The Brutal Truth About Hardware Startups and Physical AI with Aidan Madigan-Curtis of Eclipse Capital

Speaker A: M if you're not brutally honest with yourself and the people around you, you probably are unlikely to succeed. And if you're not still able to see the path and the success and speak to the success and get everybody riled up to go achieve the impossible together, you're also unlikely to succeed. I think it's true on the manufacturing forum when you're trying to manifest something that seems, seems crazy and like it shouldn't be able to be done in a very short period of time. Uh, and it's definitely true of early stage businesses.

Speaker B: Welcome back to Tank Talks. I'm your host, Matt Cohen, founder and managing partner of Ripple Ventures. And today we're sitting down with one of the most unconventional and battle tested investors in the physical AI and frontier tech space, Aidan Madigan Curtis, partner at Eclipse Capital. Aidan grew up in a tiny mountain town with no mailbox and stumbled into Harvard when her acceptance letter sat undelivered for weeks and somehow turned that improbable start into a front row seat at Ray Dalio's Bridgewater. Aidan later helped launch the Apple Watch manufacturing lines and scaled Samsara's hardware operations from roughly employee number 20 through explosive growth. She joined Eclipse in 2021, well before physical tech became the buzzword it is today, and just helped close a $1.3 billion fund dedicated to early stage incubation and growth in the intersection of atoms. In this conversation, we're going deep in the themes that define Ayden's worldview. What she learned on those factory floors that most software first founders and investors still don't understand, the systems thinking lens she took from Bridgewater and how she applies that to hardware companies and the vibe manufacturing reality she and her team just saw on 14 site visits across China. And why the US still holds decisive moats in the Think Accent stack and what a genuine US manufacturing resurgence actually looks like in the age of physical AI. Now let's jump into the tank for this week's episode with Aidan from Eclipse. Thanks for joining us in the Tank today, Aidan.

Speaker A: Thanks for having me.

Speaker B: You know, I'm super excited to have someone focusing on the deep tech physical AI world, but for some of our listeners who don't know much about you and your background, you know, can you tell us a little bit about where you grew up, your early career highlights, and, you know, how you ended up in your current role at Eclipse Capital?

Speaker A: Sure. So for those who don't know Eclipse, uh, we have about 10 billion under management and we are focused uniquely at the intersection of physical sectors and deep technology. So think about sectors like manufacturing, supply chain, energy, defense, all of the tough stuff, the dirt under your fingernails, types of industries, the places where there hasn't really been actually a lot of tech penetration the past several decades. Eclipse, uh, has been dedicated to bringing technology to these sectors and investing in the great founders who are working there. For the past 11 years, I've been a partner at Eclipse for the last four and a half. Before Eclipse I was an executive at a company called Samsara. So large scaled industrial IoT company these days, 2 billion in revenue, still growing fast. Was, uh, part of the very early team there. Took us from no product, no revenue, just a bunch of guys in a garage and the mission to a public company. In 2021, before I moved over to the dark side, before Samsara and before Eclipse, I was actually on the manufacturing team at Apple that launched the very first Apple Watch. Specifically responsible for building out the factories that produced uh, the system and package which was the first of its kind for Apple. How do you get all of that intelligence chipset into a very tiny form factor so you can make it something elegant like a watch. We were bringing up full factories that essentially were in their stud format to something producing a million systems in package a week over the span of about seven months. It was pretty nuts. Uh, lots of sleeping on the factory floor and going all the way back. I actually started my career at Bridgewater, which is a global macro hedge fund. A bunch of super smart people out in the woods trying to figure out how the world works. So that's a bit about me.

Speaker B: Unbelievable background, super excited to have you. Obviously. We had Leor Susan, founder of Eclipse on the podcast early on when we started and it's incredible to see the trajectory and how ahead of the curve you, uh, know lior and the team was in terms of where the market was going in physical systems and physical AI. But you know, you forgot to mention one part which is that you grew up in a tiny Canadian mountain town with no mailbox. You got into Harvard basically by accident because the acceptance letter sat undelivered. Can you tell us a little bit about that story and how Harvard kind of got you off the ground in joining Bridgewater.

Speaker A: Sure. That is something about my background not a lot of people know. So it's one of those things that, not to be cheesy about the taking the whole village to raise a child, but I'm um, super thankful for my community. Growing up. I came from a tiny town called Penticton. Very idyllic place Most people that do know Penticton. First of all, most people have never heard of Penticton. But for those who have, uh, it was a place for their famil. Went to vacation when they were just little. So really a town that people visited occasionally. And there was one high school, and we were sort of nestled in the middle of the mountains. And, yeah, I lived in a really small house all the way up the mountain. And what happened was I had a guidance counselor who. I'm not really sure what they were eating for breakfast that day, but decided that I should apply for Ivy League admission. Which is kind of wild because no one from Penticton had ever been to an Ivy League school. But I did it. I mostly thought it was, like, a great way to waste my winter vacation. Um, but I did it anyway. And this was back. And I was just dating myself back in the day when you still put together a physical application. So I had no guidance on what this was meant to look like. I'm pretty sure I turned it into, like, a scrapbooking exercise, like newspaper clippings and, you know, sort of a little bit about, you know, all the sort of things that were me. And I wrote some totally unhinged essay on, uh, being a vegetarian and shipped it off. And for some reason, they accepted me. I actually got into Harvard and Princeton with this cockamamie set of things. Um, and, yeah, I didn't even know. So I was totally prepared to go to University of Toronto. I'd won a Millennium Scholarship. I did have pretty good grades and was really excited just to kind of go explore the world. And I got a knock on my biology classroom door sometime, like, mid April. So I had long forgotten about these macaroni, uh, art essays that I'd sent off to these Ivy League schools. And it was that same guidance counselor who was holding this big manila envelope. And she looked at me, and I saw that the envelope said Harvard. And just this look on her face, and she said, you know, they don't normally say no in the big envelope. Oh. And she's like, did you get one of these? Like, why hadn't I? I was like, no. And, I mean, yeah, we lived really far up the mountain, so we just had a little P.O. box. And I guess somewhere in there with all the sort of, like, grocery flyers and, you know, unattended to mail, was some slip from Canada Post saying I had a larger envelope that needed to be picked up from months prior. And I just never got the admission letter. So thank goodness she decided to walk. We got One. And she walked it over to my classroom and was like, hello. Um, yeah, I showed up like five months later, I guess four months later, like, true. I don't know, something that movies are made of with like a suitcase and at the Harvard gates. And there I was. Um, so, yeah, it was an incredible, certainly, like, life changing experience. Ah, Something that I'm forever thankful for. And I think that in many ways what I do at Eclipse, I do as a pure financial investor. We take money from, uh, pension funds, from hospital systems, from endowments, from foundations. Our job is to deliver financial return. But to be really honest with you, one of the greatest joys of what I do is that I can both be really focused on being astute financial investor, but also just be obsessed and in love with the types of things I invest in, which range from new forms of battery electric storage to, you know, really critical ways to be able to build residential, uh, housing faster through something like autonomous excavation or, you know, defense tech companies, things that are for the public good. And so it's, it's quite the joy to be able to both think about financial return, but also hopefully pay some of this kindness the world has given me back.

Speaker B: It's incredible. You're not the first person actually to come from a small town in Canada on the podcast and get a full ride to, uh, Harvard. Uh, Michelle Zatlin from Cloudflare, co founder of Cloudflare, also had a similar experience, except she didn't grow up in the Okanagan in a small town. Pick Tixon, which I know about a lot of wineries just south of Kelowna.

Speaker A: It used to be orchards, a lot

Speaker B: of orchards before, and now it's obviously moved up in scale to wineries. So, yeah, I just was kind of trying to figure out, given your upbringing and how you've, you know, grown up seeing the world differently than a lot of people that come from the coast or major cities. You know, how was your time at Bridgewater, working under the Ray Dalio style of systems thinking shaped the way that you evaluate opportunities and companies today versus somebody who comes at it from purely just an investor lens or purely from just a technical lens?

Speaker A: It's a great question. I think that Bridgewater shaped my thinking in a lot of directions. First of all, I would say I found a home at Eclipse both logically and culturally. And there are a lot of things that map back to Bridgewater. So on the systematic, fundamental side, for those that haven't heard of these terms before, you know, there's lots of different ways to Invest in financial markets. People these days, uh, may be familiar with day trading and vibes trading and momentum trading. Right. It's going up. Is it going to keep going up? We don't know. Seems like yes, let's hit it and then sell it later on. Like just the opposite of how Bridgewater trades. Bridgewater is looking for sort of the intermediate and long term arcs of what's going on within a given country, within a given market. And they're really looking to understand at a sub component level what all the different pieces are that relate to supply and demand pressures basically, which is what nets out a price. And so if there's a little bit more supply and not as much demand, you can see the price falling. And that's true across currencies, commodities, bonds. And there's obviously some complexity to it, but it's probably the easiest way of saying it really trying to map out algorithmically how all these forces come together to produce a price, then indications of whether those prices would be going up or going down. And that's how they. There you go. I've just given you a whole hedge

Speaker B: fund strategy that's like the voting machine, weighing machine example, right? It's like as simple as that, right. I love when people say, yeah, that's, that's the way the stock market is voting and weighing. It's like, I get it, but there's so much nuance and time between that.

Speaker A: There's a lot of nuance to it. I have in fact given you nothing with that answer, but at least to give you some lens on how Bridgewater is structured and what fundamental systematic is. So I think that there are aspects of that that pertain to venture investing. Certainly we are high conviction investors. And so what that means is as opposed to I mentioned vibes, I mentioned momentum and I think actually there's a role for all. But there are a lot of investors out there who maybe they can't set a price on a round, which essentially means they are unlikely to want to lead that round. They are unlikely to be the first mover. Eclipse, we are first movers through and through. We are people who come from operating backgrounds. There's um, nine of us as gps. We all ran companies like Tesla and Samsara and Apple and GE and Flextronics and Flexport. So we really understand kind of what it means when you're pushing bare metal with tech into these environments. And we have our own way of getting to know a landscape. Being very thesis driven, we almost always start with uh, the notion of what's going on in technology? How will that be disruptive across different industries? How does that relate to things that we've known and seen, seen from our experience actually operating in the world? And then where do we think this means the world is headed? And how do these different fundamental elements come together to intersect and create the opportunity to build something that is deeply disruptive and truly generational? And so we'll do a whole thesis landscaping, whether it's in quantum or in collaborative robotics, or name your disruptive tech. We, uh, have a view on the space, new electrochemistry, as I've mentioned, problems with the grid. There's umpteen things to talk about across the very 85% of global GDP that is more physical, world focused. And so we come up with our views on what tech, what team, what approach. And then we go out and we hunt, we look for that. Uh, if we can't find the thing we're looking for, we'll probably put the thesis on the shelf. Um, sometimes we try and find founders actually that can go build that, that are purpose built, uh, to sort of build the thing we have in mind. Or it ends up being kind of a bit of a collabor collaboration. And we'll incubate things too. But we have that kind of very fundamental approach. And so we don't need somebody else to come in and say, hey, I'm interested in investing in this. If we believe that this is the right group of people focused on the right disruption that have the capabilities to go build this like 10x20x100x better way of delivering an end outcome into a market, we will place that bet and we will be the ones to act with conviction. So in that way it's very similar. It's all about having your own internal set of indicators that are timed. Also Bridgewater does a lot of like time testing and pressure testing. So they'll take a view and they don't just take their own views for granted. They'll come up with a whole algorithmic ecosystem for what they think is happening in, I don't know, pork bellies or oil. Um, and then they'll go test that view as infrastructure over time and over history and say, will this algorithm and set of indicators have performed well during the great financial crisis? Would this have performed well during the Volcker inflation era? Right. And sort of running it over historical data in venture, you can't exactly do that, but you can certainly try and come up with a very kind of fundamental, rigorous sense of why do we think this will work? Why do we think it's truly got an ARC to become, um, sector changing or sector shifting over the next 12 to 15 years? And are these the right people that can build that uniquely with core ip with true differentiation? Can they raise money? Can they attract the very best people? And, uh, if so, we'll place that bet.

Speaker B: I totally see the parallels between the thesis writing and analysis that you can do at Bridgewater versus what you guys are doing at Eclipse. But I'll push back and say you can't really back test a thesis in venture capital like you can in public markets or in private public investing. The way that you back test a thesis, I guess, in venture, is that you stress test it with a lot of subjective hypotheticals that you can really try to underwrite the risk return analysis on those, but you can't really see them in real life because what you're hopefully betting on is something that has never been built before or that somebody is totally creating for the first time. Tell me why I'm wrong to believe that.

Speaker A: Yeah, yeah, I don't know if I totally agree with that. I think, um, I totally agree that the nature, it's, it's kind of beautiful to be able to back test in hedge fund land because you literally can just take the same algorithm and go run it over historical data and produce, you know, it's not 100% right. And there's things you need to accommodate in data sets, but it's pretty neat as a process. I think in venture, gosh, uh, on a daily basis we find ourselves asking what is different about this era, right? Are we in a bubble? How do you think about this version of overinflated valuations versus 20, 21 versus 2000. Right. So just, uh, some tiny cheap example of. I do think that history doesn't repeat itself, but it definitely rhymes. And so really understanding, yes, to your point, what is fundamentally different about this technology? Can it truly make waves? Why can it compete against incumbents that have all the capital and all the talent and all the customers? What truly will allow for this team and tech to break into a market to allow for customers to force them, to encourage them to incur switching costs. So there's a lot of, um, patterning, I guess I would say, that you can leverage to your advantage as well as ask yourself tough questions with like, you know, why is this time really different? So these are the types of ways in which we can look to historical patterns to try and understand what is truly innovative and unique about this moment, about these people, about that tech and versus, you know what, what really isn't.

Speaker B: I agree with that completely. I think what I was saying is that uh, you couldn't like back test the outcomes that you get in hedge fund land versus the outcomes that happened prior with social networks or with the search engines and say, oh, because the back testing showed us they fail, therefore this one's going to fail as well. Because that would be the opposite job as a venture capitalist. Totally, totally.

Speaker A: Yeah. You have to see not out there. And I agree with that, that there's quite a bit of, um, well, anyway, there's also a version of that in hedge fund land where if you've got a good algorithm, it should show you that it performs the way you expect it to, even on a uh, different set of historical data that you might think would be correlated with today.

Speaker B: But that's the alpha creation right there. But you meet so many founders that are probably very technical, align with your thesis, but sometimes they don't have the, you know, launching the first Apple Watch, manufacturing lines background or sleeping on the floor in factories. What's the one thing or brutal lessons that you've experienced in the factory floors that most early stage founders for the first time don't understand? That you need to make sure they understand before you can back them?

Speaker A: Yeah, especially these days I think that there is a, uh, heavy reward being placed on storytelling and vision casting and thematics, virtue signaling, which is important. Which is important. And it's important especially because it is one of the key reasons why dollars are flying from wallets right now. So you got to know how to play to your audience. However, what's very much true of manufacturing and also true of startups is that um, you want someone who doesn't get high on their own supply or you know, drink their own Kool Aid. Right. Like, how do you actually stay grounded and know with some amount of realism what the capability of your system really is, um, where its risks are. So I think that there's this truism around the what it takes to start great truly generational businesses, which is in one hand holding all the brutal realities of why what you're trying to go do is almost impossible. And kind of a bit of, I, uh, don't want to call it a fool's errand, but like almost impossible, which is what makes it hard, what makes it venture scale. Right. And at the same time having such a belief in yourself and in the team around you and in the people backing you and in your path forward that you can actually achieve it. And when one of those two things isn't really well balanced when uh, either there isn't a strong tethering to a brutal, brutal honesty of the reality of the situation and the ability to confer very quickly and concisely with your board members, with your team, with, you know, key people you have around the table of what those realities look like, how you make it through your J curve, all of the customer turn, whatever. Right. If you're not brutally honest with yourself and the people around you, you probably are unlikely to succeed. And if you're not still able to see the path and the success and speak to the success and get everybody riled up to go achieve the impossible together, you're also unlikely to succeed. I think it's true on the manufacturing forum when you're trying to manifest something that seems crazy and like it shouldn't be able to be done in a very short period of time. Uh, and it's definitely true of early stage businesses.

Speaker B: Yeah, I call them the honest optimist. Yeah, they have to be honest with themselves where the current state of the world is today, but have the completely blinded optimism of what they probably are going to have to will into existence. Yeah, so that's very well put. You know, you mentioned you joined Eclipse in 2021 when physical tech and frontier tech, they weren't really the hot buzzwords, they are now. What was the bet you were making back then when most of the market still didn't see it? And how has the last five years proven it right or forced you to update that?

Speaker A: Yeah, so I have continued to kind of make choices with my life that everyone thinks are crazy. I probably will continue to do so. You can call me a first mover. I was just thinking about that the other day. You can call me a first mover. You can call me just like off my rocker, whatever. I'm open to any and all descriptions. So when I left college, and I'll get to the tech stuff in a second, but when I left college, you know, everyone was going to uh, GOLDMAN Sachs and McKinsey. And I was like, I'm gonna join this, like, looks like a cult in the woods full of just like some of the smartest people I've ever talked to. Who by the way, the way that they interview on campus was wild. It was like a, it was like a really funky lunchtime conversation where they come and, you know, have like a group conversation. So you've never been in a group inter, uh, setting. I highly encourage it. It definitely helps you confront a lot of awkwardness. That you didn't realize you had. It was wild. And just from the very get go I was like, who are these people and um, why are they so fun and interesting to talk to? So Bridgewater attracted me, uh, from the get go and I just felt like it was a place where deep intellectuals were tackling some of the most complex problems in the world. But I definitely was flying in the face of all things rational and where all the kind of like logical people were going. I joined well before the firm was well known. It was kind of weirdos in the woods. Um, we grew to like, you know, from just a couple hundred people to um, almost 2,000 people. While I was there, I think we, you know, really ran up in our aum up to 150 billion. It was just this wild experience to be there. And so by the end of my five year period, it seemed inevitable. It felt like, you know, clearly I was on this amazing path. And that's exactly when I decided to cut the ripcord. I go to tech, right, and people are like, what are you crazy? Right? Like you kind of told you how I grew up, like, tiny house in the middle of the mountains, like we didn't have any money, right? And so to be leaving this kind of salary and to be leaving in an environment that I really did enjoy, uh, I got a lot of skepticism, especially my mom, like, what are you doing? And I just really wanted to be on the side of the builders. So yeah, did that took a massive pay cut, uh, irreconcilable pay cut. Like it was crazy. Paid, uh, for school for a little while. And then when I was leaving Stanford, um, as a master's student and everyone was going to be like a PM at Facebook and you know, join like Uber. I was like, I'm going to supply chain at Apple. It was like, what's wrong with you? Like that's like the least sexy job you could get coming out of Stanford. Like you positioned yourself to go do these like crazy fun things and software is blowing up and eating the world and whatever. And I was like, yeah, I get that. And I was excited for that Cambrian explosion, but I wanted to be a part of bringing the power of technology into our day to day lives. And I believed that there needed to be more points of tangency between the digital world and the physical world. And so I wanted to go to Apple, which was world class. Absolutely. Especially at the time, still probably now, like best in the world at manufacturing physical world technologies. And I wanted to learn from the experts and I wanted to try my hand at, uh, building some of these things and really kind of being on the ground and seeing it firsthand. So I did that and I had a wild experience and it was really fun. And then right as the product that we'd built at Apple was really succeeding and the Apple Watch was going really well, and then everybody wanted a system and package and we were going to do it and the MacBook and the iPad. And I was like, oh, God, maybe like the sip girl at Apple for the foreseeable future. And on the one hand, again, you know, in terms of like remapping my career could have been great. On the other hand, I was like, I'm going to go to a startup. I was like, what are you doing? Like, you know, your Apple stock is flying. Like, you actually chose a good thing, don't ruin it. And I chose to not go to any startup, but I went to a hardware startup and you just go to any hardware startup. I went to an industrial technology hardware startup and people were like, are you insane? I was like, I really believe that we need to bring these, like, tangential technologies that can create those bridges into the last mile of our economy. And that's what was attractive to me about Samsara. It was serving modern technology to the industrial sectors. The sort of, these, like, last bastion of sectors that had really been not touched, pushed, by the way, are massive, massive sectors. And we had a crazy ride. And so, you know, then we're going public. And I'm a public market executive and had a pretty cool career progression as a function of that and a pretty strong financial return. And, you know, I'm like, I'm going to go to a venture and people are like, what's wrong? What is wrong with you?

Speaker B: You just love swimming upstream when everybody's swimming.

Speaker A: Yeah. And I was like, no, don't you get it? Like, this was right. And now let's take these lessons.

Speaker B: Obviously your gut was always right.

Speaker A: Proliferate them to energy and defense and the rest of supply chain and manufacturing and there's so many other places where technology can penetrate and make a difference. And so I just had this crazy experience and so that's what brought me to Eclipse. And anyway, so hopefully that makes sense in terms of why I keep choosing these things and where to from there. But I really saw Eclipse as this opportunity to take those lessons and continue to sort of push the efficient frontier. And hopefully that gives you the answer.

Speaker B: Well, you obviously like to go where nobody is focusing their attention on because that's where you seek out opportunity and challenges which it seems like you like to, to put yourself in front of is a challenge even if the opportunity is not well known to others.

Speaker A: I think it's required as a venture capitalist. For what it's worth, there's like an I love um, Sarah's name of her fun and podcast. Like the no priors thing. Right. Like you've got to go where it is entirely unclear and extremely not evident that the world is absolutely going. And by the way there's a lot of ways to get that wrong. So that's why venture is more of a bets business that, that has a power law to it. Right. But yeah, it's neat that a whole industry has wrapped itself around.

Speaker B: So when was the first time you, you went to China? Was it for the Apple manufacturing plants?

Speaker A: No, I started in China when I was a college student. I went to volunteer at an orphanage and I was just very kind of curious to. I didn't get to travel much as a kid based on the sort of economic upbringing I was mentioning. So this, this was a chance for me to do some research on international law in China and also do not for profit work. So I was 18 and I was in Beijing while I was in an ah, orphanage kind of about 40 minutes outside of Beijing.

Speaker B: So that was on a kind of charitable mission. But from a business standpoint, when was the first time you got to see like real scaled operations and manufacturing in China?

Speaker A: Yeah, you know, interestingly at Bridgewater I also covered China um, for a few years focused on like the political economy. So. Well, I wouldn't say that was. I was not visiting manufacturing floors at the time. I was taking trains like around the country looking at the way that construction was proliferating throughout different tier three, tier four cities. And you're trying to understand the sort of economic drivers there. So definitely got a sense of what was going on in the country. But my first manufacturing floors experience were definitely at Apple, uh, in China.

Speaker B: I want to talk about China because you guys released this amazing report, China Field Notes, you know, in the US and kind of this China reality check and I think it's really important. Your team just came back from 14 site visits, I believe across China you've seen like all the big robots, you know, from BYD and others and all the thousands of tele operators and Full Stack Robotics, uh, you called it Vibe manufacturing. Can you explain what vibe manufacturing is for our listeners and what exactly you were able to see there and how they're compounding the magnitude of delivery of robots um, versus the U.S. sure.

Speaker A: So the vive manufacturing concept, which was coined by my partner Charlie Mongi, who's incredible and the kind of guy you can just listen to talk all day, he's super thoughtful, used to lead a big chunk of Tesla's manufacturing and was the head uh, of manufacturing for Rivian as they were building out all of their initial manufacturing sites. Now he works with us at Eclipse, does a lot of our robotics investing, uh, alongside myself and a few others. So the vibe manufacturing concept is really pointing to China has been over the last two and a half decades or so, almost three compounding manufacturing, uh, ecosystem that started with some people decades ago wanting to leverage cheap uh, Chinese labor to be able to manufacture for greater margins. But honestly, incredible country, incredible, um, spirit of innovation, incredibly hardworking people, very clever overall economic structures. The economy is having a hard time right now, but it's still been quite impressive how they've essentially compounded on their growth curve. The concept is like, it's so far from this concept that a lot of people have of Chinese manufacturing where it's the fact that a, uh, there's 14 layers deep in the supply chain lead to being able to manufacture really quickly and rapidly, almost anything you can imagine. So these folks have received something to the tune of like $2 trillion worth of foreign direct investment over the last 25 years from companies that have wanted to invest in the capex, invest in the people, invest in the multiple layers of supply chain so that things can kind of arrive quickly and efficiently. As a function of that, um, the manufacturing companies themselves have also, people have kind of grown up on these manufacturing lines being forced to be innovative, being forced to like push harder, win against their competition. By the way, it's a super competitive environment in China. People don't understand this about China. They think it's these government backed SOEs. And sure there's a lot of government involvement, but there are multiple government types of funds that are investing in dozens upon dozens, in some cases hundreds of similar styles of companies. And each province and each municipality is looking to create the national champion out of their company that they backed, right? So you've got these crazy competitions happening within China, within provinces, within cities, amongst different companies that are all trying to do the same thing. So that pushes the innovation levers. They figure out how to write their own software codes on the manufacturing line to make things really easy to shift and move and change. They uh, have really hardworking people who are still, you know, there's still something to the tune of like 500 million people in China who are living in a fairly substantive way. So they've got people coming in from the countryside, coming to get educated, want to stay in the cities, are willing to work in pretty tough conditions, are willing to try hard things. So you've got this compounding of the human capital, compounding of the physical investment capital, and just kind of clever and innovative ecosystem with a ton of competition and good government backing. And that's what creates just this sort of like proliferation of. It's sort of like a, uh, being in an elf shoemaker shop. Like you could get anything made as quickly as you want, wherever you want, whenever you want, within the span of hours or a day, because they've just got the tooling and the capability to go from idea to end output, given all of those pieces very rapidly. And yeah, happy to kind of translate that to robotics, but that's what we said when we talked about vive manufacturing. I mean, you want a different look of an actuator, they can do that. We didn't make actuators at all basically in the United States. So our manufacturing process, forgings, casings, tooling, like they just have everything there. It's hard to describe other than to say it's like literally a trillion dollars of just equipment cap. Like it's so much that's been compounding over the years. Plus you've got people who actually know how to sort of fluidly program all of the actual pieces of technology that dictate how machines operate. They've got the consumables, they've got all the raw materials, and they've got the desire to innov and compete and win. So that's manufacturing.

Speaker B: Well, I want to just double click on the FDI part because I think that's really important for people to understand why maybe this can't even happen in US and Canada, North America, unless we change our structure. I spent one semester working at an IP patent and trademark law firm. And what their job was was to go after all the fake goods that were being sold on the streets from Louis Vuitton, Hermes, Nike, you know, all the brands. And we would end up finding out where the supply chain was coming from, from. And what we found out was that it was coming from China, but from the same factories that were producing the real goods that they were making the actual goods.

Speaker A: They call it the third shift.

Speaker B: Exactly. And so what would happen is they just don't go through the supply or uh, the right checks and balances when they come off the line at nighttime, but they basically are Made from the same goods. Right, because they're sold out the back door. But then I started thinking to myself, wait a second, this American company supplied all the capex goods for building out that factory in the beginning. Fdi, but it's not owned by Nike because you can't own property as a whole 100% subsidiary owned by Nike. It has to have a 50% local partner, I believe. Right. Uh, the government kind of controls who owns the assets. You get all of the IP learned inside China, all the manufacturing, you know, experiences given to the employees and then they come back and say, okay, sorry Nike, we don't need to actually produce your stuff anymore. We're going to produce our own brand, BYD it. And now we've got all of that, you know, talent, IP knowledge, know how and thank you very much for the free capex build out of our factories. And we still own everything here locally. We're going to take advantage of that. The chi fi is something that I think is so real now. How is America going to compete with that if they don't start nationalizing the ways in which FDI comes into their country?

Speaker A: It's a great promise. I think. It's complicated. First of all, you are actually seeing more nationalistic capital behavior than I think people even realize. Government's taken stakes in like chip companies, um, the actual CHIPS act and this is bipartisan by the way. I mean certainly there have been activities this year, but CHIPS act actually and sort of the relationship between the US government and TSMC was quite thoughtful. You know, how do you attract that relatively large investment in, on um, behalf of a foreign company? Uh, how do you structure the right amount of non dilutive capital that can go that direction as well as bring together a syndicate of uh, infrastructure investors that will get a reasonable return. What guarantees do you give? You can even look at like the MP Materials deal, right. That's going, that just happened right now and say okay, so the government's now like taking stakes in things and using tools in its tool belt across OSC and across DOE and across a couple other major departments to incentivize and also own, um, um, some of these potential national champions. So I do think that in some ways, um, first of all it's not new. I mean the DoD or DoW, whatever we're calling it these days, was actually pretty deeply involved in creating Silicon Valley.

Speaker B: Right back of course, National Semiconductor and stuff.

Speaker A: Yeah. And it was the, by and large, it was some um, early stage grant making that got that off the ground. It was also the Offtake arrangements. Right. Especially during a period of war and post war, to sort of create nuclear arsenals and protections around those. And so, um, when the government is an off taker, it's also a very powerful thing. And so that's kind of really what you've seen in some respect is like a bit of a blueprint in China. It's sort of innovation capital, directed innovation capital on the front end. And it's the government acting as a major off taker too from uh, in its selected national champions and a few things in between. So I don't think these are new playbooks. I think they are known playbooks the US can get in its own way a little bit, but I think we're also seeing in a bipartisan way, uh, a greater shift in bullying and ability to leverage some of those toolkits.

Speaker B: No, makes sense. I mean you've seen obviously the amazing, uh, amount of capital going into the robotics space in the US but you just came back from the trip in China and got to see some pretty cool shit. I mean, tell us what you saw in the robotics space happening there and where the US is either ahead or very far behind and needs to catch up.

Speaker A: Yeah, the u. I'll start with the good news for the U.S. i would say the U.S. is ahead, or at least highly, highly competitive. Uh, it's ahead in what you would actually describe as like physical or embodied AI. So how do you actually create a level of situational awareness and intelligence in operations such that a machine can truly operate and fulfill a goal, uh, without human intervention? I think about a company that I invested in, um, helped get off the ground actually called Bedrock Robotics. It's an autonomous excavation. Well, it's an autonomous heavy equipment company. So with uh, a harness, with some compute and uh, an autonomy model put on top of it. These excavators, bulldozers, wheel loaders, can actually act and behave autonomously. And that's not to say they can do a drone show. Like they can actually go about the digging and the trenching and the work that's required following a plan, a sort of broad plan. By the way. There's a million chaotic things happening in these environments. These, the way these equipment need to operate, have us be safe, must be precise. Right. So it's actually quite hard to do this level of autonomy. We're seeing it um, work as well in something like a Waymo or our company wave for self driving. Like there are ways in which US companies are actually delivering real world embodied AI outcomes and that is hard to do, it's still very vertically focused. So maybe we can do it in construction, maybe we can do it in self driving. No one is there yet on general purpose robotics. Uh, you've got some interesting folks in our portfolio as well as others pushing on those boundaries. But we have a major data uh, dearth problem in terms of the actual parameters it takes to create what seems to be fruitful for general intelligence behaviors in the LLM space. Certainly in the physical world space. Neither China nor the US have um, an efficient, effective answer to full embodied AI general purpose intelligence. Now where the US is really behind is on essentially from the ground up all things manufacturing, um, especially some of the more complex things to manufacture like magnets and actuators as well as just the ability as I mentioned to vibe manufacturer to kind of have an idea and to go actually create and iterate on that idea, physically speaking. And then some of the things that I think people aren't pricing in about China that are quite impressive are what I would describe as like the lower level software policies that are still quite critical to allow for a robot to move around the world. So stabilization policies, locomotion and navigation, they are able to, in a way that I haven't quite seen here. They have really good software on that front as well. And then the software hot rock hardware combo that makes for things like crazy fighting robots and stuff like that. So you still, it's almost like a live, a real world like Mortal Kombat where someone will be like playing with a robot and the robot jump kicks and flips and stuff like that. And you're like how is that even possible by the way? I've not seen a US company that can produce that level of robotic control. So that is something that I was impressed by in China, uh, battle the

Speaker B: bots, you know, 2028 version. It's like the new UFC is just going to be controlling robots to fight each other to the death.

Speaker A: Yeah. Which by the way it's pretty entertaining. It was also kind of scary.

Speaker B: Yes, exactly. It's entertaining until they break free of the cage.

Speaker A: Yeah, it was uh, it was both terrifying and wonderful all at the same time.

Speaker B: I want to double click on the data dearth problem because that's something we had even with LLMs. Right. You know, that's why you've got the markors and stuff taking off and things like that. And you know, what is it hugging face came around and had their own data labeling, you know, run up. What are you seeing in the data to earth problem in the robotics side how are people going to win of it? And are you even placing bets on that with people wearing gloves and cameras inside of manufacturing facilities?

Speaker A: Yes, we have companies operating in that space, one called Genesis, which is going to have a big demo soon, so I'll leave it at that. Pay attention to the Genesis big release, Mind Robotics, which is essentially Rivian's competitor to Tesla's Optimus, sort of a very capable general purpose robot that will be trained on all of Rivian's manufacturing capabilities and data, um, in addition to other robotic policies, and a company called the Botco, which is, um, Kyle Vogt from Cruise's household robot. So a series and then obviously our more verticalized AI companies, as I mentioned, Bedrock or Wave. So we are kind of heavily investing across the space. Nothing that I would describe as a humanoid per se, but, uh, certainly we're invested in general purpose Robotics. Um, the data girth is a really interesting challenge. Uh, there's a group of people who believe that you can use essentially video data in video language models to be able to produce a level of understanding and awareness that's sufficient for a robot to function generally in the world. I would not say that we subscribe to that mentality. If you do need other data, which I believe you do need the kind of data that's missing, or that companies like PI and skilled Sunday Robotics, uh, mundane, are like racing to try. And this hole that they're trying to fill is like proprioception data. So all of the haptic feedback and the physics layers and what it actually means to have data about how scenes in the world will evolve that aren't just based on visual pixelated data, but an actual understanding of the physics vectors and the mechanisms of touch and interaction with things. So that's the data set. These are the data sets that are being trained up and trying to get pulled into the models that are being made today. And it's hard and it's laborious. And actually one of the reasons why I go back to China and its advantage, one of the things that a lot of companies in China are investing in are these large data farms. And so what you would have imagined historically as a manufacturing line with a lot of lower paid workers kind of pushing and sticking buttons on a thing. These days, what are those people doing? They're like, like moving around in robotic suits, picking up a flower and putting it in a vase and like washing a dish. And it's kind of wild. These data farms actually exist and they are being to some extent sponsored, um, governmentally and also by other people who are investing in the space in China. So it's nuts. And this is a data dearth that we need. We need to close that data gap as well.

Speaker B: Yeah, I want to switch gears though, and get your sense on like the institutional LP perspective because people talk about the size of the software market versus the physical world and like, you know, robotics and you know, logistics, supply chain. There's a lot of money spent on physical things rather than just on software. And governments are trying to push people to, you know, spend money in their country. You're Canadian by background, so I hope we have still dual. Okay, you're dual, but I hope you have still some affinity towards helping, uh, the little brothers out. What would you advise, you know, the Canadian government or Canadian institutional investors on how they can get their dollars to work in this multi trillion dollar annual investment required to build out the next manufacturing industries here in Canada and the U.S. oh man.

Speaker A: You know, it's such a good question. It's something I'm thinking a lot about. I think Canada has such an advantage on all of the big moving pieces that matter today. So number one, talent. You know, we're in an insane market for talent. And when you, it's so funny, you go to any dinner in Silicon Valley, people, uh, introduce themselves like, like 30% of people are from Canada. And you're like, what the hell? And so there's just been decades of brain drain, to be honest. And it feels like there's some m. Incredible people up in Canada. A lot of incredible people have left Canada. And I think we stand at a moment in time when the Canadian government can, and I think is getting really aggressive with its immigration policies. I know a lot of Americans who are trying to leverage their like, Canadian ancestry to do this lost Canadian thing, which is amazing. So, you know, there's both like pullback. There's having. Not to be political about it, but I think Mark Carney's doing a great job leading the country. I think there's a level of pride around that that is attracting a lot of like historically left Canadians back. I think the sort of push on pulling talented people up into the country. So immigration and smart immigration policies and being really thoughtful about like hunting the best people. UBC and University of Toronto and McGill are all trying to pull great researchers up, up via, uh. I know my wife's an organic chemist and she was at University of Nevada Las Vegas, has moved her PhD to UBC. Right. Like this is happening, it is happening. As scientific fundings get drained in the United States and The Canadian universities are like, come up here, we have all the money, we have all the equipment, we have all the interest, like come now. And it's happening. So that's one way in which, like, you know, Canada can really win. And I see, I feel like, like they jumped on that quickly and it's exactly the right thing. Number two, abundant energy. So Canada has, you know, what is the fundamental currency of AI? It's electrons. Right. And so how can Canada leverage its energy abundance and its clean energy abundance to really, um, create all the incentives, the infrastructure incentives, work with the hyperscalers, work with the different players up and down the stack to say, build your data centers in Canada. We'll give you an insane deal. We're going to leverage our natural resources. And by that I don't just mean tar sands. I mean there's a lot of, there's water, there's all sorts of more renewable power that Canada has access to. The hydroelectric is incredible. And so, uh, we should be investing in nuclear. Let's leverage government incentive structures and infrastructure desire to come build out our energy infrastructure and as a function of that, attract probably the most critical strategic investments up north, which is data centers and AI. Right. So what are we doing there? I have a bunch of ideas on those policies. And then I think that we do have really great universities and universities that are still attracting talent from around the world. So how can we get the best in biotech? How can we get the best in quantum? It's such a, a peacemaker state that in some respect, if you're from, I won't name names, but if you're from one of the many countries around the world right now that sits in a hot, geopolitically contested moment, come to Canada, right? Do your fundamental research in Canada continue to move the frontier forward? I know the government's deeply interested in investing in the technology spinouts that can come from there. I know because I talked to them. So I think that it's just on every front, whether it's fundamental research across things like quantum and biotech, whether it's investing in the AI infrastructure that can cause like super accelerated low cost AI, and whether it's like immigration just to kind of pull the best and the brightest both back up and then over to the country, these are three key ways that the government can actually take action. And I see them doing that. Plus obviously the offtake that I mentioned, I know they're ramping up spending for defense. I think they should be ramping up spending and other smart strategic ways too. For, you know, better healthcare solutions and things like that. There also probably needs to be some thoughtful approach to reducing bureaucracy.

Speaker B: It sounds like you've thought about this, Aidan. Maybe we can get you on the advisory council for Carney if, uh, if you have some spare time while she's listening.

Speaker A: I'm, I'm open. I've obviously got a full time job.

Speaker B: A lot of people, uh, in the government listen to the podcast, so we'll make sure they get your phone number to call you. Eclipse just closed this $1.3 billion fund for physical AI and early stage incubation. Have you made investments in Canada in the past? And should founders in Canada be reaching out to you if they are building in the physical AI space?

Speaker A: Yes. We've got a great company called Augmenta in Toronto. We've got a great company called Jetson in Vancouver. I was up giving a talk at Mila in Montreal last year, uh, meeting with people about founding all sorts of great AI companies. I work with, talking to some folks at UBC about getting involved in some of their company creation. So there's um, a lot of active interest on our part to be investing in the best and the brightest up there, especially as it relates to deep tech and fundamental technologies.

Speaker B: Amazing. We, we got some feedback, uh, from some of your friends, like uh, Dan Preston, who said you're one of the best operator investors out there. So for any founders listening, you definitely want Aiden in your corner. Before we wrap things up, we always ask our guests for their fast favorites. So first off, your favorite podcast.

Speaker A: Clearly now it's Tank Talks. Not to be a kiss up. Okay.

Speaker B: Number two, fine.

Speaker A: I also really like Ezra Klein's podcast, but whatever.

Speaker B: I'm sure you listen to no Priors as well, which is a great one.

Speaker A: I do like no Priors. They're good.

Speaker B: How about your favorite newsletter or blog?

Speaker A: Uh, there's a TL Dr. One that I really like. Um, pretty good cutting edge tech news.

Speaker B: Awesome. Next is your favorite tech gadget. From a hardware sense, this must be the Apple Watch.

Speaker A: I hate the Apple Watch. I'm an Oura.

Speaker B: You're an OURA ring person. That's awesome.

Speaker A: I'm an Oura Ring person. I think actually my favorite tech gadget is. I'm very excited to see what happens with neuralink. Think there's a lot going on in the world of like physical plus digital as I've mentioned and like the human side of things is one of the last frontiers, one of the scariest frontiers. But I do believe in the technology's capability to change people's lives for the better. Um, so curious to see where that goes.

Speaker B: Absolutely. Next is your favorite new trend.

Speaker A: I like TikTok dances. I don't do them, but I get endless enjoyment out of watching people do them. I'm sure no one has seen that before.

Speaker B: That's hilarious.

Speaker A: Like who? Everyone. You know you like. People don't dance because they think it's not cool. Yet somehow we've had this like dance revolution and it became the only thing people did through covet. It just makes me laugh.

Speaker B: It's like, I agree. I sometimes tune into the Red Bull dance competitions. I find them very fun.

Speaker A: Some things in the human spirit die hard.

Speaker B: Absolutely. Next is your favorite book.

Speaker A: I love the biography of, ah, Jennifer Dudna, the woman who discovered crispr. That's probably my favorite book.

Speaker B: Awesome. And last but not least is your favorite life lesson.

Speaker A: Don't ever believe it's not possible because life has continued to tell me to swing for the fences and it is. So go for it.

Speaker B: Yes, you're definitely living proof of that. Thanks for joining us in the Tank today with Aiden Madigan Curtis at Eclipse Capital. Hey everyone, thanks for tuning in to another episode of Tank Talks. We hope you found today's conversation as insightful as we did. If you're enjoying the show, we've got three quick things to ask of you. First, hit that subscribe button on your favorite podcast platform so you never miss an episode, whether that's Apple Podcasts, Spotify, Google podcast or YouTube. Next. Follow us and stay up to date on upcoming episodes and behind the scenes content on social media with Twitter, LinkedIn and Instagram. And lastly, share the love. If you found value in today's episode, share with a friend or colleague who'd benefit too you. Your support helps us bring in more amazing guests and keeps the Tank Tops engine running. That's it for today. Until next time. Keep disrupting and innovating.