Exclusive: Why Mega Platforms Will Win in VC

Thanks for having me, Harry. As they often say, long time listener, first time caller, and it’s been really fun to watch your meteoric rise over the years. As you know, I’ve been watching since the beginning.

Yeah, thanks. So the big news is today I am officially announcing that I’ve joined Lightspeed Venture Partners as a partner. After seven and a half years at KP and almost eleven years in the venture business, this new chapter in my career is really exciting in the sense that I feel like I really get to step into a truly global platform and really play a role in kind of driving what was once the core of the firm forward, which is this early stage enterprise investing that really set the tone for Lightspeed’s success and what it was able to become today.

So I feel really, really grateful to partner with that team. Yeah, that’s the next chapter of my career. I think it

So I think we’re in this really unique point in time, both in terms of the technology entrepreneurship industry and consequently the venture industry. If you look back five or seven years ago, we would be talking about companies like Databricks or Snowflake or even DoorDash on the consumer side as sort of the hallmarks of venture success. I think what we’re seeing now is that there’s this new shape of company that’s emerging that really, in some sense, puts those to shame. We’re talking about going from talking about 30,000,000,000 to $50,000,000,000 outcomes to multi trillion dollar outcomes in the form of SpaceX or OpenAI or Anthropic, where Lightspeed is very proudly an investor.

I think what that means is you also sort of have to take a step back and say, Okay, so where is the alpha going to come from in the industry if that’s the case? I think what I’m seeing and sort of have been thinking about for a long time is that these platforms that have the scale of some of those that you mentioned, Lightspeed included, they get to have really interesting conversations about like, hey, what if we invest a billion dollars in one of these companies that could be worth trillions of dollars?

What if we were to succeed in terms of turning $1,000,000,000 into $1,020,000,000,000 dollars? I think that’s just not a conversation that’s ever been had in the venture industry and frankly any other asset class that I’m that I’m familiar with. And I think for that reason, it’s really clear to me that having the chip stack that these large platforms have is is really uniquely beneficial at this point in time. But what I will say is that none of that matters if you don’t maintain a lot of dedication to early stage investing.

If you’re Sam Altman or you’re Dario and you’re asking yourself, Do I want to hitch my wagon to one of these firms by raising billions of dollars from one of them? You know, you’re thinking a lot about what their brand stands for. And I think that these companies, in spite of running towards these trillion dollar outcomes, are still very much startups and entrepreneurial in the way they think about their DNA. And therefore, want to partner with someone who shares that DNA.

And I think the only way for a venture firm to sustain that is by staying true to early stage venture, working with raw formation stage companies, and spending lots and lots of time kind of trying to figure out who that next wave of entrepreneurs are. And so I think it’s really important that Lightspeed continues that focus, I think that’s something they and a few of their peers do a really good job of.

So I think the story there is still very much being written, is the first thing I would say. But I think at this point in time, to your point, these companies have margin compression, they’re high burn because they have very high CapEx in terms of their need to scale their products by purchasing and running lots of compute. But on the other hand, you’re seeing this also, this very, very promising trend play out in the form of these consumer apps and even enterprise apps increasingly that they’re starting to lean into.

And so I think if we look ahead three to five years, the margin profile, growth rates, and obviously revenue scale are going to look a lot different for these companies. Right? So at this point in time, I think it’s fair to say that you’ve taken a bit more dilution than you would if you were investing in some kind of a late stage enterprise software company. But on the other hand, I think the scale of the outcomes that are possible with these companies, and I think the types of consumer and business facing products they can build atop these models, are really just starting.

I think if you look at just how unique these companies are on a top line basis, we’re talking about billions of dollars of run rate growing in excess of 100% year over year. I don’t think we’ve ever seen that before, right? You talk about ServiceNow doing billions of dollars in ARR and growing 2030% year over year, and you’re like, is an amazing company. These companies are growing over 100% year over year doing that kind of run rate. So again, I think the margin question is a valid one.

I think the long term cash burn question is a very valid one. But what I’m seeing is I think these companies are just getting started in terms of optimizing the shape of their business by layering these products on top of this very capital intensive asset that you’re referring to in the form of these models.

model for thinking about this is actually quite similar to when we were asking ourselves when Amazon, Microsoft Azure and Google were rising, like which companies would survive that, right, as they were starting to get into multi product. And I think what we saw in that era was that there are just so many categories of software that exist that require unique understanding of the customer pain points, unique resource allocation to build the best product, and really just like a set of insights that can only come from focus.

Then I think what we’re going to see is a very similar set of dynamics play out here. So sure, in a core market like cogen or maybe enterprise search where you see companies like Glean being independent today, OpenAI famously saying, Hey, if you’re in Glean, you don’t get to be an investor in OpenAI, which would foreshadow they plan to play in that space. I think what you’re seeing is that there are some categories where these model providers are going to be very, very active.

But I think what’s so amazing about this point in time that we’re in is that this long tail of applications is so long and so fat that it’s hard for me to see how the model providers are able to play in each of them. And so again, I think what’s going to happen is that there will be startups and in some cases, incumbent companies that really figure out where those areas are, deep customer insight and deep focus is required. And I just have a hard time seeing how the model providers are going be able to prosecute each of those.

And so I’d certainly be a little nervous about how the cogen stuff is going play out today, because it seems like that is just far and away the killer use case for LLM. And therefore, one that these big model providers are going to lean into next from an application standpoint. Of course, the consumer side of things is very tricky as well, given that’s sort of the hallmark of all these companies in terms of what they put in front of consumers. But again, I really feel like that dynamic is going to kind of repeat itself.

And there’s going to be so many opportunities that exist outside of that core that you’re going to see big, big companies get built. And those companies are either going to have to be acquired by the model providers or they’re really going to struggle to compete with them because they monopolize all the customer insights that come from being first and going really deep and solving customer problems.

every firm takes a very different view on this. And I can tell you that when we were at Kleiner Perkins, was something that we it was kind of a red line that we wouldn’t cross. And the reason for that is because we invested in so few companies and we go so deep with each of those companies that in our mind, it was just too hard to feel like we could really provide that level of service to each company in a way that would make conflicts a non issue.

I think it’s a little harder for me to speak about kind of how we do things at Lightspeed in that sense, but I think to your point, you’re seeing a lot of these later stage investments get made into sort of multiple players. And I think part of that is like, there’s just so much demand for capital from these big firms that these model providers want. And I think on the other end, it’s really hard to say how this is gonna play out to the point that I think there’s a need for diversification as well on the investor side.

And one argument would be, hey, pick one of these companies, go all in on them. And you can see that’s sort of what Thrive is doing with OpenAI. You could also say the same about Founders Fund. Another would be, who knows how it’s going to play out? And I think what you’re seeing with, say, Andreessen Hsurf, is they’re trying to invest in every one of these companies. And I think it’s a rational strategy so long as you have buy in from the entrepreneurs so as to not ruin that relationship.

an LP, but like you, I do spend a lot of time with LPs. And what I can tell you right now is that it’s harder and harder for me to see how there are too many of these mega platforms that exist in the coming years. And the reason for that is one of commitment from the existing large pools of capital in the LP community. Right? So if you’re an LP that wants to fund one of these efforts, you’re really making kind of a multi fund commitment, as you know.

The sophisticated LPs invest with any manager for at least three to four funds, and they know that is the optimal way for them to capture that alpha because it’s really hard to be timing the market, as we all know. So So I think if you sort of say, okay, in the backdrop, it feels like a lot of these big LPs have committed to these platforms. It’s pretty hard to see how they can commit to too many more of them just on the basis of the physics of how big they are and how much capital they’ve committed.

So I think what’s really, really interesting about this space that we’re in in the market right now is, like you said, there are these companies that can get really, really big and consume large amounts of capital and generate potentially outsized returns for investors of those large checks. But on the other hand, it doesn’t appear as though the supply of capital from platforms like the one that I’ve just joined or some of the others that we’ve talked about is going to go up in size. And so again, you could argue that these five to six firms kind of have this captive opportunity to really play a role in helping these companies scale and be their partner.

And I think that’s a really exciting dynamic for these mega platforms.

Look, I think that’s a tough one because it really is somewhat indexed to this question of how these megatrends like space and robotics and AGI really play out. If we really are talking about companies that are going to get to some level of AGI such that they’re going to be not just multi trillions, but potentially larger than that, if we’re talking about how Starlink and the kind of assets being built around that could kind of scale to those heights, these companies may not have to go public for a really long time because there will be so much demand to continue supporting them from the private markets, which as you said, are very, very robust right now.

So the honest to God answer is I don’t know, but I think we’re going to find out as an industry over the next few years. It’s really going to come down to how some of these truly blue chip assets that are that are scaling to those levels, just how how they trend and and and how the market kind of shapes around them.

So I think if

investors right now and you asked them how many of these really, really valuable companies are going to exist in AI, in defense, in robotics, in space, most of them will tell you there’s going to be maybe a few of these companies that get to that really, really massive, unprecedented level of scale. On the other hand, I think we’re still really early in the super cycle of each of these areas. I mean, I think there’s a lot of talk about how it’s going to be Helsink and Andriel.

There’s a lot of talk around how it’s going to be OpenAI and Anthropic, maybe xAI. I just don’t think we know. And when I look at a trend like robotics that is still so, so early, but at least in my mind, I’ve convinced myself that this time is truly different in terms of the industrial applicability of these products, It really, to me, looks like we could be surprised to the upside. But again, this is sort of our job as a venture investor is to be very, very optimistic.

So I think that is the number one constraint to this whole experiment playing out favorably, will be like, are there more than one or two of these per mega category?

Mamoon, has taught me many things. One quote that sticks with me on this one is, The best companies always feel expensive. And I think that’s proven true in every example I can think of. I think it’s really about determining, one, is this one of those truly special companies? And obviously, the earlier you have to make that decision, the much harder it is and the error rate’s gonna be higher. And I think therein lies probably some reason to be a little bit more constrained on your thinking around price.

On the other hand, and this kind of dovetails into a question of, Well, what does it take to win a competitive round these days? I think there’s always someone who is going to believe more than the field. And so if that is important to the founder, then you sort of have a hard choice to make. Whereas if you think that Series A, this is a very special asset and they’re pushing price beyond kind of the way your mind can rationalize it, history would say that those companies always feel expensive.

Those companies can run and compound for a very long time. And I think especially today when we’re in a world where, let’s say if you talk about these agent products that are quite literally going after existing software spend to starting to replace human labor, it kind of comes back to this optimism I have around the size of these outcomes. I really believe that these software companies that are not just augmenting, but in some cases replacing the work that humans do and budget allocated towards paying them for that work, these companies are going get bigger, right?

So I think the way I think about things today is like when you find one of those companies that has demonstrated some path to market leadership relative to the field dynamic. My view is you really have to build conviction

Yeah, mean, the honest answer is it depends, right? Like, in a case where I would say it does matter is if you’re building Let’s say you’re building a cybersecurity company where the play is you have a better solution to something that already exists. You know exactly how much is being spent on that thing, and therefore your opportunity is to go and capture as much of that existing spend on that thing and then grow with that market. There, I think market sizing really matters in terms of just being sober about the size of the opportunity and honest with yourself about the right way to build that company and capitalize it.

I think there’s this other case though, where doing something fundamentally new. And if you’re doing something fundamentally new, the act of sizing a market is just, at least from my experience, so imprecise that it borders on being a fool’s errand, number one. Number two, I think you’ve heard many people say this, but I very much agree with it. They’re like the best founders, they’re just so creative and have so much ingenuity in terms of their ability to essentially set the rules. Right? And what I mean by setting their own rules is they get to decide what market they’re playing in.

And they get to convince customers that there’s this market that they didn’t really think about before that they should be. And I think in that sense, have this way of kind of dictating that and therefore, like trying to size it is just really, really hard. And some of that is like you build new products. Some of that is you reframe an existing view of how to solve a problem into something that kind of comprises a new market. And then there are these compound startups, right? And this Rippling is, of course, the company that everybody loves to talk about in this sense.

But I think what you see with Rippling is like, sure, you were kind of starting and thinking about it as a payroll product, but they have just been so unbelievably effective at layering on new products that have synergies to the existing core that the market size just keeps getting bigger every quarter. And so I don’t spend a lot of time sizing markets, but I do think those companies that are going after an existing market, it’s- it would be- it would be crazy not to at least think about the size of that.

The window that spreadsheet investors have to gain access to these companies is just a lot narrower in the sense that these really, really blue chip companies that are wielding AI in some interesting way, for example, let’s say, again, they’re building some kind of an agent that is going after a large existing pool of labor and automating that, they’re just raising so much money early before there is a spreadsheet that one’s ability as a spreadsheet investor to get exposure to those companies at the kind of entry price and stage that they’re used to, I think is changing a lot.

And so my view is that these spreadsheet investors are just gonna keep getting pushed later and later stage. And again, think the book is still being written on whether that’s a good thing for the industry. But what I would say is that there’s just so much conviction amongst top investors at the early stages for companies like that, that

So I can tell you that I have been involved in multiple situations where companies look back on their fundraising approach and say, Hey, we raised too much money at too high a price. Now the flexibility that’s afforded to us, and not just downside scenarios, but in some cases, scenarios, is more limited than we’d like it to be. So I’m completely wide eyed about the trade off there. And anytime a founder I work with is thinking about going this path, I try to have that honest conversation with him and say, Hey, if I were you, this is how I think about it.

Of course, there’s a way that you can raise more money at less dilution. And if that’s all you’re thinking about, then that path may seem like the right path. But on the other hand, there are so many scenarios in which preserving optionality makes a huge, huge difference. Therefore, keeping dollars in and the value of your $40.09 or your post money of your last round as low as possible is very, very beneficial. So to me, it’s like there are certain companies and certain types of founders that I think are better off and very comfortable going the kind of how do I raise as much money as efficient a dilution type of price as possible.

And then there are other founders, but I think you really, really have to help them understand the downsides of that because they’ve never done this before. But ultimately, it’s the founder’s choice. And I think in terms of in terms of how, like, late stage investors may or may not enable that, I personally would say, like, I take a view on that as well. And, like, when I believe the company’s up for it, when I believe the company has enough pull and enough potential to be a little bit more aggressive with fundraising, I’m okay leaning into it.

But I don’t feel that way about every company I work with, and I’m honest with the founders about that.

think this actually kind of comes back to the market sizing question you asked, which is there are certain companies that have a very deterministic sense of their market opportunity. Let’s say they’re going and replacing something that already exists. You see this a lot in cybersecurity, for example. I think these kinds of companies, they understand the headroom and they understand if I get this much market share in this period of time, this is how big my business can be and therefore how valuable it can be with a modest degree of confidence.

Where I get really conservative in terms of the advice that I give founders that I work with on fundraising is when companies are fleshing out a new market. Like, we just don’t know. It could be really large. It could be nonexistent. I can tell you that the cases where I’ve been involved with companies where they’ve done that wrong is when they just didn’t understand their market yet. Right? And I think if you have a poor understanding of your market and there’s a non zero chance that that market could be very constrained, optionality is without a doubt your friend.

And I’m very honest about that with the founders I work with.

I definitely did. Sure, in the case where you feel like you’re drawing dead, the ability to find a good home for the company and do good by your investors and your employees is, let’s just say, much easier to come by when you keep valuation down. And we all know how that works with corporate acquirers. I was once on a corp dev team, so I know that world very well.

But I think there is another form of optionality where if you really think that the market opportunity or the range of outcomes that you’re scaling into is so vast, keeping it lean also allows you to say, Hey, if this doesn’t work, do I really have to spend the next four or five years of my life working on this thing that I’m not sure of? Right? Like that time is so precious for great founders. And I’ve been in situations with founders before where it feels like they are kind of drawing dead and the market’s not resonating the way that they thought.

Sometimes I think having that extra, you know, two or three years of runway can actually be really, really punitive given the opportunity cost associated with amazing founders’ time. So I really try on a personal level to be honest with them about, look, is this really what you want to be spending the next two or three years of your life doing? And I think if you go and raise too much money, sometimes you can kind of be stuck with that. And there are countless companies in the industry right now where you have really talented founders working on something that’s kind of working, but maybe not working to the degree that they hoped.

In some sense, they’re kind of stuck with it. Right? And I really empathize with that. And I’m protective of the founders that I work with because I care about them and I value their time because I think their time is valuable.

Look, this is a hard one. I mean, there’s examples like Figma where, you know, they toiled in obscurity for some time and became the amazing Did you buy that, though? Because I

But I think you have a lot of very special founders whose companies don’t work, number one. And so you can say that about a number of founders and a number of companies. But I think you’re starting to see more and more evidence of these companies that just took longer than people thought. Right? I mean, there are just so many now. It’s not just Figma that I think you kind of have to stay open minded to that. And there’s another dynamic at play here, which is that I think that so much of the low hanging fruit has been picked off the tree in terms of software businesses that you can build.

You’re starting to see the most interesting companies be those where there’s just something deeply technical, like a problem they have to go and solve that’s never been solved before. And I think in doing so, that can take, in some cases, multiple years to get it right. A recent example of this is Kleiner. Kleiner is this company that’s growing very, very fast on the sales tech side. And if I’m not mistaken, it was like five or six years of very little to no growth before it took off.

And so I think you have to be open minded to those outcomes. And I personally try to lean into those outcomes because I think sometimes if something takes many years to get right, that can also be something that’s very defensible and hard to replicate. And so I think the game has changed a little bit on this front. And I think that these more deeply technical software products just take more time. And you do have to be more open minded to these timelines being a little bit different than the canonical that either works in the first twelve months or not.

Look, so I think core to any VC’s job is the ability to prioritize and ruthlessly reprioritize. And part of that is just like, what are the best ideas that I have in front of me at any given point in time? I’m going spend time on those. And right now, to your point, the bar is going up for what best idea looks like when it comes to growth rates and momentum. But I really think you’re right to say that what great looks like has really changed. I think the book is still being written on some of those companies that you mentioned and their peers in terms of just how durable that revenue is.

Again, I’m not going to repeat myself and go down that rabbit hole. But I think what you’re seeing is that the pent up demand for intelligence in all of these different areas of like work is just so unbelievably high that the growth rates in the market pull around these companies are like nothing we’ve ever seen. And I can tell you now as an investor, when I’m out there looking for new ideas, I am often looking for the signals that I saw in companies like Glean or companies like Windsurf that have just had this incredible market pull and reception.

And I can tell you that it’s almost made me at least I feel like I’ve become a more discerning investor in the sense that there are signals that I saw for those companies that were sort of foreshadowing of what was to come that I can now look for in other companies.

So first, I have to give kudos to my former partners, Mamoon and Lee Marie, for being the leads and involved in those two companies. But what I would say is those two companies in particular to me found this zeitgeist at the senior most decision maker level. If you go and talk to a CIO or a CTO of a large enterprise today, they’re furiously seeking out ways to apply business. And the reason for that is because their CEO and their board told them, You’re going get fired because the entire fate of our company depends on leaning into AI if you don’t help us do this.

So there’s this veracity and this appetite to bring new solutions that sort of comprise what it means to bring AI into business into these companies like I’ve never seen before. And I think Glean found that very early on, and the rest is history. It’s just been on an incredible trajectory and I think scaling into a true household name that embodies what the AI app business of tomorrow looks like. I think similarly with Windsurf, they did a really, really good job of going to these large enterprise technology leaders and helping them understand where the future was going and the role that they could play in shepherding them that way.

And I think you’re going to see more and more of that where once there’s like a magic moment that a CIO sees, there’s a bit of this mimesis that picks up where every CIO is going to hear about the fact that they adopted this technology successfully, and then they’re going want to go and do it themselves. And so I personally am looking for companies like that, that I think really can kind of be something that the CIO walks into the boardroom and says, Hey, like, this is that thing that you asked me to do.

I’ve adopted it. It’s working. Employees are happy. We’re getting all these productivity gains. And I think we’re just scratching at the surface of where that’s going show up. Like you’re to see it in finance. You’re certainly going to see more of it in software engineering, you’re going to see it in cybersecurity. And again, I think like Glean and Windsurf are just good cursory examples of like that trend that’s going to continue playing out and why I’m so bullish on these enterprise apps.

Yeah, so let’s unpack this by first talking about like what it takes to win in one of these really competitive opportunities. Like, I feel like there’s this sort of meme going around that it’s all about like, you know, famous VC picking you up in his helicopter and flying you to his house on an island or the courtside seats at the Warriors or Knicks game or the, you know, the Michelin Star meal with a famous person joining. Look, these parlor tricks exist. Every firm engages in them.

Depending on founders, they either respond positively to it or not. But ultimately, what I’ve found in practice is that none of that stuff really matters. What matters is that you put the work in to develop a deep connection and a deep set of insights with that founder, their vision, and the company that they’re trying to build. And that just takes time. And I think it takes time in the sense that every process I’ve been in that has been highly competitive, the winner, if it wasn’t me, was the person that had been doing that work for the better course of a year.

And I’ll give you a quote a founder told me once when an opportunity didn’t go my way. He said, The person that I went with, they just had insights about the business that not even my existing investors had. And the only way to develop those insights is from spending time thinking about the business at the level the founders do more than just about anyone but the founders. And so how this comes back to picking is you have to pick which companies and which people to do that with.

And if you don’t and you rock up to one of these processes that is very competitive, there’s always gonna be someone who is at a great firm with a great brand who has done that work, and that person will win. I see it every day. So is the future of venture domain specialization? So I think this is like a very good segue into like theoretically why domain expertise and specializing in domains matters, which is like, if the game really becomes, how do I pick which founders to spend time with and how, when I meet a new company, can I readjust my prioritization of spending time developing insights and rapport with that founder and that company?

Domain specialization really, really helps there. Being domain focused really, really helps with that picking aspect of picking who to spend time with, who to position yourself with for when they do decide to raise. And then ultimately, they do decide to raise, being able to reinforce that like, Hey, did I pick the right person to spend time with or not? Or am I just running away with this because I’ve spent so much time on it? Which is a whole nother bias you have to manage.

strategy can definitely work. But I would argue, and this is maybe world’s smallest violin, that picking is actually much, much harder in these larger firms that have notoriety in the market. The reason for that is because the opportunity set that they have access to and the number of founders that are willing to lean in and work with them is just higher. And so you just have more inventory to choose from. And I think because you have a little bit of magnetism to you as a platform that a lesser known firm doesn’t, you just end up having a lot more at the top of funnel to sift through.

And I think if you’re not very diligent about how you prioritize and manage that, picking actually can be like the failure mode of a lot of GPs at these funds.

I think ultimately it’s really hard when there’s so many founders that are coming in your door, which many firms like Lightspeed or Kleiner Perkins have the luxury of experiencing. It’s very, very hard to figure out what is the nine out of ten one and the 10 out of ten one. And I think you can talk about that in the context of who the great founders are. You can talk about that in the context of momentum. But my point is it’s a very humbling job for that reason.

And I think to say that no one has to pick that’s at an existing firm because the great companies are the great companies, that just hasn’t been my experience.

I would say the single biggest thing that I’ve changed my mind on as an investor in the eleven or so years that I’ve been doing this comes down to what I did before investing, which is I was, as I said, I was a member of Cisco’s corporate development team. And at a place like Cisco, when you’re a member of a corporate development team, your job is to go and look at markets and technology and figure out like what are the lucrative markets, what are the dynamic markets, and then what is the right technology that the company should have to go and prosecute that market.

Notice I did not mention founder quality or quality of execution in that statement. So I think when you come from a role like that and you move into venture, it’s very tempting to kind of apply that same lens to looking for great companies. And again, I can’t say anything about this that hasn’t already been said, but it’s very, very obvious to anyone who’s been doing this for a long time. Like you just get this visceral feel for it’s all about the founders. And so I would say that like the thing that I’ve really changed my mind on is I used to kind of go around developing theses and trying to figure out like where the world was going on the level of markets and technology.

And now what I’ve realized is, it’s not to say that work doesn’t matter. It’s actually very important and very useful. But really all it does is it gives you a flashlight to say, Okay, these are the areas where I want to go figure out if there’s amazing founders or not. And I think the hardest thing to do when you’re a thesis driven investor is to go do that work and have the humility and the restraint to conclude that there aren’t great founders in that space. And that’s okay because I’m going to go move on to the next one because what matters is I find those people.

So I would say that’s the thing that I’ve changed my mind on. Then to answer your question, the bad investments that I’ve made, they’ve been because I’ve downweighted the execution capacity of the team and the quality of the team that the founders can build around them at the expense of my intrigue for

Honestly, Harry, I think I’m just too much of an optimist at heart to say that there’s too much capital flowing in. Seriously? I think it’s easy to say that on the bad days, but ultimately, there is just so much amazing work to go and do. And really, we’re bound, I think, more by I think we’re bound really by like, can we help enough of these founders that come from unlikely backgrounds go and pursue their visions? And so I personally have a view that I think the industry is in a healthy place, and it’s just hard for me to say there’s too much capital chasing too few opportunities.

I feel like I’m seeing the future every day, candidly, and I just can’t spend time on every single one of these.

So I would certainly agree with this. I think it’s the single biggest challenge of making C and Series A investments right now. Let’s take these AI app categories that we’re meeting every day, for example. It used to be the case, as you said, there might be one or two players that you’d to pick between. And now it’s probably more like four to six. And so I can tell you a recent example in an anonymized way. One of the most recent investments I made was in an AI app company.

And I can tell you that we had been spending time with this team really since they’d raised a very small seed ground. And we had been really excited about the category, building a lot of altitude on it, but everyone’s product was under development. It was really hard for us to determine who had the best product. So okay, that’s fine. Then figure out who the best founder is, right? And there was a set of founders that had deep domain expertise in this area but very light on the AI side of things.

Then there was a set of founders that had very deep AI expertise and very light on the domain expertise. And I can tell you that we waited to pick the company that we pursued and we succeeded at pursuing an investment in the one that we wanted. But I can tell you the price was a lot higher than it would have been had we leaned in earlier, right? And I think what that came down to is we wanted to see whose products were good, whose weren’t. And I can tell you with hindsight, there’s this instinct that I’ve kind of honed in on, especially with these AI app investments, which is you think it’s the domain expertise that matters.

It is without a doubt the AI expertise that is more scarce and therefore matters more. And so I can tell you right now that there are these companies out there that have deep domain experts building application layer AI companies, but they’re really struggling to recruit the AI engineering talent that you need. And given how fast this technology is moving, you have to have people in the company that have these innate instincts for how to wield it and how to understand where it’s going and therefore build for that future versus build for today.

And I think if you don’t have people on your team that are truly AI native and truly AI engineers, you’re just not going to nail that the way the team that does will. And a good example of this that’s public would be Harvey. Harvey actually has the perfect union of a former lawyer and someone who has deep, deep AI engineering expertise in the form of their CTO. And they’ve been able to recruit an incredible engineering team of AI practitioners because that leader is someone who has that expertise.

The other form factor of this that I don’t think works as well is when you have CTOs that are, say, domain experts, like they built something in the category they’re working in, but they don’t have that AI engineering expertise. They’re just not able to get the talent in the building that I think you need to be a market leader in one of these big categories.

comes back to my view that I shared with you about how if you’re going to build one of these mega platforms and sustain a compelling position in the market with one, you really do have to stay dedicated to like the craft of like helping people build things from scratch. And to do that, you have to be in the pre seed and seed business. And I think the moment you stop doing that is the moment you lose the instincts for just like how fast these companies can change and how quickly the story can improve, such that you kind of sit back and are looking for perfection instead of really seeing what these companies can be.

And like, I actually think that these AI app companies we keep talking about is a great example of like, if you looked at Harvey at the Series A and you saw its product, you probably weren’t that impressed, right? It was just a very early product. The scaffolding of it was clear. What was possible with better models was clear. But it was still a very early and raw product.

And I think one of the mistakes that you could say a lot of investors have made in this wave of AI investing as it relates to these app companies is they failed to imagine what they can be as the models get better, as their understanding of the pain point gets better, as they spend more time with customers. And so I think where this comes back to this question is if you’re not spending time with early stage companies and you have that sense of just how fast things can change when companies are moving fast, start you to lose sight of that and you make mistakes where you start to take a fixed view of of these companies rather than a a more fluid and dynamic one.

And so I agree strongly with your statement, and I think it’s especially important for these really, really large funds to to to be playing there actively.

Look, would say it’s possible if you make a very, very small investment in a company at the very early stages to use that as a wedge to build a compelling relationship with an entrepreneur. But it is by no means a given. I can tell you countless examples of where, let’s just say, later stage firms or even just firms that weren’t necessarily the lead investor in the early days did something, let’s just call it low conviction, thinking they were going to get access, where it just didn’t serve them.

So I think it really comes down to this point about picking once again, which is like, hey, if you’re going to start doing that as a later stage firm, you’ve got to be committed to putting the legwork in to actually use that as a wedge to develop that relationship that does give you the access. Because I think the check itself does not. And so where this comes back to picking once again is like, let’s say you do that a 100 times, which of those 100 are you going to put that work in with, right?

And so you have to have taste, judgment, and sort of an instinct for, as these companies are developing, like which of those do I need to really spend time with? And I think that goes for any investor at any stage, which is again

what I’d say is that when a larger, let’s call it multi stage firm, does lead a seed round and decides not to lead the next round, that is certainly something that the next round of investors are going to want to understand better and think about and will sort of take as a signal. There’s no doubt about it. I’m not going to deny I will also say that any seed investor competing with a multistage firm will go very, very out of their way to instill this wild fear in the founder’s mind about the risk of that.

And I can tell you that what I’ve seen in practice is that I don’t think it’s as much of a risk. And perhaps I’m biased. I mean, I have worked at a smaller firm before and been on that side of the fence. But what I would say is that when a firm doesn’t want to go and do that Series A, it’s really just the fact that the company hasn’t necessarily achieved the milestones that they agreed upon with the entrepreneur in the early days for the company to be Series A ready.

And so sometimes that’s a downside scenario where let’s just say things didn’t go as well as you’d hoped. Maybe they took longer and you just have to go and raise because you have a runway issue. But usually what I find is that it’s just the quality of the company at that stage how they’ve executed against the milestones that they set forth that determines the Series A success rather than, say, the whimsical taste of the multistage firm that led the seed. So what I found is there are absolutely cases where a multistage firm will say, Hey, we feel good about our ownership.

We feel good about the amount of money we have in the company. Like, we want to help you raise a Series A from someone else and we’ll do pro rata. And if that company is a good company and it’s hit its milestones, that round will get done. And it does all the time. And, you know, often it’s another multi stage firm that comes in and does that. But I can tell you that like there is this meme sort of where a company doesn’t perform to expectations, and then the fact that the Series A comes together seems to kind of be blamed on the fact that the

I think the

somewhere in the middle, which is that, of course, multi stage firms, by virtue of being multi stage, they want to be first. That’s where the generational returns get made. And we’ve seen that time and time again, if you look at some of the best venture investments in history. They’ve been these seed investments in these companies that have gone on to grow really, really large. They’ve come from an investor who’s continued to concentrate more and more capital in that company over time so that they have a large amount of ownership at the end of the day.

That is where I think the competition lies, that’s obvious. But I also think that the great C firms, one, they’re heavily reliant on collaboration with the multi stage firms to be kind of sources of capital for downstream rounds. And conversely, I think that every good multi stage firm is humble about the fact that there will be companies that were either non obvious to them or not visible to them that these seed firms will find and will be good partners to. And so I think for a multistage firm to alienate those seed funds and say, hey, we just don’t wanna work together at all, just be crazy.

And I can tell you that none of the multistage firms that I know well do anything but seek to be very collaborative and follow what the seed

So I think for me, the thing I’ve changed my mind on the most in the last twelve months, specific to AI, is how to think about the net revenue mix of companies like OpenAI Anthropic. And so, look, I remember when investors were considering participating in a $30,000,000,000 valuation round for OpenAI. The question in the room was like, should we even value this ChatGPT thing as anything? Right? Like, is worth anything? Or is it just a proof of concept to show what the model is capable of?

And conversely, I people were looking at these API businesses where they make the model available to developers and saying, Hey, this is going to be like Stripe or Twilio or, you know, the next great API driven business. And I think what’s happened in practice I didn’t anticipate, which is that it turns out that these API driven businesses are really, really tricky. One, you have this kind of 100x year over year decrease in prices per token. So you have like downward pricing pressure that is inevitable driven by the competition.

Two is you have this like almost zero switching cost where let’s say, you know, Claude releases a new version of its model and I think it’s better and performs better against my evals and I’m currently using OpenAI, it’s not terribly hard in most application areas to switch to that model. So you have downward pricing pressure, you have zero switching costs, and you have a lot of competition, And then you have this kind of ever increasing sense of CapEx to drive that innovation forward that I think just like kind of is a death by a thousand cuts kind of impact on on the quality of those API revenue businesses.

And so I think what you’re going to see is that those businesses where you have where a company has the flagship model that everybody wants for a given use case, like you could say Anthropic has been very dominant, for example, in cogen, those businesses will look very healthy. But the moment another player releases a better model in that use case, like the revenue is going to yo yo. And so what I’ve really come around to is that it’s these products that are going to be built on top of these models, even inside of the Model Labs, not just by third parties, that are going to be much, much more compelling businesses long term than, say, the model APIs themselves.

Does Anthropic not have to acquire then and move into the application layer? It seems very likely to me that in some form or another, every single one of these frontier model labs will begin building both business and consumer facing products. And I think what you’re seeing with OpenAI is certainly foreshadowing of that. You could argue X is trying to build its own super app around its models, of course. And it’s hard for me to see how each one of these players doesn’t kind of end up in that place.

Now, there is a unique play here, which is SSI, Ilya Hsieh, Hsieh, and Daniel Gross’s company, which said, We’re not going to do that. And you could say, Hey, that’s a really contrarian move, but what it allows them to do is keep a very lean team and kind of go straight shot towards AGI, as they famously talked about.

would make would be today, there are a lot of decisions that trickle into the R and D and research organizations of, say, an OpenAI that are in service of helping them build better products in the short term. Right? Like the model needs to behave a certain way, so ChatGPT can be better or Deep Research can be better or whatever it may be. There are for sure short term optimizations being made there. I’ve seen it with my own eyes when I talk to people that work at those companies.

The argument for not getting caught up in that is you can just be entirely long term. You can make bolder research bets. You allocate resources differently. You can maintain a smaller team that’s more focused, as you said. So I think that’s really the argument, is like having to generate revenue and build a business that you can take public someday is going to come with short term thinking. And if you sort of say, Hey, assume we have access to capital that we need, and let’s say we can just kind of sweep all that aside, what does that buy you?

I think it buys you less distraction and more ambition. Did you do SSI? I suspect SSI will announce their fundraising and their investors and talk more widely about that.

My sense is, without going overboard on conflicts of interest, being in as many of these companies as possible, given what I know today, is very rational. But I also think it’s rational and becoming to say, Hey, we believe in one of them. We’re just going concentrate all of our resources in that. I think there’s no middle ground. You either have to be in all of them or you have to pick one of them. That’s my view. Can I ask, what do most people believe about AI that you think is wrong?

It seems like most people believe that this path to AGI is just this ever increasing upward slope. Again, I don’t have strong intuitions on whether true or not, but my instinct is we should be much, much more open minded to the possibility that we could arrive at some form of plateau and still have an incredible outcome for society and an incredible outcome for entrepreneurs and venture investors. So in a sense, you could say, Hey, AGI is already here in certain pockets. There are clearly things that these products can do that an army of the smartest humans in the world would never ever be able to try.

Now that we have test time compute, the same thing that brought us AlphaGo is sort of bringing an equivalent of that in all these domains where unlike the human brain, these AIs can just try hundreds, thousands of different paths to get to the best possible answer, whereas we kind of have to think about the best one. So in a sense, humans are still predicting the next token, whereas I think that these models with test time compute infrastructure are able to figure out, how can I try this a thousand times and then decide what the best next token is?

Right? And so I think as a result, like to me, AGI feels like it’s here in a lot of ways. And even if we were to kind of say, hey, the capabilities that we have today are it and it’s not going to get any better, We are so early in bringing these capabilities to bear in society and in industry that I personally think it’s an incredible wave of technology, even if the progress halts today. I think there’s a lot of people kind of holding out for this moment of singularity.

And not sure it’s as black and white in that as I’m not sure it’s as critical as that for the industry to play out favorably for all these different stakeholders.

It looked as though pre training was no longer as lucrative a scaling dimension as it was originally positioned as. I think there were a lot of people that were rationally saying, hey, like progress is going slow down and things might be the way they are now. Then this test time compute paradigm came along, and now we have things like post training and reinforcement learning that are presenting additional scaling dimensions that I think it’s just really hard to say. And it’s really hard to say because there are these amazingly talented people inside of all these research labs that are every day trying new things and trying to figure out what that next scaling dimension might be.

So I think the right way to frame the answer is when do we run out of new scaling dimensions? And right now, that appears to be very unlikely, at least in the near term. But at some point, it could happen. But I also have just learned to never bet against human ingenuity. And given the best and the brightest are now so heavily concentrated inside of these big labs trying new things every day, I personally believe there will always be new scaling dimensions. Whether the next one is as steep in terms of progress as the last is hard to say, but it’s really hard for me to

this is fundamentally different. And if you compare it to cloud, for example, adoption was fairly slow in the beginning, right? Kind of gradually, then suddenly. Here, I think the difference is that there’s this broad based consensus that failure to embrace AI to the fullest extent as a company is just like existential to its existence. And I think if you start to kind of ask yourself, like, what does it mean for the entire industry to conclude that if I don’t adopt this technology, I’m gonna be left behind?

What you start to see is this, like, voracious appetite to go and adopt it at all costs in every nook and cranny of the business. And I think this honestly explains why you’re seeing these companies grow faster than we’ve ever imagined. I think it explains why investors are so bullish on these app layer companies because they’re going and talking to these CIOs and hearing things they’ve never heard before. Like, I have to adopt this everywhere. I’m going to get fired type of urgency. So you’re just seeing unprecedented appetite and urgency because of this view that if I don’t adopt this technology, my company will perish.

And again, I think that’s just yet another reason why we’re in this really, really unique point in time for entrepreneurs and investors.

I think Mamoon is known as someone who is very metrics driven, and I think that could not be more false. He has this incredible taste in people. Stack rank. Okay. Market, traction, team. Team traction market. Why? I’ve already told you why I think people are more important than everything. Some of the biggest mistakes I’ve made as an investor is when I’ve ignored traction because I’ve had reservations about the Traction is so hard to manufacture, and I think that traction signifies a quality of founder that is good enough to figure out how to overcome any market size limitations that exist, and so I’ve learned to overweight that.

I do not think it’s a permanent structural challenge because I believe that in the same way that great entrepreneurs flock towards white space, there will be solutions and new products that get devised by investment managers to solve for these

in which both ends of the barbell really, really continue to rise, meaning small dedicated specialist firms on one end and large platforms on the other end. And I think that uncanny valley in between is going to be an increasingly challenging place to be over time if this notion of trillion dollar companies

been very generous to me over the years is the person that actually hired me into Venture. It’s this guy named Mike Dauber, who’s a partner at Amplify Partners, which is a wonderful new firm that’s really, really gotten off to an incredible start. Again, with focus on both early stage and these more deeply technical software businesses. They’ve since branched out quite a bit. But what I can tell you about Mike is he is just an incredible mentor and he has imparted so much belief and wisdom on so many young people in this industry and I think helped a lot of people gain the confidence and the clarity of purpose around what it means to do this job well and why they’re capable of doing it.

And I’ll be forever grateful for Mike for just always believing in me. Every time I’ve made a change in my career, every time I’ve had a tough decision to make, I’ve called Mike, and he’s really, really been someone whose feedback I value immensely.

first thing I’d say is we both love having each other in our lives. I think that’s the core thing that I think is just so important for that founder investor relationship to have. And obviously, there’s some derivatives of that that create that dynamic. One of those is, does this investor have a deep curiosity in me? Do they have deep and unwavering conviction in my vision? And do they have the instincts as to when to be helpful and how to be helpful? And most importantly, when not to be helpful and when to get out of the way?

I’d love to see Lightspeed continue to assert itself as a truly global, multistage, generalist platform. I would love to be a part of shaping that legacy in the sense that I have driven and help driven excellence in the early stage enterprise investing that the firm does. I don’t

to do it again soon.