every five to seven-year period, every vintage year kind of thing, the track’s gotten tighter and tighter and tighter and tighter, and there’s the search for runaway names that occurs. And we’re seeing it at its most extreme form to date, and just our kind of collective histories to date.

And it’s happening for a few reasons. One, ultimately, the expectation and valuation of these power law assets are bigger than we’ve ever seen them. They’re like we expect, they’re lagging the public market valuations.

So now you have multiple trillion-dollar-plus valuation public entities that are driving demand for their stock price, and their stock is ultimately driving the stock prices up. And then ultimately in their, I call it the magnetic pull or you know, planetary gravity pull that they have in their ecosystems, where they can pull along infrastructure spend, application spend, the traditional cycles we see it’s just larger than we’ve ever seen.

You know, it’s to the point where old nuclear factories are being turned on to power some of these capital expenditures. So, it’s back to Sathya saying I’m good for my 80 billion in 2025. Like this is kind of the comments we’re seeing.

Mark MacLeod:

Was it Microsoft that got a nuclear plant to be powered back up?

Amar Varma:

Yep. The one in New York State. There’s another one, I believe, in Ohio that’s being fired up, and in Iowa. And everybody’s looking at everyone now, right? What’s there, what infrastructure is there, what kind of reactors were in there, what’s safe, what’s unsafe. There’s all this analysis, and it’s partly…

Mark MacLeod:

Your sister is a nuclear engineer, so maybe she’s in the safest job category around right now.

Amar Varma:

Possibly. And this is one of those arenas where there’s unlimited demand driving unlimited exploration of this stuff. So it’s kind of irrelevant again in terms of Homer’s job. So that’s kind of one of the things I was looking at.

Mark MacLeod:

You know that parallel thing has always been the case, right? And I wrote about this a few weeks ago, with Lovable going from like zero to a hundred million revenue in eight months. And it used to be like the best in class for SaaS. And this is again another Lemkinism, where you would triple in one year, you would triple again, then you would double, double.

So like, I don’t know, you have 1 million, then you go to 3, then you go to 9, then you go to 18, then you go to 36. So, like four years in, you’re at 36 million. Whereas lovable just tripled that in eight months now. And that’s just an illustration of the non-linear, completely explosive growth that’s possible with these new, completely disruptive companies.

And anyway, my observation about all of this was just, like, as a VC, you can’t unsee this now that’s possible now all you want to do is back the next Lovable before it takes off, right? And so like, I feel like the bar for these merely mortal companies, like oh, oh, you’re just going from 3 to 6 million? Like, ugh, I don’t care. Like, I don’t know. What do you think?

Amar Varma:

Yeah, this is great. So let’s take a step back into the power law. So the greatest, the goats that we’ve seen, and there’s a handful of goats that have been around for multi-decades, traditionally came from sizes of funds that were in the few hundred million dollars. Now you’re seeing across the board people sitting on billions of dollars in funds, and then they have these growth vehicles that are multi-billion dollars, and all of this capital has to go to work.

And unless you’re in one of the 10 names of this vintage, you won’t have the ultimate returns, DPI, whatever metric that gets used to raise the next fund next time. And that’s kind of where you’re seeing this massive chase in the ecosystem. And the barrier to compete with these folks is getting harder and harder and harder.

So I think that’s just kind of the Power Law that you’re seeing in terms of size. In terms of the opportunities, Lovable, I loved your article on that. No pun intended. I think it was just highlighting how quickly this thing can happen. You see it, and if I take a step back, what’s happened is the substrate and the barriers to entry to get into the substrate of using AI is the equivalent of using search on a browser. It’s a box. So that has opened up a huge swath of the Internet population to start try and use and you start to compound your learning and you start to figure stuff out and then once it’s unlocked is there was always this cast of people who wrote code, didn’t write code, and there was two sides of a line and that line’s been blurred now.

And everybody feels like everybody should feel like they can write productive code that goes into production, and that’s kind of opened this floodgate to allow people to do it. And you see how much people are willing to pay to get that stamp to do it. They’re willing to pay Lovable that kind of money in that short period of time. And incredibly, they’re not the only company. Like you see.

Mark MacLeod:

No, that’s right. You got Replit; there are probably others. Yeah.

Amar Varma:

And you see these companies come up and up, and then the question ends up being what’s the network effect? What’s the mode that you create with these tools? Am I just going to use the next version of Lovable for the next versions of apps, or am I going to use this? Like, it’s just a series of tools, and if you’ve used them over multiple projects over a longer timeline, you start to see, okay, clearly Copilot and get something you’re going to use in the Microsoft suite you’re going to end up using.

So that’s kind of one of the ones where there’s just a built-in network effect in the moat. Do these other folks have it? Can they be long-standing? Because there’s one thing to chase the growth from 0 to 9 figures, which is amazing. Can you get from 9 figures to 10 figures in that period of time? Is there a growth or is there a petering off, and you’ve bought at the top? These are some of the questions that are there in terms of some of these applications.

If I switch back to the infrastructure investment that’s just steamrolling, that’s still just compounding, and I don’t see slowing down. I feel like the cost per token is going to get to a place where that infrastructure spend is just going to make even more sense. You’ll see them just continue to some Moore’s Law equivalent for making the cost of tokens cheaper.

Mark MacLeod:

You don’t think the big providers, once we’re hooked on tokens, are going to jack the price? They’re just crack dealers here. Your first hit’s free, and then afterwards we’re going to hit you up.

Amar Varma:

You know, it’s always a worry with different types of players in the system, right? But if you look over time, cloud providers were there, and they were forced to bring down the cost of their services, and like smart people, they added new services that you paid more money for.

So getting into the cloud was easy. Firing up a new service costs you money, and leaving one instance for another instance was incredibly expensive. So those are the types of things where they’ve just, they know this stuff, they know this game, and they’ll use Porter spy forces to navigate you down the gauntlet.

Mark MacLeod:

So I’m curious, through your vantage point, not just as a pundit, you’re now a part-time pundit, but as an operator, as a CEO. Because, on the one hand, you see these companies raising just ridiculous amounts of capital, and then on the other hand, what they’re building is actually enabling most CEOs to be more capital efficient than ever.

Like, there are so many headcount functions that you no longer have to scale linearly as revenue grows. And thinking about developers, support staff, et cetera, there’s a trend with many companies now, seed strapping, I don’t know if you’ve heard this term, which is sort of like skipping the seed round and just bootstrapping your way past that, which AI enables, right?

It enables efficiency like we’ve never had before. So there’s sort of like these two opposing things happening. There’s a cohort that’s being super capital efficient to the point of bootstrapping, and then a cohort, a very small cohort that’s like sucking up all the capital in the ecosystem. Do you feel pressure to go one way or the other? Like, how do you react to that?

Amar Varma:

Yeah, it’s a great observation. The ability for all of this capital investment to enable seed strapping, I think, is the term that you’re pulling up, and I understand it; it makes intuitive sense. Now that you bring it up, it’s amazing. If I look back in history on two vectors, one is thinking of maybe some of the home automations we’ve had over time, dishwasher, laundry, those types of things. You didn’t go back to washing your clothes by hand unless you had something very delicate, maybe you’re washing fine china that doesn’t fit in a dishwasher, things like that.

And that’s maybe the crude analogy that’s happened here. If you switch into maybe more of a tech lens. You know, the huge cost of infrastructure projects, big physical chips, systems, over time led to creating lighter-weight tools, open source tools, software, and databases. Ruby on Rails was one that was influential for us, maybe 17 or 18 years ago.

And when it kind of just made web development way, way more approachable, and you start seeing that play itself out at the scale it’s playing itself out now. And the cost to start companies that went from tens of millions to millions to hundreds of thousands, to ostensibly whatever time you can dedicate to this stuff, the marginal cost is phenomenal.

I’ve never seen it in history. I speak at tons of classes, entrepreneurship classes, and today is the best day in history to start something. There’s never been a lower barrier to getting started.

Mark MacLeod:

I remember my first startup, I was finance theater for, they had to raise 3 million, mostly to buy Oracle databases and shit just to get started.

Amar Varma:

Yeah, look at Oracle today.

Mark MacLeod:

Pretty wild for such a big company, right? And they had crazy guidance. I think they said they’re going to grow their cloud business by 70% next year or something. Mental.

Amar Varma:

When you see these data points in public companies and you put together that whole ecosystem of the various vendors that fit into a deployment, the infrastructure, the software, they’re creating the groundwork for this next revolution of companies that are coming through. You’re seeing it in the Lovables and the Replits and the Cursors and all this. That category of companies, there’s a bunch of others. So not trying to single them out.

Mark MacLeod:

Do you feel tempted to raise more capital than you need in this environment?

Amar Varma:

This is very interesting. So I’ve been looking at our team here at Mantle. We’re around 20 people, and for the amount of throughput and things that we’re doing versus maybe a company like when we did Autonomic, maybe seven, eight years ago, it’s probably a third of the staff we have versus the output that we’re getting versus back then, maybe even less.

You know, I’m just trying to think out loud; it might even be less than that. And that’s just a tribute to the tools, tribute to our team’s ability to execute all those things, but we’re not unique in that kind of equation. The numbers might be different, but everybody has that opportunity.

So the question is, what are you raising for? And if you’re raising for the ability to land grab market share, like the traditional venture types of things, that’s really where you’re having to look at the money. And that’s why you’re seeing the barbell distribution, or basically the plot point on the chart of everybody in this category, the 10 companies getting 41% of the dollars, right?

They’re just going and land-grabbing this whole thing. And then following on to it, I’m in an application layer. What parts of the applications get subs by the models, and how do you interact with the data? And fundamentally, the question I ask myself, I ask all the entrepreneurs, is what is your moat today? Because it’s certainly not the ability to write software or create great systems of code.

Mark MacLeod:

It’s not features.

Amar Varma:

It’s not features. It’s the human decisions that impact the success of your product. When I break it down for us, it’s our customer choosing to use us to do something.

Mark MacLeod:

So it’s a human connection to you that is maybe implied. Like, there’s a brand element there.

Amar Varma:

Yeah. Without that human checkboxing, signing, or agreeing to do something with our company, if that didn’t exist, the moat would be… And I have a blank. It’s probably a longer, deeper philosophical discussion, but it’s like, what is that moat? Because if you don’t have the moat, you might get a really great liftoff. But if you can’t sustain the liftoff, you don’t end up being a longstanding business. You might end up being an ephemeral thing that goes and comes.

And I’ll just give you a quick example. I use this one, and maybe even used it before here. But we went to brunch, me, some of the kids, and they really like playing trivia at one of the places we go to, the Solo House just down the street here. And we went once, and somebody was using the board game. And I was like, they’re like, I was like, no problem. I went in, I created a trivia app in, I don’t know, all of 20 seconds. And we played trivia for at least an hour and a half, having the time of our lives for brunch.

And at the end of that session, I never looked at that thread again. And you know, it was gone. It’s ephemeral, it’s missing. It didn’t need to be a standalone app in the App Store with distribution and ads, right? It was gone. I was like, oh, apply that across consumer applications, enterprise applications, and then just SaaS is really one that I’m really taking a keen look at and being like, what is it that people need to do in the interaction? What’s the human in the loop?

And then after that, when the human in the loop is no longer used or required, what is that user interface? Is it voice? Is it a text box? And then when these decisions are being made, not so much by just AI as we think about it on a machine, but in physical AI, now we’re in this other one.

So if anybody uses Waymo, if you use self-driving in any of your vehicles, like, you are using physical AI today, and it’s making decisions on your behalf that are pretty important.

Mark MacLeod:

So it’s kind of life or death decisions.

Amar Varma:

Your life’s in their hands. So these are the types of things where it’s like, okay, like, where’s the end of these things? And it’s not so much the end, but what can you do as an entrepreneur to grab onto one of these waves, ride it through a quest, and see where you can come out on the other end.

Mark MacLeod:

Is the role of design changing for you in light of all this?

Amar Varma:

Incredibly, I love this. So we have the cast of designers that we’ve worked with over the last couple of decades through Extreme Labs, Pivotal, VMware, Autonomic, and all the companies that people have gone on to, I think, like Ecobee if you went there, and then Ritual. And they’re just spread all over.

Some work here with us again, and their ability to articulate what the product or outcome should be is usually very, very, very good. Like a cut above. And they’re able to leverage the models better than the average, just because they’re by design rules, they’re very articulate, and very opinionated, and they can use the models to output things at a pretty high level.

They’re now also contributing code for quite some period of time. And it’s getting better and better and better, it’s really impressive. So the design has gone from just a visual aesthetic and a user experience aesthetic to a, let’s build that. Let’s not just have it in a prototype; let’s build it, let’s have it on a local host. Let’s put a PR through there and let’s actually get it into production. It’s really impressive.

It’s taken on a whole expansion of the role, and the interaction has never been more important because if you don’t get that first experience tangibly correct, like I use the term, when you walk out of the Apple Store with a box handle and there’s a machine, you have a really nice feeling carrying that box. Like you have to have that kind of experience when you enter someone’s product, or else it just falls a bit flat. So it’s never been more important.

Mark MacLeod:

So going back to this, the 41% and of course, it’s going to the big horizontal players, OpenAI et cetera. And you made an observation, and that was one of the catalysts for like, oh fuck, we need to do another episode, which is like there’s a new kind of ecosystem that’s being run off OpenAI alumni. So I would love for you to just expand on what you’re seeing there.

Amar Varma:

Yeah, look, I think as previous lives have been active investors, and so do some personally, but it’s like there are ecosystems over time that created a lot of next-generation opportunities. If you go back in time and look at all the folks who came out that created a lot of the chip companies in the late 80s, 90s, they all came out of AMD and some of the other long-standing chip companies. You know, then all the Telco folks came out of Lucent.

Like, if you just kind of go through it over time, and you know, then PayPal for software. Yeah, everybody started coming out. PayPal has it in, probably the best-advertised alumni network on this stuff. You can look at doing beat a dead horse here, but there are just a few that have been great, obviously, a lot of Google alumni, et cetera.

But I’m looking at what’s happening now in folks who have come out of model-based AI companies, and they move. The speed at which this is moving is unprecedented. The who and the tag you can give somebody in 30 seconds, it’s like, oh, hey, this person worked on this model doing this piece, or owns this part of the thing.

For example, this person who worked at OpenAI was part of the training and a mixture of experts. Like all of a sudden, if you could follow that nomenclature and that tagging, you’re like, okay, I can visualize where and what they can accomplish because they have fluency, they have adaptability, and understanding.

And you’re like, okay, that’s a person who can move this forward. And if you can bet on people who are going to move this state-of-the-art forward, you’re going to find winners in that. And that ecosystem is happening in a super tight, super fast network. And if you’re not in it, you’re on the other side.

And it’s kind of analogous to, maybe even more so. I think even this week was the YC demo days. It’s like every company’s funded before the demo day, right? You didn’t approach them, you didn’t have the list, or they didn’t find you. Good luck with that. So that’s kind of how the game is evolving. And this is why you’re seeing a lot of fast capital early. You’re seeing a lot of bigger funds put money to work early because it’s become a bit of a numbers game early in that capacity.

Mark MacLeod:

I was reading today about a new AI company from some prominent OpenAI alumni. I think it’s called Reflective, like raising, seeking to raise $5 billion. And they’re ostensibly a seed-stage company. So I don’t know, I’m old and crusty, but I don’t figure, I don’t quite understand how that works.

Amar Varma:

Well, look, here’s kind of maybe some framing for some of the folks listening. You know, there’s always this term Aqua Hire, right? And there’s been you can acquire a company, you can buy a product that fits in a business line, or you can buy a product line. Those are usually the depths of them. Or you could just merge equals, which is rare.

The Aqua Hire used to be, and still sometimes is, hey, come work here. You know, we’ll take, we’ll give your investors some stock in our promising place, and you’ll join it here as a proven prominent person. Then that ranges to sometimes Google was doing Aqua Hires and very nice sums and figures over time. And then the Aqua Hire is now just at a level that I’ve not seen. It’s like the ones that Nvidia is picking up, or these people are like OpenAI is picking up some of these things. These are not trivial numbers. They are large numbers because the value that they add as a puzzle piece to the ecosystem is so big.

And that’s why you were seeing some of those offers that Zuckerberg and Facebook, Meta, were throwing around earlier this summer. It looked like an NBA contract with fewer years or a baseball contract. It was like, oh my goodness. Like, there’s just that much of a run. There’s always going to be value in the scarcity of talent, and we’re seeing it at the next level.

Mark MacLeod:

Do you think all of this is a bubble or evidence of a bubble?

Amar Varma:

Underlying this is a fundamental shift. This is equivalent to an industrial revolution. Now, some people want to call it the fourth Industrial Revolution. My history is not perfect, so I’ll follow somebody else’s tag on this. But this is fundamentally where Ford Motor Co. comes in. Having bought one of my companies, which was one of the quintessential industrial revolution companies, and you’re seeing AI as the next jump in revolution.

So that’s going to happen. You will always see bubbles in these things because things will get overfunded. How many car companies were there and how many are there today?

Mark MacLeod:

Probably hundreds, right?

Amar Varma:

Hundreds to tens. That’s happening. So you’ll see it. But I don’t know how, as an entrepreneur, you’re not excited by this opportunity. Greed will always have some level of impact, and it might be one of the stronger impacts in a lot of these bubble deals. But as an entrepreneur and investor, you have to go in with conviction, and you have to be there and kind of to some degree be greedy about it and be greedy about the outcome for everything. And if you’re not, you risk not being in the wave moving you forward. You risk being on the undertook part of it.

Mark MacLeod:

I mean, Sam Altman himself thinks we may be in bubble territory, but at the same time, he’s enabling pretty big secondaries for his employees, cashing in.

Amar Varma:

You know, we’ve been around the block a few times, right? People always want to focus on one line that someone says instead of the arc and their actions. So I think you have to follow the actions and really do a fundamental thing on that. And if you said it, you said it, great. But having the secondaries at those valuations is enabling it.

Mark MacLeod:

Which goes back into this notion of OpenAI alumni. Like, part of what made the PayPal mafia work was they all had this experience, but then they had capital to invest in the next generation. And now you’re going to get a whole bunch of OpenAI employees or alumni with capital to start or invest or both.

Amar Varma:

There are two things in the journey in life that I’d look back on. One is when I was young, and maybe not even so young, I had a naive view of opportunity and outcomes. I was like, there’s no ceiling in this world. Like, I can just go do this thing. That signal is usually stronger, generally the younger you are, which is good. But you kind of have to maintain that through life to be in this world, in this business.

So you have to have that ability to think, I can go do it. And that fuels a lot of leaps of faith in what you’re doing. I’m just reading, like, an old friend’s book. Barry Diller put out his book recently, so it’s really exciting. For those who know Barry Diller, who’s like a Hollywood exec, and he kind of stumbled into doing movies on TV, and he just never feared failure. He didn’t fear failure.

And part of it, coupled in there, is the naivety of knowing something can work. And the other is not fearing the outcome. Like, you sell your stock at whatever valuation and you’re financially independent, your ability to fear a negative outcome is either reduced, limited, or eliminated. Like, you’re just like, I don’t fear the outcome of this, which is the most awesome place to be, because those folks do the most ambitious, generally kind of swing at things.

Mark MacLeod:

100%. All of my CEOs who’ve taken secondaries along the way or had a prior exit, and now they’re back in the ring, all of them are just like, de-risked. And to your point. Yeah, just have either no or little fear because their downside is taken care of already. They don’t need this to work. They want it to work, but they’re willing to swing to the fences.

Amar Varma:

The first win, right, is always what people are chasing. Some people are like, oh, I was a founder and then they become like an associate at a venture firm, and it’s like they haven’t had the win, right? And maybe they’ve thought they can go get the win as a venture investor, then come back to doing something later. Everybody’s got their own path.

But it’s like once you have that first win where people go, oh, you’re on the scoreboard. You never want to play somebody else’s marker game, but in your mind, you’re like, I have the win. I don’t need this output of the scoreboard to be relevant to me. I’m going to drive my own rules, my own engagement. And that’s when you just see people move with a different speed and purpose.

Mark MacLeod:

I love it. I’m curious, you’re an operator, but you’re also a very active investor. Does this environment make you want to invest faster or slower, or is there no change?

Amar Varma:

This is a long, a great question with a lot of depth, but I will try and synthesize some of the thoughts. I’ve actually spent a good chunk of time looking at Power Law over my career, and then some of the venture investments. And then on a risk-adjusted basis, I’ve started looking over the last decade or dozen years, what were the returns in private markets? Risk-adjusted vs owning the S&P. Owning names you used in your life every day, including tech.

Like, I’ve been asking folks who’ve started setting aside these education savings plans for their children, whose kids have just started to go to school. So they’re calling it 17, 18 years old, like, how big did that get? Right? And the answer varies based on what they invested in. Some people were in index, some people are like, I YOLO’d it. We’re like, I did Tesla. I met a friend last week for dinner at old Facebook. Right now, living in Costa Rica, I was in town and it was like he YOLO’d it. And now his child’s got high-class tax problems for a few years post-school.

These are high-class problems, right? So I really started looking at the risk-adjusted basis. And the answer is like, if everybody’s chasing the 10 companies with 41% of the dollars, you either need to have been in those and you need to keep doubling down in those or you need to be in your mind, smart enough that you’ve caught an early wave where you can still make the real power law return starting early in something that’s not necessarily a known category yet.

And that’s where infrastructure we know drives all sorts of applications and other investments over time. So there’s kind of a historical flywheel there. But will it happen again? Will we be right? Will it happen in a shorter time period with a compressed cycle? Will it take longer? Will the subsuming of apps change what this looks like? There are just all these things that start to come about.

Mark MacLeod:

Let’s talk about that subsuming of apps. That’s another observation you had recently. Tell us about that.

Amar Varma:

Yeah, look, for anybody who’s doing productivity work, whether it’s for personal or professional use, it’s like, I’ll give you an example. I had to pay some city taxes. I was like, it’d be great if that could just get done for me, right? And it’s like, well, you could automate it. You could put a calendar notice for yourself every six months, or however you want to do it, right? Tasks like that can just get taken off your plate. Organizing a vacation in a trip, right? Like, do I need to go into any of the travel apps now, or do I ask the model to curate an agenda for me?

You know, click here to buy a ticket. You know, there’s just a bunch of these things that are becoming co-pilots, concierge services, however you want to think about it. And it’s fascinating for me. So, do I go to vacations.com anymore, or am I looking somewhere else? Am I starting somewhere else? You know, okay, Gemini or ChatGPT, or Clyde, create the agenda for me. There’s this conference coming up later this year, and it’s like, hey, what dates can I go? What flights can I take? What do I need to bring? These are all exciting things that I’m like, huh, okay, when I’m gone, what kids activities do I need to have covered? Just all of these types of things as an individual and a family.

And then when you get to work, it gets amplified, right? All of the email marketing automation, all of the contract negotiation, a lot of the development, like anything that was kind of friction, a lot of the friction’s starting to come out using these tools. And sometimes the apps get eliminated. Yeah, that’s just part of the subsuming. I’m trying to think of a big one that’s just getting eliminated. Like, if you go through the stack, Gmail’s not going anywhere. Copilot, GitHub’s not going anywhere. Some messaging apps are not going anywhere yet. Email’s not going anywhere. But you know, do I need, do I need slides?

Mark MacLeod:

Blogging’s probably gone.

Amar Varma:

Yeah, yeah, blogging, slides. Do I need to have a disparate app for slides, or can I use a canvas that says, hey, create some slides for me? And it has subsumed the slide-creating capability, and it does it. OpenAI Canvas does a good job of this to allow you to kind of edit and view on both sides of the screen. So, this concept is coming, it’s there, we’re using it. But when I’m in this like every day, like a frog, we’re getting boiled. So I’m sometimes losing track of what’s happening. What are the jump functions and the jump steps?

Mark MacLeod:

I just got Perplexity’s Comet browser yesterday. So it’s just another example, right? Like, just behavior patterns moving. How much of your time, like your work time, do you think is spent engaging with an LLM versus inside an application standalone?

Amar Varma:

Taking a moment to answer this because it’s a bit of an involved answer for me. For quick task-oriented things, it’s probably a quick ask into the model. I spend a lot of time in our applications, and it’s a bit of a morphed one because we use a lot of automation in the applications. That’s probably not the right benchmark, but more than zero times a day, I’ve asked models for things in a work context. It’s kind of always…

Mark MacLeod:

Like, I run a one-person business and I have three employees, which are ChatGPT, Claude, and Perplexity. And I use Claude for creative work. I use Perplexity for research because its origins are search, and I use ChatGPT for everything else.

Amar Varma:

And what’s actually super interesting is that a lot of people will do the first layer of the work. And in terms of using these models, what’s super interesting now is that as the context windows are getting bigger, you can throw more at it, but as the memory capability is starting to come up, you can have the starting point for analysis, have a bunch of context and a bunch of memory on situations, and it’s super interesting.

We use this for initial views of any kind of contract negotiation. We’ve fed all of the historical negotiations in all of that context, and we now can take any contract we get and get a green, yellow, and red on any term. And we get that as a starting point for which is then go to Chad and team to get things finalized, but it’s like, it happens super fast.

Mark MacLeod:

All right, let’s switch gears. I’m hoping this will be a fun one. You were talking about James Bond villains and how they’re kind of across all these different industry categories and taking over in real life. Do you want to expand on that?

Amar Varma:

Oh, wow. Okay. This has always been fun for me. I’ve used this for those who’ve known me a while; this is kind of an overuse now. But if you’re listening for the first time, this is fun. So I always felt like the movie industry, generally across the globe, had a bit of a leading indicator of imagination of how the world could unfold.

And in this race for AI, if you look at a movie like Ready Player One, where it depicts the world as a place where you only want to be in the virtual space of it. You know, who are the people out there and what are the personas, right? And I think Elon’s one to pick out. He’s got energy, he’s got media, he’s got space, he’s got transportation. He’s got AI now; he’s got a lot. That’s not necessarily the Bond villain per se, but everybody will have a different opinion. But I look back and I go…

Mark MacLeod:

Sometimes he seems like a villain, I don’t know.

Amar Varma:

Yeah, everybody’s entitled to their opinion on that. And it’s like, if I look at media, right, we went through this era of misinformation. What’s the source of truth? You know, I think it was Tomorrow Never Dies, Elliot Carver, right? The chip revolution. This was one of my favorites. I remember going to the theater to watch this one. Semiconductors, if you remember Max Zorin vs. Sinjin Smyth in A View to a Kill. This is to destroy Silicon Valley and hoard the entire semiconductor chip inventory.

We see that happening today. Who’s hoarding these things? Who’s got contracts and access to these chips? Water sources are becoming incredibly valuable. They’re becoming increasingly more valuable as the AI Industrial Revolution is starting to come through. And who remembers Quantum of Solace? We had Dominic Green, right, that character. Energy Alternatives. This one’s a bit older. The man with the Golden Gun. This one’s harder. I had to look this one up. Francisco Scaramanga.

And these books aren’t new, these movies aren’t new. And the Spy who Loved Me, Transportation, we had Karl Stromberg. The list goes on and on and on. It’s really fun to do this. All this to bring back the whole context of though we see things as new, and maybe for the first time, somebody’s imagined it already. Somebody’s already looked up. That’s kind of like James Bond is a lens. I’m sure there’s the George Orwell lens. There’s a bunch of other lenses he could use on this.

Mark MacLeod:

Well, some people bring up Terminator.

Amar Varma:

Terminator, it’s a great one. Terminator, the headset, Total Recall, now with these Waymo Robo taxis. Like, this is fun stuff, right? I remember that for Bumtop, Anand did this 3D user interface. Anand Darkerwall, if anybody knows. Yeah, we did a 3D interface for touchscreens. And talking about it, he used that movie with Tom Cruise. What was it where they were using the hand interface?

Mark MacLeod:

Minority Report.

Amar Varma:

Minority Report. And he goes, it’s that, but for computers. And I was like, oh, that’s a great analogy. You know, it’s just always fun to kind of do that, just as a good gut check.

Mark MacLeod:

Yeah, that’s fascinating. So, not to put you on the spot, but where do you, if you have a position, are you in The Terminator is going to come true camp or it’s not going to come true camp?

Amar Varma:

So this physical AI element, and there’s a lot of talk about it. So let me, let me frame it for how I think about it. You know, AI in the virtual world, like using ChatGPT or any of these, like, the number of times we’ve mentioned that in this podcast is unbelievable. Like, anytime you’re using like an interface for a computer, where this is happening, that’s very fun. When you start using it, interacting and interfacing with it in the real world, it’s interesting.

So, like, think of the iRobot or again, another movie. But it was just kind of one of these things where it’s like, the models now at scale have open weights, right? And I can use them on Sunny’s hosting at Grok Sidibhadra. I can do that. I can maybe use Aliasaria’s stuff, put it on my laptop. There are just a bunch of ways to put these models together.

And exoskeletons are now at a point where they are incredible. I’m sure you’ve gone on YouTube at some point, or somebody sent you some kind of clip where you’re like, robots can do that, right? And it’s like navigating this incredible terrain holding a tray of glasses or something, right? You’re like, wow. And like, you start putting those two things together and start saying, hey, learn to do this, learn to do that. Something simple, like unloading a truck. Do something more complicated, drive the truck, right?

Like this, an area where people are doing incredibly fun things and impactful things. It’s not just fun, it’s super impactful. You know, North America, we’re probably behind Waymo in our depiction of seeing a robot in the wild in other societies. You see them kind of outside, right? In China, you see clips of it, and if you’ve been, you’ve seen it, and it’s like, okay, so it’s a bit of a human change management that we got to figure out here.

It’s coming, it’s coming. I don’t know if they’re going to get programmed in the wrong way and learn things and be nefarious. We all hope that that’s not the case. And we’ve got to put in some safeguards and understanding of how and why, as a community, we can advance together.

Mark MacLeod:

I tell you, I got mixed feelings about robots. I think I would only have a robot in the house if I could turn it off.

Amar Varma:

Okay, so let’s use one. You and I are out there slinging on the tennis court, and I have that ball machine that could learn how to make you a better player. Would you trust it?

Mark MacLeod:

Absolutely. Yeah, yeah, I’m all over that. But it’s more like a robot around the house. Like, I don’t know, making me a drink, opening the door, doing stuff, maybe cooking me dinner. Like, I want to be able to turn it off at night and not feel like it’s going to come in and strangle me while I’m asleep.

Amar Varma:

Yeah. So like Jetsons styles, right? Like you know, some. You know, this is an old cartoon. I’m baiting myself. But The Jetsons was a cartoon where they lived in the future. Robots were taking care of all of this.

Mark MacLeod:

For what it’s worth, my wife brings out the Jetsons pretty regularly because they would apparently, I don’t remember this detail, but the food was just a capsule and it had all the… She hates eating. She’s like, if I could just take a capsule and have all of my nutrients, I’d be great.

Amar Varma:

Yeah, yeah. People are trying that. Yeah. This is super interesting for me. If the robots could take off a bunch of the workload for people, that’d be nice. That’d be nice. Yeah. I’m thinking something super basic. Like, I was on the freeway and I saw people cutting grass on the freeway, and I was like, okay, like, if that was just a machine that did its thing and we knew it wouldn’t turn around and try and slice someone.

Mark MacLeod:

And even on the cars, like, I was driving to my parents’ last night, and it was mostly straight on a highway. I would have loved it if we could truly have autonomous driving here; I would have taken a nap. I was pretty tired. But got to keep your hands on the steering wheel, and you’ve got to be looking out.

Amar Varma:

There’s a whole human element of design and rules and stuff. Like, I think you could. You could cheat the camera by looking at your eyes before. It’s gotten better now, even if you’re wearing sunglasses. They’re learning. They’re learning. People used to put weighted bags on the steering wheel.

Mark MacLeod:

Yeah, yeah, we tried that. All right, I’ve got two more topics for you. One quick one, and then we’re going to bring it back to your company. So, you’re an active investor in companies and in funds. And a thing I was bringing up with you is like, hey, is VC pattern recognition out the window? Like, so one pattern is Jason Lemkin’s triple, triple, double, double. That’s a pattern of best-in-class expectations for SaaS companies, and other patterns might be when you look at the top-tier companies, the time between rounds compresses.

Maybe it was 18 months from the seed to the A, and then it’s like 12 to the B and 9 to the C. Because you just get this momentum, right? So there’s like certain patterns. And then of course, VCs looking at all these companies to identify the few to invest in use pattern recognition as their primary filter. And I wonder, with all of these changes, does that filter become less effective?

Amar Varma:

In the history of mankind, pattern recognition has proven to be valuable. Pattern we’re recognizing as being rewritten in some capacity. But I’m sure when we look back in a couple of years on the other side of this chasm, it’ll be like, oh, that was obvious. It was the same thing we did last cycle, right? It happens every time. But the observation of triple, triple, double, double, that’s just a filter now.

That’s not really an observation of like to me anymore, where it’s like, I could see that happen in that company that’s sitting at $1 million today, and it will triple, triple, double, double. And you’re sitting at $35 million in four years, right? Like maybe, right, but what’s the pattern? It’s like, okay, I’ve seen super cycles of investment. I know road maps of the incumbents are moving at light speed, but there’s a warp light speed of like the next roadmap items they need.

They think the roadmap is going to go there in two cycles. So you’re building for this next cycle of the roadmap, and you either connect with their roadmap and vision or you miss. Like, that’s an obvious pattern to me. And if you can hit that next roadmap item, you have a chance to catch the wave, right? That’s a pattern recognition thing I’ve just seen over time: always be successful. Easy to say, hard to do, but it’s just changing.

I feel like that network effect of being inside one of these closed networks that has the ability to influence and move markets. Like, there are companies out there that can adopt a solution and anoint a company overnight, right? Like, it’s always been the case, but it’s more prevalent now, just given the scale. Data center ops, GPU management, like there’s a whole ton of these things that were kind of quirky things that have just grown at scale.

You know, how do you quickly create physical space for a data center in the desert? Is it a tent, or is it a concrete wall? All of these things are getting rethought, rewired, and that’s a super cycle spending on a pattern. To your point, maybe the SaaS revenue one is an observation and not so much a pattern to watch right now.

Mark MacLeod:

Well, I’ll tell you, I’m not going to go raise a venture fund anytime soon.

Amar Varma:

Well, let’s put a minute on this, right? Like, you’ve been asked to join a number of firms since you’ve chosen to do other things, and likewise. And it’s like, if you were to do it today, let’s just frame it here. What would you do? What would you raise? What would your thesis be? And I’m just like, you could either be so small, so nimble, because you know the 10 people that matter, which is rare, or you have to be so large that people have a gravitational pull to you because of your scale.

And your referral and networking engine need to be on point, right? Like, that’s really hard to do. Two vintages back to back. I’ve seen very few people be able to be at the top for multiple vintages. So you can always come back to it during different timelines and stuff because you’d be like, hey, I was successful in this part of a super cycle. Or, would you raise $10 million? A hundred million? A billion? 10 billion? Would it be at about $100 billion? Like, it’s just, I…

Mark MacLeod:

Yeah, I’ll tell you. So last time I was a VC, it was a seed stage. And I stopped doing that for a number of reasons. One, I didn’t believe in the industry at the time; location mattered far more than it does today. And I didn’t feel that I could out-Sequoia Sequoia from Montreal with a tiny fund. But also at that stage, product mattered a lot. And I’m not a product person.

I now spend all of my time helping CEOs grow faster than their business, and have observed many who’ve achieved real scale. So if I were to pick a focus, for me it would be getting a fund that is focused on the growth stage. You’ve hit escape velocity, now you’re scrambling to keep up. And my entire focus is actually on the CEO and like, how do I help them grow, you know?

Amar Varma:

Love it. So, betting on the jockey, so to speak, right? Like, who’s taming the horse, who’s going to bring it around, those trackies, that sports analogy. Okay. In 2025, I would suggest that you either need to have fast, nimble capital at any scale. Plus, at scale, you see these firms have all of these different product offerings that help to generate alpha now.

Like it’s gone past just recruiting, and you know, it’s like, I have an investor, and they’re great in craft, and they have like, how do you do product-led growth? They have sales-led growth. They have all of these, not just playbooks or blogs; they have teams, they have people, they have programs, and they have unlimited cycles to spend with you.

So to the extent that you can and you should spend time with them, that’s part of delivering the alpha and attracting these deals now. And it’s become a very functionally broad business versus what it was. You kind of need to have that. And I love that you’re focusing on the CEOs at these ventures because that’s generally where your alpha is going to come, their ability to scale or not.

Mark MacLeod:

Get the CEO right, they’ll get the rest right. I don’t want to overly simplify, but look at all the biggest outcomes, they’re founder-led.

Amar Varma:

Yeah, I think that’s hard for people to digest. Maybe, not in your audience, but just let’s go through it, right? Like Bill Gates, to the extent Larry and Sergey were there, Salesforce, and Nvidia.

Mark MacLeod:

Zuckerberg, Benioff.

Amar Varma:

Meta, Tobi’s amazing. These are amazing examples of founder-led. And I learned this in Silicon Valley when I moved there just after finishing school. All of the companies were founder-led, and the companies were either public or startups. And it was all founder-led. And you saw a different attention to doing things from the founders that were running the place, because there was always an attention on getting it right, no matter what the detail was.

There was never this big company, like, the system will take care of it, we’ll be fine. Yeah, it was always like, we got to get this detail right, no matter how big or small the consequence was. And it was this attention to detail that just got people grinding to get it right.

Mark MacLeod:

Yep. Yeah, totally. All right, final topic, Amar, and I threw this one last minute, like when we were just riffing before we went on air. But you know, there are probably many folks watching or listening, building in a category that has an incumbent, and that’s your reality, kind of in your space, kind of in this whole kind of cap table management place. I’m pretty sure most, well, actually, I’m not gonna make an assumption you’re gonna say.

But anyway, when folks think about that category, they probably think about Carta. And so I’m just curious, as you were thinking about this, what did the presence of an incumbent motivate you, make you afraid? Like you were indifferent, you didn’t care, you just saw the opportunity, you were scratching your own itch? I don’t know, just turning the mic over to you, whatever thoughts come to mind.

Amar Varma:

So I think a lot of people building companies, there’s the $0 billion analogy, right? You’re either in a market where you’re going to, if you’re correct, you’re going to go from a zero billion dollar market to a much bigger market, and then there’s the shrinking multi-billion dollar market from 4 billion to 2 billion or something like that, where you’ve driven a bunch of efficiencies.

And then there are ones that can get both, right? They can nail both. Generally speaking, you’re always up against an incumbent, which is the existing way that something is working. The current mode of operation from these textbooks, right? How are people doing it today? And I think if you fear the incumbent enough that you’re like, I’m not going to do this, that’s a data point.

If I told you 25 years ago that this company that left for Nvidia would be, I think it’s like maybe 10 times, maybe 20 times more valuable than Intel, you would be like bro, there’s no way, right? Like you would just argue. We’d be like, no, come on, man, you lost all credibility. It would be like this crazy conversation.

So if you’re not willing to take on incumbents, if you’re not willing to see where the market is going, don’t play. Like that’s a simple thing. No crying in the playing field. I love that we have in our business, obviously, for cap table spreadsheets was the thing, and Carta created a tremendous business around it.

I got a lot of time and respect for that. But I feel like for that business we understand where that market’s going and we think there’s room for not just one, there’s room for more, and it’s not just to be second, we want to eventually win this marketplace.

As we look across the entire spectrum of alternative assets and private assets, which is where Mantle plays. Whether you’re a company, a founder, an employee, a general partner in a fund, or a limited partner in a fund. Like, we’ve launched a second product in limited partners now. Like, we think that the infrastructure here is in its infancy.

And with the power of automation, we’ve just been able to create a lot of it very quickly. And if you look at some of the tailwinds we have, what happened in America, there’s an executive order to allow private assets into 401ks, and 401ks are, I don’t know, $12 trillion? I don’t know the exact number. It’s probably a very large number.

And then BlackRock came out and said, well, instead of owning 60% stocks, 40% bonds, what about 55, 35, and 10? And it’s like. And the 10 is now a private alternative asset bucket. Okay, that’s a huge number of dollars that’s gonna start going into that. And then now you have to look at all those dollars coming in.

What are the products, what’s the infrastructure, and how does it get tracked? And then what’s the tooling for that? You can’t log into your bank account and see all your private assets. There’s no way to transact it or no way to understand what you have. So this is not about anything other than just being able to get your handle around what’s there as a starting point, creating infrastructure for transparency, and then ultimately, what happens to it over time is the market will dictate.

Like, this is why Bloomberg works so well. There’s that transparency underlying the public market driven by the Q SIP code, which is a technical term. And it allows everybody to track each and every single asset and allows them to have communities and applications and all these fun things.

Mark MacLeod:

Fascinating. I love it. Well, I definitely think there’s room for you to come in and win this market. And yeah, excited to see you do it.

Amar Varma:

Well, thanks, I appreciate it as always. I love watching your episodes, and I love coming. So it’s always a pleasure to be on here. So thank you, and keep doing what you’re doing, and I love it.

Mark MacLeod:

Thanks, Amar. Always learn from you. So great that others can learn from you as well. We’ll see it. We’ll have you back on soon. Take care.

Hey, thanks for listening to the Startup CEO Show. If you’d like to connect with me, be sure to visit my website at markmacLeod.me, or follow me on LinkedIn at The Mark MacLeod, or X account @markmacleod_, and if you want to tune in again next week, be sure to subscribe on YouTube, Spotify, Apple, or wherever you get your podcasts. We’ll see you next time!