VC: How Benchmark Picks AI Winners - Max 10 Bets a Year, 5 Partners | Chetan Puttagunta (GP)

We are roughly a $600 million fund with five partners and each of us makes about two investments a year and so the fund is making on the order of 17 investments a year. What AI native companies are going to be the winners of that window period of time? Primarily doing seed and series A investments, their high-conviction investments were taking board seats. So the last investment idea was data centers in space was Star Cloud. Jathan Furgunta, General Partner at Benchmark. How does that kind of speed to revenue change the way that you are looking at companies as an investor? Our company is now going to dramatically decrease the amount of time to get to $100 million probably. I mean I think when work day and service now and sales force, $100 million dollars they had compressed the window relative to on-prem vendors and had taken advantage of the cloud. What do you see in specific founders, whether it's their manus, playing chain, L'Agora, any of the companies that you're backing? They're like this is the right person to take it. All right, we're back with another special episode of the GTM now podcast. This is the VC bonus edition. I'm joined by my General Partner, Paul Irving, Paul Husgoing. Doing well. I'm really excited to talk about this episode as I was listening to Jathan talk. I think probably as furious as I was taking notes in any episode that we've done so far. So I'm really excited for people to listen into it and learn from some incredible investments. Unsurprisingly, that benchmark has made over the last couple of funds. Yeah, they've been on a tear recently for sure. We're in a couple of great deals with them. One that just got marked up. Not sure if we'll be announced yet when this airs. Phenomenal, what they do, he's a great follow on X. Always very insightful. And yeah, I mean, talk about a little wave in the ecosystem. They've gone from the fund is changing. They got the Jack Altman move, was a big move a couple months ago. Now they just announced the $2 billion fund, which is very different strategy than they've always been kind of the historically like, nope, we're going to stay in our kind of half a billion dollar range. They're kind of their model. And so there's definitely a lot of commentary on them, doing a bigger fund than ever before. I mean, what are your thoughts on that? Is that just the sign of the times where the outcomes are much bigger? So the fund size should grow with that. I mean, I saw posted the day that said, and we've spoken about this many times. But again, the 2010s, you were like a $34 billion company. Was like a big company. That was like a big startup. It took 10 years to get there. Now it's like in five years, you can build a potentially a $45 trillion company. Like we'll see where Anthropic nets out, right? So, you know, is that just inflation? I think when you see the headline, because I mean, one of the most historic and revered funds in history venture capital for good reason, you can see it across the ledger of portfolio companies that they've had through funds and in the door over the years. And it's their first growth fund. They've never done it before. And I think on the surface, you see the headline, $2 billion having the first growth fund, one of the last kind of pure play elites of a certain stage in strategy now becoming more of a platform. But then you, you know, as more news came out, you start double clicking on what the strategy is going to look like. It's still very benchmarking in the sense that the intention is to have from the sounds of it, you know, five or six or seven companies only in the growth fund. It's going to be highly concentrated. It's not going to change the sort of partner or platform model, so to speak. Like it's still a small team making really concentrated thesis oriented bets and ideally working with, with of course, generational founders. And I think you just see it in the amount of capital that gets raised, you know, they were series A, I think lead and series B for Saribris company that went public earlier this year, fairly recently and is absolutely ripped in the public markets post IPO. You know, they did multiple rounds of funding, you know, pre IPO post the benchmark investment rounds. And you start to just see opportunities to put more capital to work into incredible companies that you have an inside track on. And to your point earlier, if we're compounding value to companies that are an order of magnitude larger than we saw in the previous generation of, you know, cloud B2B companies or, you know, even the generation before that, then, you know, having another sleeve of capital to be able to put into those companies. So you're not getting diluted and you're owning more and more of iconic businesses makes makes a ton of sense. Yeah. And we did end up talking quite a bit about GTM in this episode. Sophie did a phenomenal job with it in my place, but, you know, we talked about two things, which is POC to trial as an accelerant. So getting, you know, getting a demo and a trial as fast as possible needs to kind of bigger wins faster. And then the power direct sales in the AI era. You know, now software is just so much more customizable than it ever has been. So you really do need kind of those forward deployed engineers that direct sales person. The person who's going to kind of, you know, in the past, maybe be a master's of ceremony moving the deal forward. And now they're more of like, Hey, here's how we help you like help the software customize for your needs. And tailor it to a certain degree. What were your thoughts on kind of those two pieces? Because I think that those were the big like, I don't know, pull the string on GTM parts of this episode that are really like. Yeah, because I think it's easy in the first generation of AI native companies that have grown really quickly. So many of them were PLG or sort of single user signups. So you think of cursor, lovable, replete, you know, these companies that have grown as fast as we've ever seen. But it's a lot of individuals signing up who are, you know, sophisticated enough buyers and swiping a credit card and getting to work. What Chaitan mentioned, which I thought was fascinating is that this next gigantic wave of AI companies is going to build a ton of value with direct sales. And then there's ways you can accelerate it. So to separate those two things, to start on the direct sales side of the house, any mentions LaGora is a good example of it. And I think it is a good example where, you know, there's big corners of the economy, whether you're building a horizontal or a vertical specific platform, that know that they need to adopt AI. These are C level board level mandates to pull their business into the AI era or they're going to get left behind. So there's budget available, but there's not necessarily playbooks that everyone can replicate on how to do that and what it looks like. And so if you're building the best AI native product for that particular category, direct sales becomes is very consultative. And this is why we're seeing the explosion of forward deployed engineers as well. I've like, let me help you take your business into the AI era. And there has to be trust. People are investing inventors for what should be multi-year relationships, where it's not just what does the product do right now, which should be magical to the point of demos and POCs. But what can you do for our business over the long term? How can you transform it? And that's just a really difficult motion to execute if you're PLG. You need the executive buy-in. You need the trusted direct sales relationship. And then you need this post sales FDE pretty hands-on implementation and go live just to make sure that they're having success because this is a huge investment as it should be on both sides of the equation. And then, I think you're getting as a leader these days, a million different AI pitches. So you know, we talked about how hard it is to kind of break in and how you got to separate yourself and differentiate yourself. And then if you even do get the the meeting, you know, the demo is that kind of that hook that draw, hey, can I get in? Can I use it? Can I see like my product in the demo? And this was something in the 2021s. You saw a bunch of companies. It was demo stack, reprise, arcade. There were a bunch in this space where it was like demo environments where you can like show somebody a demo of your product in real time using their product or their information and stuff like that. And now with AI, it's like, I think all those companies might have been early. Like you can finally kind of do this at scale. And while all those companies were early, I mean, it would be interesting to see kind of how it comes to revision text boxes, another one in that space. So yeah, I think there's just so much more you can do in the sales process now where you're engaging the buyer in a more customized manner. And you can convert them to being in the product very quickly and have a hyper customized product. Yeah. And that's where the compression comes from because you're still seeing, so now we're seeing the generation, you know, Lagora being the example that we talk about in depth in that actual episode, but of AI direct sales companies, with the direct sales go to market motion growing incredibly quickly. And the compression comes from this, okay, instead of doing this long 180 dcl cycle where we do, you know, enterprise wide deployment and we'll train everybody and have a notation. What's a POC that we can start with? And then to your point, you always talk to our founders about this, which I think is such an important call out of, don't just say let's start a trial or POC, be very clear about what the goals are, what success looks like, what conversion looks like when we hit all those goals and metrics and then what it looks like to expand, you know, post that first set of gates. But if you can do that correctly, and let's remove some friction. I'm going to show you a magical demo. You're going to be inspired to, you know, at all the different ways you can roll us across your business and how much value it's going to create. And let's just start somewhere. And if we can start somewhere, the expansion, as you implement and you roll it out and you invest a bunch of resources in implementation is really, really compelling. So, you know, we see it in our portfolio as well, these starting contracts, which they're not small in their own right either, but the speed to upsell and expansion is, you know, not with, not a 12 month cycle anymore. Like we're seeing it in one month, two month, three month, four month increments, where these accounts just continue to grow. Well, really excited to dig into it because it's a great episode. So without further delay, we've got Chase in general partner from benchmark on the show. Let's get into it. Our LP base spans from individual operators to institutional allocators. An angelist has been instrumental in supporting all of them. They handle everything from investor onboarding and accreditation to distribution and tax documentation, creating a seamless experience across geographies and fun types. Plus, all of this is available on a single modern platform. For an LP base like ours with over 300 C suite and VP level operators, this kind of white glove service and seamless workflows is so important. Also, instrumental that we support our institutional LPs that we're fortunate to work with an angelist is able to do so every step of the way. If you're looking for a platform that can support any type of LP investing in your fund, learn more at angelist.com/gtmfund. Chase and welcome to GTML. Thank you. Yeah, thanks for having me here in the benchmark office. It's great to be here. Welcome. Yeah. Well, let's start off here because you meet a lot of founders in this scary office and one founder in particular that you met by the name of Max dates back to early 2034. Believe it might have even just been down the hallway from here. That company, LaGora, now in 2026, just raised a $50 million extension on their Series D is now valued at $5.6 billion. And you met him very early. So take us back to that conversation when you're sitting down with him in the office here. You and Peter. Yeah. We met Max originally, I think February of 2024. The company was five people. It was still in Y Combinator and it hadn't gone through demo day yet. And you know, Max came in and had very specific ideas about how the legal AI market was going to evolve. The couple of things he talked about, which really resonated with us, was that the intelligence layer was going to get a lot smarter than what we were dealing with at the moment. And the right way to go about building an application for lawyers was to really ride the curve of intelligence over the next year or so and beyond. And the company executed beautifully against that vision. You know, we invested in March of 2024 and in the product launched in October of 2024. And at the time that the company launched their direct competitor had raised an evaluation of $3 billion. They had been market in market for a really long time. And so from that moment of launching the product in October of 2024. And 18 months later, at the end of March, 2026, the company grew from a million dollars of ARR to a hundred million dollars of ARR, setting all sorts of speed records for enterprise selling. And it's been a remarkable journey. But if you fast forward to that moment in time in 2024, the thing that Max talked about that really resonated with us was purely that product differentiation and a technology trend and betting on that technology trend. And at that moment in time, it was unique insight. If you'll recall just two years ago, what everybody was doing was building custom models. The models were not that capable yet of doing agentic tasks. And everybody was trying to build lots of custom frameworks on top of these models. And Max just had the view that the models were just going to get really good. And so just betting on the foundational models was the way to go and build that application. And that just turned that to be the right bet at the right time. But it's been an amazing journey to date. Definitely. And I mean, you met him in February of 2024, you said you invested in March of 2024. And they've just taken off. And that's not the first company like far from, you've got many different examples of times that you've built conviction quickly to win that partnership with a founder and then subsequently help them scale considerably to the likes of Lagos Gross and so forth. So another example, which is a bit of a crazy story is the minus one. So we'd love to hear a little bit more about when you met Manus. What was that meeting like? And how did things progress there? Yeah, originally met them in Tokyo. And the company at the time was pre-revenue, pre-launch. And you know, launched, I think soon after that meeting, probably a month after that. And they scaled zero to a hundred million in eight months. And so that was a different scaling factor because they scaled primarily through pro-calate gross. So that was users coming to the website, signing up, putting their credit cards and paying for tokens basically. LaGora in contrast is enterprise selling. So they have AEs, they sell directly. And so I think the interesting thing to zoom out is just how fast companies can grow in the AI era. And you know, there's lots of companies growing with just pure PLG motion. But what's really interesting to me in this AI world is that the direct sign can move so quickly. And so I think that's more about the macro poll in these different industries and different verticals of companies for the first time wanting to change their software stack to be AI native. And so they are much more willing to experiment. They're much more willing to try. And so you're looking at verticals that have not adopted a new stack in maybe 15 to 20 years that are finally willing to adopt solutions from startups. So what that opens up is this dramatic velocity opportunity for new entrants. And that's all, you know, that's the thing that's really cool about this moment in time with AI is just how fast you can move through the market. Even if you're going direct through salespeople and AEs and just doing direct selling, which could be considered a traditional playbook. But in the AI world, it's just experiencing a velocity that's like previously software companies in experience. Do you see an endpoint there then because we're in this window where people are looking to disrupt their stats and go AI native? Well, that window shot. So if you look at the internet and the cloud in the early days of both trends was this huge acceleration of the companies that established themselves as the dominant enterprise software players. And there was a moment in time where if you captured a certain category, you actually were able to build escape velocity and dominate the market. So a very specific example and perhaps the best example of the cloud era is of course Salesforce, which when it showed up, Sebel was the dominant CRM provider. And so if you just looked at how fast Salesforce grew in the initial innings of cloud, it kind of looks like the growth of AI companies. And what's amazing is how capital efficient Salesforce was through that entire growth cycle. Like they were an extraordinarily lean team selling licenses at an average of $5,000 and $10,000 and $20,000, which at the time was an unbelievable price point to be selling enterprise software. And so if we just look at how fast that velocity went for Salesforce, it looks a lot like initial innings of AI. And I think AI is even faster, especially at the foundation model level, like those companies are growing at unprecedented rates. But if you're in the early innings of these markets, I think the velocity is actually much faster. And I think one of the analogies to the foundation model is growth themselves is if you just look at the cloud, hyper-scaler growth in the initial days, so if you look at the US growth in the initial days, Azure growth in the initial days, Google Cloud growth in the initial days will actually in the Google Cloud more recently too. There are these early windows in a big macro trend where I do think a lot of enterprises re-examine their stack. And those are unique opportunities in time. How long will that window last? You can sort of approximate that based on the addressable market, both from a software spend perspective and addressable market based on services. So for example, we're talking about legal AI. Legal software as a market is about 40 billion per annum. And legal services as a market is about a trillion dollars per annum. And so the windows open until somebody can capture some significant share of both the software plus services market. And so in 2026, we're in the very, very early days of that market transforming. You know, like five to 10 years from now, I would say the window is going to be much narrow. or than it is now. 'Cause a lot of the incumbent stack providers will have been replaced by the AI native ones. Yeah. And those companies will now become the incumbents. And so you'll have that stack replacement. And then at that point, those companies, the customers of the AI native companies, they're gonna be pretty happy with what they're doing. And then to dislodge them again, is gonna be much tougher in five to seven years' time. And then I'm sure there will be another disruptive way of in a decade, it's time. Yeah. And then you'll have another chance to stack. So it's not that it's a, sort of a window that closes, it's just a window that sort of narrows over time. Mm-hmm. And so if you just look at what happened, from my perspective at Cloud, between 2009 to 2012, was like our great window for applications. And then the application window just became narrower from 2012 to about 2022. It just became narrower. And so you got more verticalized apps, you got more narrow apps, you got more focused apps. And then of course, AI's open things up. And then you can now do horizontal software again. Which is exciting. That is, yes, yes, certainly. And so now there's this window in period of time that we talked about. So for founders that give building, it's a very exciting time because you can capitalize on that window. But what comes with that excitement is also just a surge and the amount of companies be built for yourself assessing many different companies, but only allocating capital towards a very few select companies. Like how are you kind of deciding and looking at what AI native companies are going to be the winners of that window period of time? So to give you a sense of benchmark, we are roughly a $6 million fund with five partners. And each of us makes about two investments a year. And so the fund is making on the order of 10 investments a year. We're primarily doing seed and series A investments. Their high conviction investments were taking board seats. And so in that, in moments of great disruption, all of us are very much tuned to founders coming up with visions. I think it has to be technology, a lot of technology led vision that seems really interesting and perhaps radical and perhaps the sort of spear that can cut through all of this enterprise's demand and have this huge pull through in this moment in time. And so I think if you look at the common thread amongst all the investments that I've led here over the last couple of years in AI, I think that's the common thread across all of them, which is that the founders have come up with some technically interesting angle at the market they're going after in a very AI native way, which then has created a demand pull from customers and has created just an immense velocity for them. I think that's broadly true of all of the investments that we've done over the last couple of years, which is that there was some kind of foundational technical element to it. So you look for that technical differentiation and then that technical differentiation ultimately creates a distribution advantage. So of course the large incumbent software vendors and hyperscalers have way more AEs than any of our startups can ever get, period. And when we go to an enterprise account like Coca-Cola, some the incumbent vendors are going to have account coverage teams of maybe 100 people against one specific use case. And our startups can most put five people on that account. And so there is no way you're going to be able to outsell an incumbent unless the customer realizes the value and is pulling you organically. And so that has to start product out. And so there's only so much distribution, muscle you can build and hacks you can do and all that kind of stuff. But at the end of the day, if you don't have differentiated technology and differentiated product, it's going to be really hard to break through incumbent distribution. And right now because of this AI wave, these large enterprises are all being forced to reexamine their stack. So there's some kind of top down mandate. So maybe somebody in the board or maybe the executive leadership team has told everybody in the company, we need to reexamine stacks. Let's go AI native where we can because we see great ROI, we can save money, we can expand revenue, whatever it is. And so they're telling everybody in the organization, examine your stack, examine all the AI companies. And so that's unique in that everybody's now coming out and saying, OK, what are the interesting AI companies across these different functional areas inside the company that gives you a reason as a startup or chances of startup to pitch them. And then how you run from that initial pitch to a closed contract to a production deal, that's really interesting to me. And can this moment in time allow you to accelerate that time from initial contact to production in a way that wasn't possible five, six years ago? So traditionally, you would go selling enterprise. They would put you in a pilot, be a pretty big and complicated pilot. You'd get through that pilot. There would be some kind of security review at some point. You'd get bogged down the security review. By the time you put the system in production and integrated with all their internal systems, you would look at these sales cycles in our board meetings. And it would be like 180-day sales cycle, 360-day sales cycle. Yeah. Can't do that. If you're especially at the kind of speed that these AI companies want to go at, that is sort of like as friction full of a sales motion as you can get. And so the key then becomes how do you design a sales motion that is in concert with the enterprise that wants to try things? So that's a big challenge right now. And I think this is like a moment in time where there's a lot of innovative things happening and go to market that could really help startups break the distribution and the energy that the incumbents have because they have better products. And what are you sharing in your board meetings around how they're actually consolidating that time to value of going through the pipeline for potential customers? Like, what are they actually doing to speed up the sales process along that way? So none of this is going to sound particularly complicated. But I think implementing it is tough. I think number one, you have to have a clear value proposition that you can show people within five to 15 minutes. You have to understand that enterprises now in 2026 are likely getting somewhere between five to 10 demos of AI products every single day. And so they see everything. And so you've got five to 15 minutes to make an impression of what does your thing do and why is it magical? And so there is that hook of the demo has to make sense. The demo has to be magical. The demo has to really deliver. And then immediately, one of the interesting things a lot of companies have done is how do you go from demo to a quick pilot? And then how do you set up that pilot criteria in a pretty defined way, such that in a 30-day pilot or a 60-day pilot or a 90-day pilot, you have a clear rubric of what's going to be measured and why this is better than an income-muntz product? The part where traditional software was very frictionful, if you had a great demo and you wanted to do a 30-day pilot, there would be an integration phase where you had to go in, get a security check, security compliance review, all this kind of stuff. And then you'd have to integrate into their enterprise as data systems, and then you could run the pilot. Yeah. And so you have to figure out a way of can you run a pilot in sort of like a safe zone environment or whatever, where those integrations may not be necessary? And if you can do that and then run a 30-day pilot and show value, you've now collapsed what was previously a six-month process between the initial contact to the end of pilot to now. Hopefully a 15 to 30-day process. And so now then, and then your pilot goes well, you show success, and then you're now on the other side of that where you have a champion in your organization, and this enterprise is saying, we need to use this. I'm going to clear a path to get this into production as quickly as possible, because I see value, et cetera, et cetera. So these are not particularly complex techniques, but it is about taking a 180-day sales cycle and turning it into a 30-day sales cycle. And part of it is ultimately, it's humans doing business with humans, and I think that's the part that may be lost, is that ultimately you as a startup have to realize that we're all vendors, commercial vendors, helping a company do their business better. And so you have to become a trusted commercial vendor to the enterprise. And when you do that, you start to get these Big lump, big, excuse me, big jumps in efficiency. So you can go from 180 days to a cycle, 30 days to a cycle, you know, 100 days to get pilot interproduction, to two days to get pilot interproduction, integration time that used to be, you know, 50 days, 60 days, integration time now is seven days. So, yeah. It's all of the stuff put together is what makes this thing so much faster. Well, yeah, compression of time to value, all the other areas too collapsing, from a timeline perspective, would it go a little more specifically 'cause they had an enterprise sales motion? - Yes. - What did you see that kind of amazed you about their process and how they were able to scale, at the tremendous rate that they were able to scale? - So the big thing that they did initially, but I think created a big differentiation for them is that they, as I said, we invested in March, and they didn't go GA with their product until October. And the company had been investigating AI for lawyers for like a year prior to even the seed round. And so, they were just doing a ton of research on the industry itself. So you had the founders were three computer scientists, three AI people that didn't know law. - Yeah. - And so there weren't lawyers by background, they weren't lawyers by training. And so the legal sector was new to them. And one of the things that they did before they even started a whiteominator is that they basically set up an office inside of a law firm and shadowed lawyers and just tried to understand what lawyers did every day. And what was like the daily tasks that they did and what were the tasks that they needed to do every single day? And how could AI do stuff for them? - Right. - There was a ton of internal data collection that they did on just the industry itself. Like they just wanted to learn about what legal professionals were doing in these lawyers law firms. And then once from March to October, in March they had a working prototype. And so the journey from March to October was really about taking a working prototype and making it production-ready. In this whole time, because they were embedded in customers for so long and doing so much research about how lawyers evaluated software, what they used it for, how they did their jobs, et cetera. That allowed them to set up pilot programs and demos in such a way that resonated with lawyers in a really spectacular way. And so that doesn't happen accidentally or with some kind of like strike of genius. It was just purely that they had spent so much time with lawyers. I mean, we're talking like well over a year of just interviews with lawyers and just understanding what lawyers did every day to come up with these insights to say, okay, when we meet a specific kind of group inside of a law firm, here's what we're going to demo them. And we're going to say, with authority, because we've done so much research, these are the five things you do every day. And this is what we can automate with our software. And we know that you don't believe us right away. And you shouldn't. You should be skeptical. So set us up in a pilot and here's the rubric that you can follow. And we'll just give you this pilot for 30, 16, 90 days. And it's not going to be a traditional software pilot where you use the software and then the 30 days expire and whatever sort of like internal assets you've built to make this pilot work, the lights just get turned off and you've wasted all this effort. Their view was if you build IP around this thing, if you build integrations around this, it's a software or you build a workflow around it, whatever documentation you build to use our software, that's yours. And at the end of the pilot, if you don't like what we're doing, it's okay. And you've learned to use AI as a result of working with us for 30 days, great. It's fantastic. We're glad to have helped. And so there was a collaborative nature to their approach, which was you're looking for commercial vendors and AI vendors to take you into the AI world. And we want to be that trusted vendor to take you into the AI world. And so how do we design pilot programs that enable you to do that? How do you design a pilot program that enables you to trust the commercial vendor? Yeah. And then there's also a big investment that we made under data security, data sovereignty. And I think this is part of the advantage of being in Europe is that they were keenly aware of how regulatory complexity around the legal world, just how complex it all was and how you had to design the software from day zero to be compliant with anything. And because they started in Europe, they weren't building legal AI just for Sweden. They were building legal AI for Sweden, France, Germany, the United States, Singapore, et cetera. And so from day zero, they were thinking multi-jurisdiction. Yeah. There's a lot of nuance. Yes. And they were thinking data complexity and lots of data compliance around the world, depending on these regions. And so they were thinking multi-jurisdiction from day zero. So they would come to these lawyers and say, you are in a complex industry with complex requirements and we know that we're keenly aware of that. We've built this product from day zero, thinking through all of that. So all of this establishes you as a trusted vendor pretty quickly, run the pilot. And because you're running a very tight pilot, there are days with pretty exacting criteria. You come out of the other side and the customer is absolutely love you because they now understand how much homework you've done before you shut up with a product. Like none of this is just kind of happening. Like we're not throwing spaghetti at a wall. We're not figuring it out on the fly. We're not surprised by any of the things that you require. If there's anything that we don't have in the product that you need, we're probably aware of it already. We're already telling you, like you probably, you know, are thinking about this specific thing and look, it's in our roadmap, it's coming out next week or in two weeks or whatever. And so just being that prepared with that much knowledge is such a big advantage to somebody just showing up into a new vertical and just saying, I'll learn on the fly. And not just, you know, I, it could be luck, it could be skill, whatever. It just so happened that the Ligora founders had spent so much time researching the legal industry embedded themselves in a law firm for so long that now it's just turned into a huge advantage. Yeah. And the other part that they've done internally is they've hired a lot of lawyers to be the people that are doing deployments for customers. Right. And so they, they're, you know, they've created a role called legal engineers and they're essentially highly technical lawyers. These are lawyers with law degrees that have practiced for a couple of years at the top firms that also have a new very technical people. So they can implement software very easily. And they're coming into this kind of forward deployed engineer role. 100%. Like the trusted partner and interesting. So they can sit with, you know, the senior partner at one of the world's leading firms and that partner can describe their workflows and their work. And this person who has both a foot and product and engineering and in the legal world can translate that into AI powered flows in a way that I think is really, really spectacular. And so they're working on this motion where you're taking AI and doing so much of the last mile to deliver value to the customer that you actually end up making the customer business better. And, you know, everybody that uses LaGora sees much more efficiency in how their lawyers use time. Their lawyers are able to now stop doing all the wrote manual work. They're able to spend way more time on the strategic stuff. They're able to deliver way better services to their clients. You know, benchmark uses LaGora. We use LaGora with all of our outside firm partners and, you know, every legal firm that does work with benchmark does it via LaGora. And so the collaborative nature of LaGora itself, you know, just makes everything so much faster. So when we're, for example, you know, just earlier today, I was reviewing an agreement that we have with one of our companies because they're raising more money. Normally that would have been several emails back and forth. Yeah. But you can just sort of see it evolve in LaGora. And if I have any questions about where the document is, what were the changes that were done in the last week, I'm now talking to LaGora and not having to sort of these semis like black, these emails into black holes and waiting for responses. And so, so it's a complete, once you get used to working in this system, it's just such a differentiated experience that you now have completely different expectations for all your software vendors. And again, looping back to the beginning, which is like that ends up creating even more opportunity for startups. So I think once an enterprise buys one AI software vendor, they're going to buy way more AI applications. And I think it's one of those things that just ripple. And we saw this in Cloud, which was people originally bought Salesforce, then they bought ServiceNow, then they bought Workday. And once they bought all three, it created this exponential effect. And all of a sudden, it was like, we went from three SaaS apps to 50, to 100, to 500 inside one enterprise. And that similar trend is going to happen in AI, in my view, which is people are going to start with a Lugora, a Sierra. And they're going to have these AI applications. Of course, they'll have chat to you. They'll have a cloud. They'll have Gemini. They'll have these AI applications generate real value for them. I'm really excited about it. And then it'll just spark this again, refactoring of their stacks. And they'll go from having five AI vendors, in my view, to about 100, very quickly. So I made a couple of areas I want to dive down there, but you mentioned Workday. It was one of the first companies that created this wave and ripple effect. And Sierra and Lugora, now being part of the more current, worked it over five years to hit 100 million in AR. And Sierra and Lugora both did it in less than two years. You know, I think we're in a world where our company is now going to dramatically decrease the amount of time to get to $100 million. Probably. I mean, I think when Workday and ServiceNow and Salesforce went to $100 million, they had compressed the window relative to on-prem vendors and had taken advantage of the cloud. And then now the AI vendors are taking advantage of the AI window and getting $100 million even faster. I think what's interesting to me is not so much the time from $100 million, which is obviously happening really fast. But what's really interesting to me is how long it takes you to get to that first million. And that may take a lot longer in this AI world. And I don't know. It's one of these things that could be true, which is that because these last mile problems, these AI problems are perhaps more complicated. It may just require the founders to just spend way more time doing research and building before launching than in the cloud world. And I think the part that we saw a lot in traditional SaaS, which was you could have a very strong viewpoint on a market, building a vacuum, not have it touched the reality of customers, and yet scale really quickly and get to $100 million in five, six years, whatever. I think the AI world, the amount of time it takes to build could end up being longer. And then once you're out, the amount of time it takes to scale is faster. That's really interesting because that's almost like the opposite of some large narratives around product build time is collapsing. And now it's the actual scaling part that's longer. So I'd love to hear more on that because slowly orange is something of LaGora at having done some research ahead of time. I think if you just look at what these coding agents can do, it's absolutely magical. And so the ability to generate code is trending towards zero in terms of marginal cost. So the marginal cost of generating the next line of code is basically trending to zero. In that world where you and your competitors can generate a lot of code really quickly, the customer's expectations of how good the application is when they first try it is now ratcheting up very quickly. Gone to the days we old MDPs. That's right. So if you're selling the sophisticated users, they themselves can vibe code simple applications. And so the ability for a company to show up with a simple application and really get somebody to pay anything for it is essentially gone. And so if the cost of developing a product as viewed by the customer is effectively zero, the vendor has to build something sufficiently complex that a customer says, okay, I see value in paying real dollars for this because for my team to reproduce this and maintain it, even if the marginal cost of generating the next line of code is zero is just way too high because of maybe the time that it takes to maintain such a product or like all these last mile services they figured out, it's like, wow, you figured out a whole bunch of stuff around the code that I perceive to be valuable. Therefore I will pay you serious dollars for the species software. I think that's actually really hard now. And so there's a lot of innovation here where you'll see some companies building custom models as an example for their application. And that's a mode of differentiation because those custom models are very different than the foundation models. So there's like custom image models, custom video models wherever it is. Those are differentiated against the large language models. And there's a business being built around the differentiation of the model itself. And then what we've talked about with companies like LaGora is that they've built so much tooling and expertise on the last mile of what lawyers need that it's not the next marginal line of code that people are paying for. It's really that expertise of helping lawyers do their job better, become better service providers to their clients. So that stuff I think actually takes a lot longer than the code itself. I think generating an application now maybe an hour's time which used to take months. So the value then moves away from just the product build to what value can you articulate. And so therefore it actually might take longer to get customers to perceive this value in your application because they just don't ascribe any value to an application that they can think of. And ask Claude or Codex to just code for them. Or what I'm kind of hearing from you is the value accrual now accurs in the proximity to a customer and understanding their pain and helping them do well in their job as opposed to the software in any kind of capacity. Absolutely. I think if you just look at traditional business models, you have application vendor business models and I think you had service provider business models. I imagine the software in the future looks a lot more like service provider business models in the sense that service provider has to make your business better for them to get revenue. So they get revenue either on success or some outcome or whatever. And you can easily imagine a future where the value of code or the perception of value of code disappears and all the perception of value is transferred to what service do you provide me. Yeah. And how are you getting my business to be better as a result of the service? And that's probably where it trends in my view. And who knows, in six months this might change pretty dramatically too, just given how fast everything is moving. But it's certainly not that you can ship an MVP and that MVP is going to generate a lot of revenue and a lot of sales immediately. I don't think that's how it's going to be. Do you see a trend between more of a sales-led motion and PLG motion for companies excelling in the AI era? I think sales-led is here to stay. I think the, it's, it creates these like really durable relationships between the vendor and the customer. And I'm a big fan of that. I think the direct sales motion now, especially in a time of great disruption, it's, from my customers lens, it's very chaotic about what's happening. So you don't want to make a mistake by picking the wrong stack or the wrong vendor or some wrong set of tools, which then sets you behind your competition. So the competition makes the right decisions. You make wrong decisions. You have now harmed your own business. Not good. And so, and so, it's time of great disruption. We're really looking for trusted vendors to take you into the future. And that's where I think direct sales enables you to create these relationships in a way that can actually be very durable. And so I think if you just look at a number of these incredible AI companies. their product engineering teams are incredibly efficient and lean, but they're go-to-market workforces look a lot like traditional software and those don't seem to be pretty good at changing. Maybe they're more efficient, maybe each person can have you know a higher ratio of quota to OTE, maybe they're able to close a lot more deals, etc., etc., maybe they're able to just have a much higher economic upside because they're able to just go through these deals much faster, but the number of people I think is just going to be really really big because ultimately I think enterprises would want to buy from vendors they trust, especially if you're committing high-ticket budgets, you think like million, two million, five million, ten million dollar ACV contracts, you really want to be able to buy some insurance from your vendor to make sure that they're the ones that are going to be protecting you from all these issues that come up with AI, and what guarantees are you providing me and can I trust you as these intelligence as the intelligence layer gets better and better and better every month, every week, whatever, will you be able to deliver a service that enables my business to get better and also de-risk my ability to take advantage of all these new models. I love it. One thing you said there's take me to the future, and I think that kind of rings very true of everybody's looking for that guidance and support, especially as things are moving so quickly, very interesting there. And you know, one last question on the, and we're the founder side because you mentioned you look for differentiated product and that will lead to a distribution advantage which we've just talked through, but of course, you know, everybody's got a lot of grand visions on the product side sometimes, and how do you know and how do you tell if it's a right founder to actually take that vision into a reality, especially when they may not be coming from the industry. You know, this could be broadly, it doesn't have to apply to maximum. LaGora team, for example, but, you know, in that circumstance, you're sitting with, you know, three people and or five people that don't have legal background. I was like, why then? What do you see in specific founders, whether it's, you know, Manus, Langchain, Langora, any of the companies that you're backing, they're like, this is the right person to take it. You know, a lot of it, you're betting on the founder and betting on the people, and I'm not sure there's an exact science behind it. Yeah. Part of the great thing about early stage venture capital is that we make a number of investments in a number of founders, and some companies will work spectacularly well, and some companies won't, and that's okay. And it's part of the risk curve that where we invest, which is that not everything is going to just rock it to $100 million. Like money goes in 18 months later, we're at $100 million. It's not how it works. And of course, the journey is not done at $100 million. Of there are the journey continues. Yeah. And one of the things that we have a great deal of experience around the table is not only getting $100 million from zero, but a billion and beyond. And a lot of things happen between 0100, 100 to 200, 200 to a billion, etc. Like a lot of things have to happen. A lot of things have to go right. It takes a lot of time, a lot of conversation, a lot of effort. And not every company can fly through all these gates and get there. Some will in their spectacular, but not all can. And that's okay. And so part of what I've spent a lot of time thinking about is that the founders ultimately come up with these insights, and then they're able to turn these insights into spectacular products, these spectacular products then create distribution advantage. That distribution advantage creates great businesses. And they're all kind of linked. And then the cycle has to continuously happen, which is the product insights have to keep happening. So it can't just be a single, a single insight. And then you built a business for a decade. I don't think that works. The insights have to keep coming. Right. New products have to keep coming. You have to keep delivering value, etc, etc. So like the cycle has to happen constantly. And so you have to see the founder spike in some way. Yeah. And of course the founders are going to grow, evolve, get better. They're also going to grow as people as they build their businesses. And so it's a journey that's unpredictable. And part of being an early stage, a early stage focus firm is that you accept that. And you just go. And things are evolving. And when you're on the board, you're seeing all of these new issues come up daily or weekly. And you just try to puzzle through all of them the best you can. And just keep going. And that's it. Just keep going. I love it. Well, I got a couple last quick questions for you. And what's a company that you wish somebody would build or space? I think that we have a pretty strong limitation that's pretty obvious in how many data centers we can build in the United States. And I am really excited about every company that's unlocking data centers everywhere. So the last investment I did was data centers and space was Star Cloud, which is along this theme of just unlocking inference capacity wherever it's possible. And I think we just have a lot of room in the world to just really think of very aggressive ideas about where can we do inference for AI. And I think five years ago, somebody said we're going to do inference in space that would have sounded outlandish. But in 2026, that sounds pretty realistic. And it's pretty easy to imagine the next decade. A lot of data centers giving built in space. Definitely. And I think anybody that's thinking through all of the physical issues around AI continuing to grow. And I'm locking up are all really interesting companies at the moment in time for me. Very cool. And other than yourself, obviously, who's an investor that founders and operators should follow? Well, I think I would recommend everybody follow my four partners, Eric, Peter, Jack and Ad. Each has their own angle on the venture business that I think is really, really spectacular. And one of the fortunate parts of being in this business for me is I get to sit with them and learn from them every single day. And any content or podcast or tweet that they put out, I'd encourage everybody to follow that. Yeah, I would echo that also. And you know, you listed everyone there. Those will all be in the show notes in addition to yourself for anyone listening. But the newest addition to the benchmark team is Jack Altman. I'm curious, you know, what is it like now working with Jack who is also Sam Altman's brother, incredible investor. I'm curious what that dynamic's been like changing at benchmark other than having great podcast. Yes. I met Jack a long time ago. I met Jack. I believe for the seed round and the series A round for lattice. And so I've known Jack for a really long time. And you know, he's just a really great entrepreneur, a great investor and has just been such a positive influence in this ecosystem for a long time. And so I've been able to observe his journey for a decade prior to him being my partner. And I'll tell you he's one of the most spectacular people in this entire ecosystem. And so I just feel really lucky to be working with him every day now as opposed to, you know, trying to partner with him on investments or opportunities or whatever as we were trying to do before. But he brings a lot to the table. He has an incredible network. He has an unbelievable depth of empathy for the founder journey. And obviously he's a great investor. So yeah, really excited that he's here. Very cool. Yeah, you've got an incredible team at benchmark. And where can people find you and follow along to learn more? I'm Chethan P on Twitter. And so I tweet from time to time. So that's a good place to follow me. Perfect. Perfect. I thought we'd also be in the show notes. This has been a ton of fun. Thank you, Chethan. Thank you. Really enjoyed the conversation. Thank you. That was another fantastic episode of the VCU series on the GTMNow podcast. Head over to Apple, Spotify or YouTube and give us a like and subscribe. And we'll see you on the next one.