The Exchanges

Q I was going to ask about that. So these, these new managers, are you using them as scouts? Like, do you keep it bifurcated?

A Keep it bifurcated. Right. So, um, homebrew, Satya and I are individual investors in Screen Door as our Um, the other sort of VCs who helped get it started. We are a minority of the capital. The majority of the capital comes from large endowments, foundations, so and so forth. Um, we don't collect a salary from it, anything like that. So the only, my only upside is in the fund performing. I'm doing it because I care a lot about it, and I think it's fun. And the, the annual reports, the quarterly, there's, they go to ScreenDoor. They go to the team of three, um, who use it for their own underwriting and investment decisions and reporting. I don't have, you know, structural access to that. We're not processing that data to fill up our CRM, so and so forth. Is there collaboration, you know, opportunities amongst the portfolio and a homebrew, the portfolio and a forerunner, the portfolio and a precursor, obviously, and I hope they are, um, over, I think at the beginning, I probably sent more deal flow their way than they sent to me, but like over time that should, that should change. But I think of that more as like, um, uh, Like-minded investors collaborating to the benefit of founders rather than, hey, the cost of taking, you know, our money is, you know, you have to syndicate with us type of thing. I would never, you know, look to do that. Because ultimately I'm trying to, I'm t…

AI assessment note: “Keep it bifurcated. Right. So, um, homebrew, Satya and I are individual investors”

Q interesting to kind of understand the scope of how much it takes to fund something like this, um, for you and the, you know, the smaller scale, the micro, um, units that are only powering a thousand homes, um, what is the cost to build that versus, let's say, the really, really large, uh, nuclear facilities that we all see on the internet and are kind of like the symbol?

A Yep. Um, so the, I mean, the, the really big ones that we have now are on the order of a few billion dollars for reactor. Uh, a tiny little one that I'm working on is on the order of tens of millions of dollars, but you know, everyone's design is, uh, quite a bit different, but it's way easier to finance, uh, a much smaller unit like this, of course. Uh, so I think that's really exciting. And there's no reason that, uh, the software control systems, the digital twin, the regulatory work that we do couldn't apply to the larger scale. So I think micro reactors are in this really great place where they don't need a lot of fuel. They don't need a big facility. We can fit ours inside of the old EBR to dome at Idaho national lab. And in this hermetically sealed enclosure, Surrounded by these nuclear facilities and scientists, we can operate it in the most safe place possible to test the technology and then have that technology apply to all the other sizes and scales of, of reactors. So it's, Awesome. I'm super excited to be working on it.

AI assessment note: “really big ones that we have now are on the order of a few billion”

Q Were you poached by the team at USV or did it just magically happen over a couple of dinners or lunches?

A A couple months after Uh, Fundera was acquired by NerdWallet. I kind of Normally what happens after you sell your company in one of these businesses is you, you work really hard for a couple months, and then you're kind of like, ooh, things are kind of on autopilot and things are going well, and you gradually extract yourself from the business and start thinking about what a succession plan would look like. Um, and while I was doing that, I reached out to Andy Weissman, who's a partner here and been a friend of mine for a long time, who is actually the, the lead investor in GroupMe's first round of financing when he was at Betaworks. And Andy and I have also sat on the board of a company called splice, which was founded by my group, me co-founder Steve Martosi, who is a two time usb founder. Um, so we've known each other for awhile. I reached out to Andy and I, I said, Hey, can I be your intern? Um, I've got some free time. Um, I'm not going anywhere for awhile, but I would, you know, love to come in. Occasionally and help you and hang out. You know, I had been angel investing, uh, in tech startups for over a decade then. I had been an investor, uh, and an LP in USB's funds for over a decade as well. So, was intimately familiar with the firm and the people here. I was very curious to learn more about how USB worked internally. I felt like I had a really good idea of what everyt…

AI assessment note: “I reached out to Andy and I, I said, Hey, can I be your intern?”

Q How'd you get connected to them and what was the composition of the round? Was it mostly institutional? Was it strategic?

A A little bit of a, a little bit of both. Um, originally they actually reached out to me four or five years ago from the kind of industrial side because they wanted, um, to just learn about You know, new advancements of manufacturing technology, data and AI in manufacturing. So I just built a relationship with them of just giving advice to the industrial side. And then when they started the growth fund, started a discussion there, and it wasn't really until, you know, August last year, when I reached out and said, Hey, I think we're at the point where we're going to do around. Are you interested in, in looking at this? And that's when conversations, uh, really started. Uh, so in total, I mean, they were the lead, but we also had, uh, customers who invested. We also had, uh, Flat Capital, which is the investment company of the Klarna CEO and another partner, recurring capital partner who also came in and pretty much all existing, uh, investor had participated in a note prior to that, that converted into this round as well.

AI assessment note: “originally they actually reached out to me four or five years ago”

Q Right. And I'm curious. So are there specific manufacturing processes? I know you've mentioned a couple, um, are there specific manufacturing processes or industries that Odin is particularly well suited to address? And if so, could you share more examples of that?

A Yeah, it's really the ones that are kind of operator dependent. So like paper and pulp is a fantastic one. Really all types of plastics processing, like compounding of the actual material itself, extrusion, which is how you make anything that's long out of plastics. That includes wire and cable, packaging, building products. Um, we've started doing more and more metals. It's like metal drawing and different types of, uh, casting processes. Um, it tends to be a lot about how the data behaves in the process, because at the highest level, manufacturers care about the same thing. They want to know Utilization of machines, performance of machines, quality of the process, material consumption, and maybe some other efficiency metrics. And then underneath that kind of high level, we all care about the same thing. They have different like metrics associated with achieving those results. And so there are industries that are more Discreet in nature where like the data can be difficult to track from one step to the process ones from one step to another, uh, that makes it difficult for us to really show value on the whole stage. So that's why we tend to buy us more towards kind of continuous batch, continuous processes. Cause that's where we have a lot of variance, a lot of, um, a lot of cost of quality as well. Oftentimes it's also like the machine landscape needs to be pretty fragmented. …

AI assessment note: “paper and pulp is a fantastic one. Really all types of plastics processing”

Q Gotta love a perfect storm. So with this series B funding, what are you going to be, uh, investing in for your, yourself with the company? Like what, what are you going to be growing?

A A lot of exciting things. I think like there's, I'd put it in a couple of different categories. Um, one category is just scaling process AI. Um, and that means investing in the team and iteration on, you know, you know, continuing to iterate and advance that technology. But also in the delivery capability, the speed at which we can deliver it, um, the amount of recommendations we have available, the value of those recommendations, and just go to market for it. Um, the other categories that we're really focused on are what we call data enrichment. So how we actually use machine learning to fix some of the inherent problems with manufacturing data. And I can give a specific example, uh, cause since you're tying in data from multiple systems, those systems don't really have an accurate view of the world. And especially some of them might come from manual entry. So like you have one system that is reporting the raw values coming from the machines and you have another system where an operator is saying what is happening. Like I started product X at time, a, and I started product Y at time B and Every operator does that differently with like a variance of five to 30 minutes. And so if you just take that data, combine it with the time series data, it's garbage. Instead of having the operator do that, which is a non-value added task for the operator, we can have an inference model that…

AI assessment note: “one category is just scaling process AI... other categories that we're really focused on”

Q you know, overnight boom switch. Everyone's out of jobs. This will be a long, if not delayed kind of response of transitioning into whatever that next cycle will be. Um, have you noticed that people, uh, tend to overestimate, underestimate in next cycles? What would you say is like the biggest realization having done the studying of history throughout time of Reactions to these different kinds of booms and busts.

A It can, there's so many examples historically of major, major world changing technologies that took years, if not decades before they were really moving the needle. My favorite examples are the airplane, which is probably, you know, a top three world changing invention of all time, just completely transformed humanity that even the Wright brothers themselves didn't think it really had any commercial uses. Um, both the Wright brothers, their main, uh, priority after they've built the airplane, Was introducing it to the US war department. Um, because they, they themselves thought like that was really the only use case of the airplane was going to be dropping bombs out of it. And, uh, it, it, it took, it took decades from the, from the time the Wright brothers flew before there were like commercial civilian use cases for the airplanes. It's the same for the automobile. When the automobile came out, it was, it was inferior to the horse in every way when it came out. It was less reliable. It was slower. It was dirtier. It could go fewer places. And so even when the car came out, even if you are a car enthusiast back then, by and large, virtually nobody foresaw what it was actually going to become decades later. And the, the car first really came into use in the very early 1900, but it really wasn't until the 19 fifties that you saw the booming of the suburbs, which is really the, th…

AI assessment note: “major, major world changing technologies that took years, if not decades”

Q know, they're pumping up early stage valuations. They're kind of just tearing through the asset class in general. There's overfunding. There's overfunding of managers. There's just excess capital everywhere. It seems like we are still trying to reel ourselves back from that and get back to basics, but what do you think happened there? Like, what goes on in the mind of, uh, investors when there's just too much capital?

A I think, uh, on one hand, it was inevitable. It, it was as soon as you get to a Zerp world and then that much capital sloshing around, the odds that this would have happened were, were 100%. And I think why that is, is at the most basic level, every company's valuation, whether it's a startup or a public company or even a bond, is a number from today multiplied by a story about tomorrow. It's always what it is. You have a current number of like, here's earnings per share on a stock, and then you multiply it by a story about tomorrow. And the story is, I think this company is going to grow by this amount. It's, it's, that's a story. It's not a fact. It's just a story. And so I think once you have zero, then the story side of that equation becomes pretty much the only thing that matters. I mean, in any, in any, you know, discounted cashflow model, if you don't have, if you don't have, if your discount rate is zero, then earning a hundred dollars, 10 years from now is just as valuable as earning a hundred dollars today. So the story side of the equation became the most important part of it. And particularly in a, in a, like a social media driven world, the stories that people can tell are ridiculous and amazing and inspiring. So any story that you can make up about this company is going to be this in the future. It was like, great. I believe it's a great story. Therefore it's wort…

AI assessment note: “if your discount rate is zero, then earning a hundred dollars, 10 years from now is just as valuable”

Q Specifically with this round, what was the construction of it? Was it existing investors? Was it strategic? Was it debt? Was it equity?

A So we had about. 60% of the capital was new investors. 40% of it was inside investors. Andy Grayson from Construct stepped up with a huge check and has joined the board. And obviously Catherine from Andreessen, Diane from Founders Fund, Brandon and Josh from Lux. All stepped up as well. Put about 60% new capital and 40% insiders. Um, and then an entirely new debt stack as well, which is great for us because we're a capex-heavy business. So we like to spend debt instead of venture dollars on factory equipment. And then what happened was, uh, we were talking to multiple of our customers about strengthening our partnership. And that, that might potentially lead to an investment. So in December, it became clear that, uh, RTX ventures, which is the venture on Raytheon also wanted to invest. So we ended up, uh, extending the round out to give them the opportunity to invest in the same, round. And then what I can share is that more likely than not, there will be more announcements of that sort of a theme in the next couple of months. Yes, we're very lucky. Incredibly, from a fundraising perspective, and I know your audience is like a lot of people in startups, we're very lucky that it was an up round, clean terms, and we had both new investors, existing investors, and customers, or potential customers, um, who invested in the same round, so we're incredibly lucky. Um, and that's how i…

AI assessment note: “60% of the capital was new investors. 40% of it was inside investors.”

Q huge deal. Um, And Hadrian is at, you know, almost the core of it, right? It's like, it's like Andrel, it's Palantir, it's Hadrian. You're, you're one of them. So let's talk about this vibe shift. How are you feeling it internally? Are you feeling it with your customers? Are you feeling it with investors, um, with recruiting? What does it feel like, uh, being the boots on the ground?

A Recruiting is amazing right now. You know, two years ago, we were like, Please take this interview with us. Now it's like, oh, this makes sense. When can I interview with you? Um, both from Hadrian, but also like software engineers or talent in general, understanding that this category has got real momentum and is like valuable. And I think without founders fund Andreessen, like leading that charge from a narrative and capital perspective, the ecosystem just simply wouldn't be here. Um, like the, the level of momentum that Catherine Boyle, Trey Stevens have created narrative wise, capital wise is like insane. Like, you know, within two years, they've kind of taken something from like, don't know if I want to go work at Palantirand rule because like, uh, America bad to like the whole country is like, cool. We're building hard things for the national mission. Is, is insane. And like, it gives me a lot of hope for like, am I going to be better at this? Right? So we fill it on the ground with recruiting. We fill it on the ground from customers who are like, yeah, huge problem. We're going to take a bet on you way earlier. I'm going to help. And investors, I would say like, there are true believers who are doing it for the national mission, who have been investing in this category for five to 10 years. And there's this new wave coming into the ecosystem, which, um, some of them are …

AI assessment note: “So we fill it on the ground with recruiting. We fill it on the ground from customers”

Q So maybe, maybe we just even start from the beginning. Um, and how it came back to Rivet. So, how did you start your career, and what was the path to thinking about founding your own company?

A You know, I've been lucky enough to work with founders sort of since day one. Um, I started my career at Product Hunt. My first real job was working at my mom's restaurant. Um, but my first tech job was at Product Hunt. Uh, Neve Drawer recruited me to work on Twitter. Said, we need a guy to do our Twitter. You seem to enjoy Twitter. Here, have fun. And, um, I was really lucky to get to work with him. I was an intern. I eventually joined Product Hunt full time. I was the first hire after AngelList acquired Product Hunt. So, I was the new face in the office, in the AngelList office, alongside the rest of the Product Hunt team. I was there for two years. Um, I got to work across all of our social channels, our newsletter, I wrote that for a year. The AngelList newsletter, that was a really fun one to write. Uh, as well as some of our sort of like early, uh, SaaS platforms that we sold at Product Hunt. One was called Ship. I got to work on that. That was, that was really fun. Um, Nev and Ryan were sort of the beginnings of my career. I wouldn't be anywhere near where I am today without them. Um, so I spent two years there. I eventually ended up in venture, uh, working with Niamh at his venture fund called Truck Capital. I was there for fund two. Uh, fund two was a fifteen million dollar fund. He's now on to fund four, which is, you know, very exciting. Um, I was there for fund two.…

AI assessment note: “I started my career at Product Hunt. My first real job was working at my mom's restaurant.”

Q And so you're doing venture capital. More recently, I've been seeing your name and your face pop up on plenty of podcasts. So you have one with Jason Calacanis for This Week in Startups, and you're also doing one with Eric Thorenberg at Turpentine. So how did this VC route transfer into podcasting?

A Yeah, it's great. I'm also, I'm ad retargeting you, so I might be, I might not be on as many, many podcasts as you might think, but, uh, but appreciate the kind words, um, first with Eric. So, uh, as you know, Eric started a network called Turpentine, um, and for, for many years, people had told me to do a podcast. I didn't want to do a podcast because I think it's, you know, I'm big on, on coming up with Creative differentiated strategies. I didn't think at its, at its face doing a podcast was anything too clever or too unique, but me and Eric had a three and a half hour dinner at my place in Miami. Um, I think it was the beginning of last year. And he basically said, you know, we, we basically both decided that it would have been a good podcast, our conversation, uh, just, we weren't recording. And then, uh, he basically said, you know, would, would you like to, would you like to start a podcast together? And because of Eric and because of his incredible media reach and his reputation and all those things, I actually gave it a serious thought. Um, and I came back and I said, you know, what if we start a podcast interviewing limited partners and focusing on the very opaque world of limited partners? Um, you know, limited partners control, uh, our, our biggest part of the venture ecosystem and the venture ecosystem in many ways. Creates the future of humanity. So it's an import…

AI assessment note: “me and Eric had a three and a half hour dinner at my place”

Q a rational exit window. Let's start to sell it now. It might not become twenty billion dollars when it IPOs, or it might not become like five billion dollars when it IPOs, but Looking across our portfolio, five hundred million dollars, seven hundred and fifty million dollars, that looks pretty good. Um, is that when you would be considering this kind of strategy or would it be mostly multiples based?

A You mean valuation versus multiples based? I think you have to do both. So if you think something could only be worth 1.5 billion inside a billion, and there's a lot of risk, I think you, you have to sell all of it. That's really your, your analysis, right? I'm talking about some things that five hundred million has 10 X more potential. I think you sell 20% of it at that point, just to return. And one of the things that people don't think about is what is your LP base and what do they want? Your LPs are your customers. Maybe, and this is rare, but maybe your, your family offices want you to get, you know, a 12% return predictably. And maybe that means that you keep it in if it's, if it's growing at 20%. So it also depends on the LP base. Typically you are the riskiest part of your bucket. Uh, Chris Presti has a very long last name. Uh, uh, he, he manages three percent of one hundred forty billion dollars state of Wisconsin investment board, right? So he's paid to take risks. Um, so if I was to deliver an eight percent to him, you know, that would be, that would not be the point of that asset class in his portfolio, even though, even if his entire portfolio is made to return seven, eight percent. So I think it really depends on your LP base as well.

AI assessment note: “You mean valuation versus multiples based? I think you have to do both.”

Q So I've definitely started to see this narrative catch on, which must feel very validating for you. Although at this point you have won two bets against me, so you're on a roll. Uh, have you started to see others catch on to this? What's going on in the broader techverse, Twitter sphere, all the good stuff?

A Yeah, so a good close friend of both of ours, Delian, um, you know, recently on a, on a podcast, you know, had mentioned that, you know, SAS funding is down to a five-year low. Um, and then, you know, a layer deeper than that, he's also saying, you know, that VCs are finally waking up to the fact that zero marginal costs in the short term means zero margin in the long term. Um, and then, you know, transitioning over to something a little bit more along the lines of the dropping costs of software, and we're, we're seeing that happen, um, is that, you know, Christian Keel, Pronounce Kyle, um, is saying that the SaaS era is over. Um, and his reasoning is that it's really ridiculously easy to build and to copy companies or really easy to copy and build applications that are already built. Um, so what that really means is that your only moat as a business at that point is just distribution. Um, and distribution really at the end of the day is sales and marketing. And so, you know, starting to see sales and marketing budgets for, um, Test companies like go up pretty significantly. So there's not really a moat around the technology anymore. It's really around like the marketing and the distribution. Um, and then, you know, a big narrative shift that we've also seen with, uh, just like the more broad kind of general market, um, especially with David Friedberg and the all in pod is that…

AI assessment note: “a good close friend of both of ours, Delian, um, you know, recently on a”

Q I think it was Tesla versus Snowflake. We could, we can get into that later, but, um, in terms of this overall narrative, is SAS dying? Do you think AI is going to save software companies? What happens to all of those out there? Do they adopt this new component into their product suite or rebuild, or maybe this has happened before and we're being like a little bit dramatic.

A Yeah. So I, I've actually seen this story play out once before. Um, and so where I actually have like a fairly good insight here, um, is that my, my first company I started was a, was, was a predictive analytics company. Um, and so what I'm actually seeing with this current flavor of AI is almost identical to what happened almost 10 years ago now, a little more than 10 years ago, uh, with predictive analytics. And so this was the big rage back then. Uh, might've been before your time, Molly, but Um, you know, there was a, a huge run of acquisitions that Salesforce and Workday were going on. Um, and so, you know, Prediction.io was a, was a machine learning, um, as a service business that was open source. Uh, they were then bought by Salesforce and tucked in there. Um, and then Salesforce went out and bought a sales and marketing related, uh, predictive analytics business called Relate IQ, which is now their Salesforce IQ product. Um, and then Workday, obviously, you know, sitting on all of that HR data that they have, uh, was able to, you know, acquire a handful of predictive analytics companies, and funny enough today, actually, um, in, in how the comparison between predictive analytics 10 years ago with, uh, the, the current flavor of AI that we're seeing, Ramp just acquired a procurement business that was focused just on AI as, like, their moat to their business. Um, and so w…

AI assessment note: “I've actually seen this story play out once before.”

Q Great. Yeah. Well, hard tech has seriously taken LA by the reins. Everyone seems to be loving it. Um, however, there is a narrative violation here. In the previous, like most recent LP letter of yours, you did mention the shutdown of a popular emerging manager. What's going on there? Is this going to continue? Is this an emerging trend? Should we be concerned?

A Yeah. So you're talking about Countdown Capital, um, and Jay Malik, who's a good friend of mine. Um, and so I, he really outlines two, two main reasons for, for, you know, winding down the fund. Um, the first one was that pricing for aerospace and defense companies and industrials, um, was getting too expensive. Um, and then the second part was that it was just getting harder and harder to compete with multi-stage funds that were now coming into the space. Um, so I definitely agree with Jay on the pricing, uh, discussion. We've seen a lot of, um, supply of capital for very little, you know, high quality demand, um, in, in businesses and, in companies that are getting spun up. And so there are a, a, a decent amount of companies that are getting funded today, um, that I would say would fall more on the science project side, uh, less on the actual, you know, near term commercial viability side. Um, and so that, you know, that really kind of, Puts into question, you know, who's investing and how are they kind of taking what we had as a little cottage industry for the last, you know, three or four years, and now it's becoming a really main theme for a lot of different, a lot of different funds, um, and a lot of different, just like investors coming into the space. Um, and then kind of dovetailing on, on, you know, more investors coming into the space. Um, when it comes to big multi-…

AI assessment note: “So you're talking about Countdown Capital, um, and Jay Malik, who's a good friend”

Q And how did you pick Texas to start the company?

A Well, uh, Texas is the energy capital of the country and, and really should be the energy capital of the world. It's, it's in the middle of the sunbelt. It's in the middle of the wind corridor. So you have tons of, uh, wind and solar resources, but it's also a competitive market. It's a, a, a, a really great place to develop energy technology. You have price signals and kind of, um, Pro competition regulatory structures that allow companies to come in and innovate and take market share. Um, that said, we think that Texas in a lot of ways is the canary in the coal mine for the rest of the country. And as a lot of the rest of the country moves into, uh, a load profile that looks a lot more like Texas, the demand for batteries in those states will go up and we look forward to bringing our technology to the rest of the country over time as that happens.

AI assessment note: “Texas is the energy capital of the country and, and really should be”

Q Why is there such a backlog? Like, like, it was pretty obvious how much satellites were gonna get built out, so why didn't the U.S. build out ground stations for this?

A I, I guess it depends on the customer segment. So for, you know, a government customer segment, uh, they tend to procure their spacecraft and their ground systems separately. So, uh, space systems were the major focus for a long time. Uh, and ground segment was kind of just the afterthought. Like people think about what's the thing that you're putting up in space. And, you know, for a number of programs, the ground was just not really contemplated. And so that they started running into issues for that. And they were like, oh man, we need to, Actually account for the ground in our build out so that we can serve that capacity. Uh, and then on the, on the commercial operator side, it really is more of like a supply chain issue where you have, uh, a bunch of different, um, suppliers and then the, the actual operators of those ground sites, and then ultimately the, the software that's stitched on top of that. And for each of those parts, uh, of the value chain, It was just really challenging based on that structure to be able to be responsive to increased, uh, spacecraft demand. So as a lot of the operators are looking at launching new constellations, um, there's this moment of, of turnover where, uh, you have legacy operators who have their assets that are up in space right now, and they're thinking about, okay, what does our next generation look like so that we can be a part of th…

AI assessment note: “ground segment was kind of just the afterthought”

Q So how are you doing full stack ground support?

A It's a lot of disciplines that you tie into, uh, one company. So, uh, we cover everything from, uh, antenna development. Uh, we cover a lot of electrical work, a lot of digital signal processing, um, all within the, the hardware that we build. And then we have a whole sites team. So we actually think about how we manufacture and deploy those antennas, uh, not incorporates all the way back to the original design principles of what we build. How can we make things that fit in small form factors and that can, you know, go out and be shipped to any part of the world? Um, and they have to basically, you know, construct really simple versions of sites. So that's a whole different set of discipline. You have to think about how do you operate in different countries? So, you know, if you're landing a site in, uh, Japan or in Argentina or Iceland, all those wildly different contexts, how can you make sure that that service is like legally operating? Um, and also, uh, technically able to, to function in those regions. Um, so there's all those disciplines. And then on top of that, you need to have the software that can tie that all together, both from, you know, a user interface perspective, networking, making the hardware work. Um, so it's a lot of different disciplines that have to come together to make that possible. Uh, I think, you know, we really focused on top of the line talent acr…

AI assessment note: “we cover everything from, uh, antenna development. Uh, we cover a lot of electrical work”

Q Elon space ground stations and about, you know, the sensitive data and interception there. So how do you think about security and privacy and I guess, yeah, national security when it comes to ground stations and how are you fortifying on both the physical side of this and then also the software side of this?

A Yeah, yeah, no, it's a great question and it's something that we really think about up and down the stack. It's, it's all the way from, Uh, the networks, the hardware, the software, uh, to make sure that we are making it basically impossible for anyone to write anything to our local systems, um, or to tamper with our hardware at site. Um, I think we've really taken a software first approach with a lot of the security stack. Um, you know, we have conversations with customers who have historically really Uh, hardened up each of their individual sites. Think of like, you know, guards and gates and things like that. Um, we're taking a bit more of the SpaceX approach where, uh, they tend towards proliferation. So, uh, we have a bunch of sites and we have that diversity of sites so that we can make sure, um, if any one of those sites, uh, goes down, you know, not necessarily for some kind of malicious reason, but let's say there's like, uh, you know, a crazy hurricane that like wipes out The area, uh, we can continue to have continuity of coverage. So, um, yeah, we think about both at a local site level, being able to, um, have basically no intervention possible at site, as well as diversity of sites, um, for being able to, to cover for any site downtime.

AI assessment note: “we've really taken a software first approach with a lot of the security stack.”

Q Well, I guess before we get there, you went public before other space companies, one of them that recently went public. I don't know if you know SpaceX. I heard of them. Well, what was it like, and why did you make the decision? Well, what was the process, and why did you choose to go public when you did?

A Yeah, so, um, for a number of reasons. One, um, you know, if you can reach a certain level of maturity and go public, um, the capital markets enable you to do things you couldn't do in private markets. Buying Iridium. Like we could have never, ever done that as a private company. So, um, you know, it was, it was the right time for us to become a public company. And then the other part of it, which is a little bit selfish is I didn't want to build a space company that just lives for the lifetime that I'm running it. It needed to be multi-generational and go on and on and on and on. And being a public company really enforces that discipline of profitability and profitability is survival. You know, the, the rockets and satellites, and it's all cool, but at the end of the day, it's all for nothing if your economics aren't there and you're not comfortable.

AI assessment note: “the capital markets enable you to do things you couldn't do in private markets”

Q Why did you make that acquisition, and how does that change the future state of the company?

A Yeah, it's funny because I thought everybody would have picked that well early, right? Um, and I was, I was surprised that people were surprised that we would be doing that because if you just think about it logically and you stand back and you go, okay, what is the biggest TAM in the space industry? It's comms. Okay. If you ignore broadband because, you know, two really well capitalized people are taking care of that. If you ignore broadband, what's the next logical thing to do in comms? And it's like, you know, safety critical, defense critical services. That is Iridium. And, you know, it's a super quintessential Rocket Lab deal. Like, we don't, we don't come with hopes and dreams and bore a big hole in the P&L. Like, Iridium, you know, adds profitability to the entity straight away. It also buys us time to put up a new constellation. So it's like, it's, it's literally the quintessential Rocket Lab deal. Like, we don't, we don't buy companies that are, like, startups that maybe they'll do something one day. It's like, you have to be Proven and successful, and then you're the right kind of target for us.

AI assessment note: “adds profitability to the entity straight away. It also buys us time”

Q Can you explain what Ghost is? I found this completely fascinating.

A Yeah. So, um, uh, essentially, uh, you know, obviously we've got three launch sites, um, for, for the Electron vehicle and, uh, you know, our government customers, um, mainly around the haste missions, where we're really keen to see, um, you know, us go into very discreet locations, uh, to do some of that testing. So it sort of, it was driven from a requirement from the customer that can, can you build this mobile launch site? And, um, You know, now that we're doing Neutron where, like, even the smallest thing needs a crane, you look at an Electron launch pad and it's tiny now, so it's like moving stuff around is not a big deal for us anymore. So it's pretty, pretty elementary for us to, to build these mobile launch sites and flunk them down.

AI assessment note: “a requirement from the customer that can, can you build this mobile launch site?”

Q Yeah. Ben, I have to ask you, what is your hottest take right now?

A My hottest take right now is that I think AI is gonna converge, and, ah, we are going to build a digital twin of nature, and just like we have prediction models for, like, everything from, like, missile defense to, ah, tsunamis and earthquakes, we're gonna have that for nature, and we're gonna understand in real time, like, how nature is affected by every decision that we make, and that's gonna be Incredibly enlightening and terrifyingly scary, but I think we're going to build a digital twin of nature, we as humanity, uh, where we will be able to truly understand the consequences of our decisions, and then we will be able to leverage technologies like what we have to productionize species to combat some of the adverse effects of those decisions.

AI assessment note: “My hottest take right now is that I think AI is gonna converge”

Q What do you think the biggest mistakes are that people make during the acquisition?

A The biggest mistake that people make during acquisitions is underestimate the intelligence of the people who built the business that you acquired. Because sometimes you take on the imperialistic attitude. I bought you, hence you must work for me. Our attitude is you kicked our ass. Come tell us what you did wrong. Come run this for us because you did well without our money, our resources, our scale. So you must have figured something out. So we spend our time trying to understand what they figured out. Find a way that we can make them part of our culture, make that, absorb that capability, and we let them run it. And that sometimes really makes it hard for my teams because they suddenly have a new boss in a category where they thought we were acquiring something. But it's just, as I said, you have to approach this as humility because there are people out there who are smarter, faster, better resources, more resourceful than you in certain categories. And if you can embrace them in the right way, it allows us to build a durable business.

AI assessment note: “The biggest mistake that people make during acquisitions is underestimate the intelligence”

Q Did you always have that intuition, or was that also built up?

A I think, um, that's, that's definitely built up. It's definitely built up. I, before joining App Labing, um, you know, in my previous job, I already carried both the engineering part, also the product part of the work. Um, as I, as I shared in the beginning of the podcast, when I joined App Labing, I saw that the engineers were kind of disconnected from, from business. So one part of my job was try to help our engineers to understand the full context. Instead of assigning tasks to them and tell them to do exactly what they have to do, I show them, you know, how their work is impacting the business. And because our, our, you know, engineers are very talented, once they are provided with that kind of context, they have a much better sense of what to build.

AI assessment note: “that's, that's definitely built up. It's definitely built up.”

Q What do you think is going to happen to them?

A Uh, a lot of times when companies reach that point where it's tough and revenue multiple and potential cashflow multiple, if you do big layoffs ends up pretty low, they start getting accumulated by private equity or effectively plucked off the public markets taken private because the only way to recover it is a complete restructure. You fire a lot of people, you cashflow it, you lever the business up and that fits the private equity model. It doesn't fit the public markets all that well anymore. Once you can't Convey that you've got pricing pressure in a big market on the other side, you tend to lose all your, your, your multiple in the public markets, and then you're not a good public company.

AI assessment note: “they start getting accumulated by private equity or effectively plucked off the public markets”

Q How has AI changed the making of ads?

A It's early. It's not like, like, you can get a good clip, a short ad, out of some of the large language models. The challenge is to get them to put together a 30 to 62nd video advertisement for a brand that doesn't mess anything up, because you can't, like, anything wrong in that video, the brand's not gonna approve it to run. And so, can't mess anything up. Have to be engaging. And for a pretty long extended time frame, not that easy yet. So it's not at a place where you're just going to get a massive amount of inflow of advertisements that are diverse and creatively inspired and going to create lifts and campaigns, but their tools are available now so that humans can create ads much faster and much lower cost. So we've seen much more ad content coming into the system. It's just not at a point where the average marketer can go type into a box and say, here's a great ad and off we go.

AI assessment note: “tools are available now so that humans can create ads much faster and much lower cost”

Q Sean is incredibly involved with the El Segundo crew. Now you've expanded to Torrance, but what is it like working with Sean?

A Sean's amazing. Um, I, I think he was so far ahead of the curve on hard tech in general. Um, I mean, it, it really like is a true passion for him where, you know, a lot of VCs have kind of jumped on the hype train recently. Um, but when I met Sean three years ago, even before he invested in NEROS, you know, he was explaining to me his thesis of hard tech and why it's so important to invest in Manufacturing and how, you know, manufacturing has left America, but we need to bring it back, and all of these things that are really, really popular ideas now, like Sean was way ahead of the curve on, um, and he's just been a fantastic partner, so it's, it's great to be doing another round with him.

AI assessment note: “Sean's amazing... and he's just been a fantastic partner”

Q Obviously Sierra is in a different bucket too, but like, how do you fit in within that world?

A Yeah. So we are a hundred percent focused on voice AI infrastructure. So we don't do anything at the application layer. Where we focus, you know, is the models. So we create models that are Amazing voice models are amazing for all those use cases I spoke about. So healthcare, we have medical focus models, uh, drive through voice ordering, contact center, and note taking. We then build out the inference around those models. So to handle four X, the amount of YouTube volume on a, on a day, a hundred, twenty million conversations a week and growing over a hundred percent year over year. There's a lot of infrastructure we have to build out to make sure everything scales, is available, is like super fast, is low cost for customers. So we build out a ton of infrastructure, infrastructure on inference around our models. And then we also do a lot around the orchestration layer. So if you want to build a voice agent, if you want to understand speakers, if you want to translate data, we have a ton of, a ton of software at the orchestration layer that, that companies can leverage. And then, you know, above that, it's, we have this amazing developer experience, agentic coding experience, so agents can easily build with our stuff. And when we say infrastructure, I think a lot of times people think like, oh, you're just creating the model weights, right? Like you're just creating like models…

AI assessment note: “we are a hundred percent focused on voice AI infrastructure. So we don't do anything at the application layer.”