Atoms, bits, and the trillion-dollar manufacturing race

About the episode

For 40 years, the U.S. economy has been optimizing itself for the digital world while exporting its physical manufacturing capabilities. Now, the rise of AI is leaving a massive physical footprint in the form of data centers, power plants, and supply chains. 

In this first episode of Critical Capital, Crux CEO and host Alfred Johnson sits down with Aidan Madigan-Curtis, partner at Eclipse Ventures, to discuss why the next generation of trillion-dollar companies will be built where "atoms and bits" connect. They trace the industrial arc from the Gilded Age to the modern age of physical AI, exploring the geopolitical "dance" with China and the radical energy shifts required to power the future along the way.

And they ask: Who wins when the digital world runs headfirst into the physical one?

{{cta}}

Episode transcript

‍Aidan Madigan-Curtis: The vast majority of global GDP is still focused in the industries of making, empowering, and moving things. The United States has really set itself back on being anywhere close to leading. You know, we are lagging horrifically in our ability to play effectively in those sectors as innovation leaders.

And in order for us to become globally competitive, we really need to leapfrog, uh, when it comes to how we make and power and move things.

Alfred Johnson: For 40 years, the American economy has been optimizing itself for things you can't touch. Software, financial services, information. We celebrated companies that could scale without building anything physical, and we exported the capability to actually make things to whoever would do it cheap.

Now comes AI, which requires a massive physical footprint, and it's unclear if we can build fast enough to keep up.

So who wins when the digital world runs headfirst into the physical one?

Aidan Madigan-Curtis: It's a really interesting moment in time, perhaps a flashpoint.

Alfred Johnson: This is Critical Capital.

Alfred Johnson: Welcome. I'm Alfred Johnson, the CEO of Crux. This show is about the forces shaping energy and critical infrastructure, the markets, the policy dynamics, and the capital flows that most people see in pieces, but rarely see as a whole. In each episode, I sit down with someone who's at the intersection of all of it.

In our first episode, I sit down with Aidan Madigan-Curtis. Aidan is a partner at Eclipse Ventures, a venture firm that backs startups building physical systems. Eclipse just announced a $1.3 billion fund focused on physical AI. The firm's thesis, which is also Aidan's core belief, is that the next generation of trillion-dollar companies will be built where atoms and bits connect.

For roughly 40 years, the American economy has been optimizing itself for services. The destruction of US manufacturing isn't a surprise, but many people have been surprised by how effectively China has filled the gap. China isn't just good at building cheap consumer goods. It's built advanced capabilities in EVs, robotics, factories, high-speed rail, and critical minerals.

Now comes AI, which isn't just a software story, it's a profound physical story. The leader of the AI race won't just be determined by the most advanced models. It will depend on how quickly we can build data centers, power plants, and the supply chains that support them.

Aidan has been navigating the world of manufacturing and industrial systems for decades, and she's been traveling to China for 23 years. I have loved talking to Aidan over the years. She is a rare person who brings deep knowledge about history, geopolitics, technology, and investing. So today we're tracing the arc from the Gilded Age to the age of AI.

Aidan Madigan-Curtis, what a joy. Welcome to Critical Capital.

Aidan Madigan-Curtis: Should be fun.

Alfred Johnson: So here's the concept on this thing. We started it with the notion that we have all these tectonic forces that are shaping the way that we live in the world today. And those fall within these big categories. There's markets and economics, there's politics and policy, and there's tech and innovation. And the basic idea was that each of those domains is dramatically complex, right? Like just understanding any one of those domains is a lifetime of work. Yet there is this intersection of the three of those big domains that kind of guides the way that things happen in the real world. And so the concept here was how do we create conversations that match the most fun and interesting conversations that I've ever had in my life across the domains with people that know how to talk about all of those domains.

And as I was thinking about the people that I have treasured the opportunity to talk with over decades, you were at the absolute top of that list. And so it is, it's a joy to have you here as one of the first guests on the show. Uh, and I thought we would just start with the you of it all. Um, so you grew up in Penticton, in British Columbia. Uh, as you have described, on the side of a mountain, uh, there with no mailbox. It is right up the side of a mountain.

Tell me about that. And the thing that I'm really interested in is you've described yourself as this frontier woman in frontier tech, and so tell me how you became that at the, at the very basic level.

Aidan Madigan-Curtis: Sure. I grew up so far up the side of a mountain, in a pretty small town in the interior of British Columbia [00:05:00] that, two stories resonate. One, um, we lost a number of cats to cougars, unfortunately. So just to give, give you that visual. Lovely. I mean, it was awful, but it kinda gives a sense of just how far up there we were.

And then second, I had this guidance counselor who sort of like woke up one day and decided that I should, and I, uh, this is grade 12, mind you, like apply to go to Ivy League schools. And no one from this town had ever done that before. And I thought he was legitimately crazy.

Um, long story short, I ended up getting into a few of them actually, but none of the admissions letters actually made it to me because we just had this little PO box, 'cause we were on the side of a mountain. And, uh, I just remember it was like mid to late April, and I was in class. And the guidance counselor came to our biology classroom door and knocked and kind of pulled me outta class and I thought, what's going on? And they had this large manila envelope in their hand with the, with a Harvard stamp on it. And it was like, they don't normally say no in the big envelope, so did you get one of these? And I was [00:06:00] like, no.

Alfred Johnson: You'd never — you hadn't seen it. It was a total surprise.

Aidan Madigan-Curtis: No, it was a total surprise. It was super shocking. Um, so yeah, just to give you a picture of where I grew up. Not necessarily a remote place.

Alfred Johnson: Yeah. And so let's play with the frontier woman and frontier tech thing. And one of the things that I've admired about you is you have always had this foresight that is ahead of your time. So you left Harvard and you went to Bridgewater working on the macro economy with the CIO of Bridgewater, right?

As the global financial crisis was happening and global macro became everything. And then you could have had this path into financial services, but decided that you would go out to Stanford, uh, for an MBA and pivot into technology. But when everybody else was pivoting into going to be PMs at Instacart or whatever, you went to Apple and then [00:07:00] you were ahead of IoT before IoT was a thing.

And then you made the shift into being a venture capitalist in this very, uh, hard tech space that you currently operate. Tell us how you made those transitions across all of those different domains and what the underlying motivation and thought was that led you to make those choices.

Aidan Madigan-Curtis: Yeah, I appreciate you calling out each one of those transitions. 'Cause every single time, my friends, family, and loved ones looked at me and said, are you crazy? So, yeah, it wasn't easy. I remember leaving Bridgewater. Uh, I was surrounded by the smartest people in the world.

Uh, thinking about some of the hardest, most intractable problems, how do you win in global markets across, you know, so many different asset classes where the best in the business are only right, like, 55% of the time.

It was such a fascinating [00:08:00] cerebral problem and I just felt like, on the one hand, I wanted to get in the game. Uh, I had, I just had this like really visceral sense that I wanted to be one of the people trying to help shape it.

Not just sort of commenting and thinking about it. So that's what prompted my move out to the West Coast. It was just — out west. Um, I was fortunate enough to get to Stanford, which really in my head was a bit of a Trojan horse into the Valley. It's just really, you know, one of the core heartbeats of tech. So time there was amazing. I got to spend most of my time either hanging out, having global and macro conversations with you or across the street, banging my head against the wall in really hard engineering classes that I was in no way qualified for. But that was really kind of the transition in.

And then the Apple decision was another one of those moments where so many people were like, what are you doing? You know, you have this front row seat and this huge opportunity to go grab some software PM job at Facebook or something.

Like, go do that, right? This is the cool stuff [00:09:00] to do. And I felt, like, to Marc Andreessen's point, software had been eating the world, that there's this absolute Cambrian explosion of productivity and innovation and new ways of doing things.

But it was all trapped, like horrifically trapped behind these like rectangular prisms that are laptops or our phones. These very limited portals. So what I wanted to do was build more points of tangency between the incredible sort of productivity that comes from tech and the quote unquote real world.

And in my head at the time that looked like sensors that were in, you know, hard-to-reach places, generating data, beginning to bring typically offline assets and industries online. That became the Samsara bet. Before [Samsara]. And just as, you know, six years in, I had taken this crazy bet, it was going insanely well.

And that's again when I decided [00:10:00] to go kind of make a big shift and get into VC. And I just had fallen in love with the Eclipse team, a very small group of people who had been dedicated for the prior five, six years, right around the same amount of time I'd been at Samsara, to investing in hard tech in the physical world.

And I knew how much impact that could have, uh, 'cause I'd seen it firsthand. And so that was really my desire. It was to take all the scaling and the product-market-fit finding and help hundreds of other companies over time in really tough sectors, whether it's energy or transportation or logistics.

These really intractable places to go build generational breakthrough businesses that marry atoms and bits and really change the way these industries work.

Alfred Johnson: So, in this concept of being somebody who has been pretty good at seeing where the world is going before the world goes there, one thing I think I know about the way that you get there is to think about how the world has taken shape in the past. So you're a, [00:11:00] you're a student of history and you're a student of the ways that technological transformations have changed the way that we operate as a global economy and our political systems.

And I know that you think a lot about how we got to where we are today and what it tells us about where we're going in the future. You've written about the Gilded Age in particular as a moment of dramatic industrial transformation that cascaded in all these political and market factors.

What do we learn from that period of history more than a hundred years ago?

Aidan Madigan-Curtis: Okay, great. Um, I think it's right to point out that history may not repeat itself, but it does rhyme. And the Gilded Age is an interesting place to start because you've got the United States at a moment in time where certain innovations, whether they be regulatory, certainly technical, and also social, were all converging.

So one of my favorite stories is around Andrew Carnegie [00:12:00] and the Bessemer process and how actually understanding how to turn pig iron into steel at a 10 times cheaper rate, which wasn't his innovation by the way — he brought it from some scientists and industrialized it — it led to his ability to lay rail track at such a cheaper and more productive rate that he was able to sort of set off on the path that we, you know, now associate with him just being this industrial magnate. And at the same time, you had folks like J.P. Morgan doing a similar thing on the banking side. And so you had this combination of new tools on new types of projects, you know, new technology enabling orders of magnitude greater productivity in how we were building and transporting and shipping things around the country.

And that's really what led to just this, you know, incredible explosion of productivity and of growth. And I think that we're, you know, we saw another moment like [00:13:00] that in the World Wars era, in the '40s in particular, where the Allies were losing the war. And in the United States, there was the creation of a piece of technology called the heavy press.

And the heavy press was one of several technologies — I mean, not even to talk about the atomic bomb obviously — but the heavy press, somewhat less appreciated, very critical piece of technology that allowed us to, in a much faster and more efficient way, essentially stamp out the housings and the bodies of aircraft. And that allowed us to actually go create, in a — you see this playing out again now, you know, how do you create faster, cheaper, more [capable] systems? I mean, we've been trying to create faster, cheaper, more [capable] systems in warfare since the dawn of time, right? It was rocks to, you know — so, you know, just these historical moments. And not to mention the, you know, the most poignant form of competition there is, which is war, right?

[00:14:00] Civilizations actually fighting each other. Um, people who stood for very different ideologies, very different ways of life, believing that their way of life was at risk. And so it just spurs this incredible innovation cycle, um, 'cause people are quite literally fighting for their lives. And — Um, you know, at the same time you also had, uh, again, you know, fiscal policies and other regulations that were coming about, other ways of helping the country become more productive and more efficient, like FDR's New Deal. And, you know, the laying of major highways, which allowed for the movement of, you know, folks around the country, allowed for people to co-locate that shared thoughts and views or that wanted to work together on hard problems in much more efficient ways.

Alfred Johnson: Let's talk about that, 'cause the heavy press, interstate highways, the advent of nuclear energy, all of those things contribute to this period of remarkable industrial economic transformation where productivity over that period was going up pretty [00:15:00] dramatically. And then in 2010 or around there, it stopped. Why?

Aidan Madigan-Curtis: I think it's — that, um, and I think that in the early 2000s, really, actually since — I would argue sort of since the late '80s, I think you have to go back a little bit further.

The '70s, '80s, the country really transitions from being a builder economy, a manufacturing economy, to a country that was really innovating through more services and financial structures.

So, you know, also in the '90s you had NAFTA and the free trade agreements that began to make it more economic to manufacture in other places. That sort of started off a multi-decade hollowing out of the United States' core capability in manufacturing.

And you end up in 2010 where the real focus of the economy is software [00:16:00] and financial services, technology, and, you know, not hard things, not building. It wasn't, uh —

Alfred Johnson: Yeah. Well, so let's first go into the other pieces of the puzzle here, right? Because this also coincided with the rise of China. And in that period, in this period that we're talking about, you had a particular lens on it. So when we're talking about the 2010s, that was the period within which you were working at Apple and starting to create the Apple Watch and working on the processes that were inherently globalized to create a supply chain to produce a product like that. And so what did that experience teach you and show you about how the global supply chain functioned at that time, and how has it changed between then and now?

Aidan Madigan-Curtis: I've been going to China for 23 years at this point. For me, it actually started in the early 2000s, when I was in college. I was just so hungry to [00:17:00] get out and see the world and ended up working in some not-for-profit organizations that were doing work in China, largely focused on helping children who had birth defects get care. And I spent multiple summers in China. I took a semester off and lived in Shenzhen right as it was really evolving into a modern Chinese tier-one city, but it was at the beginning of that evolution. I just developed an incredible appreciation for the challenge and the wild opportunity that was China at the time. You couldn't walk around the city — I spent most of the time in Beijing back then — without feeling the insane pressure, the thundering pressure of the population.

There was so much of a hill climb that China needed to do in the '80s and '90s to move from what was really sort of a devastated economy post the Mao Zedong era. [00:18:00] and the Cultural Revolution into, you know, the China that we see today, which is a global superpower, right? So seeing that in the early 2000s was just mind-blowing, wrestling with these questions of how do you do that from a policy perspective? What are the effects economically speaking?

I had this lens of China from the early 2000s in the not-for-profit sector, looking at things from a socioeconomic lens. And then I had this, you know, maybe five, six years later, this more economic lens that I put on looking at China at the transition into the Xi Jinping era.

And looking at sort of the political economy of things in the early 2010s. And then fast forward four or five more years, I'm looking at China through the techno-economic lens, and I am really in-depth in their manufacturing sector, which obviously was such a critical part of their growth story from the 20 years prior.

I just got back from China like four days ago [00:19:00] where I was looking at not only where things have gone with their manufacturing capabilities, but also looking at the growth of AI and physical AI, robotics and autonomy in China, which is now becoming a really formidable player in the tech space as well.

Alfred Johnson: Yeah, so I think there's this really interesting dance that we've been in with China over this period that we've been talking about. Right. And in the period that you and I have both been working in and around the space, we lived through this moment of direct intimate connectivity of the two countries, right?

Where the supply chains were deeply interwoven and the financial markets were deeply interwoven. In the period that you were at Bridgewater, I was at Treasury, working in the part of the department that issued treasury bonds to the market. And at that time, China was by far the largest buyer of US treasury debt.

And that was a very [00:20:00] specific dance, right? US consumers would buy products from Chinese manufacturers that would create dollars in China that needed to be used, or the government was using to deflate the price of their currency, uh, or lower the price of their currency.

Aidan Madigan-Curtis: Keep the goods cheap.

‍Alfred Johnson: And that required investment back into US treasury bonds, which lowered the cost of interest on our debt.

And it was this very like symbiotic system that existed at the political level, the industrial level, the financial markets level. And so fast forward to today. You were just there. You've seen the level of innovation that's happening over there right now. And we're in this period where we're in a different form of dance. And what does that look like today from your perspective?

Aidan Madigan-Curtis: I think the average American has no idea how much value we've created for ourselves on the backs of the Chinese workers. And that interplay you just mentioned between [00:21:00] — it really created a moment, over, let's call it 15 years in the United States, where goods across the board were so much more accessible for so much less.

It's like kind of a tragedy. I think that that's a little bit lost, like just how much our economies have been tied and in many ways, like, mutually beneficial. While those dynamics were, I think, extremely beneficial for prices in the US, they also certainly took a toll on competitiveness of US manufacturing.

So actually maybe it is an interesting lead-in to how China affects US mentalities today. And it almost hearkens back to the conversation we were having about the innovation cycle that comes from global competition that runs so deep that it could almost be considered, you know, dangerously competitive. And the kind of innovation you get out of people when, you know, referring to the wartime scenario — I said when they're fighting for their lives, you know, I wouldn't say [00:22:00] we're exactly there, but we're certainly a lot closer to those dynamics.

I spent some time — I had a company called Pony AI in Shenzhen and it was really interesting. Like, they're sort of the Waymo of China. We spent time with the CFO and he's a brilliant, brilliant guy. Super sharp. I think he was educated over here actually. Perfect English. Worked at Baidu in San Francisco for many years, and I think they started Pony over here actually. And then they just determined that, in the current geopolitical climate, they were gonna have to pick a lane where they could be most competitive. And they chose to develop this technology in China, and they're doing a great job bringing it up and getting to scale.

Designing and building the hardware for autonomy, for robotaxi businesses, out in China and delivering that service in a handful of tier-one cities. And I think in a, at least break-even kind of way, which I don't know if any US company can [00:23:00] claim that right now.

So it's interesting. I think we're right back there in a new wave of tech, and maybe with some other geopolitical underpinnings that have, you know, brought us to the place where China is the other superpower. Right. Over the course of 40 years, like, now China is the other superpower. We've helped them build themselves.

They've helped us in a lot of ways too, and now we're competing with them as our near peer, potentially also global adversary, quote unquote. And yeah, I think a really interesting innovation cycle is coming outta that.

Alfred Johnson: So let's go there. Okay. So, looking at the historical examples, right? We have the Bessemer process and the heavy press and these things that have changed the shape of manufacturing and capabilities in industrial manufacturing processes. You have, in the prior century, the cascade of political effects of that, and even geopolitical tension that has come as the result of that. We now have AI, [00:24:00] which is arguably — I think probably definitively — as transformational as some of those technologies that you referenced. And we have a war in the Middle East and increasing tension in the global economy. And so, play out our current reality from the perspective of the history that you take and allow us to see the perspective that you gather on where that means we're going.

Aidan Madigan-Curtis: I just wanna double-click quickly into where China is and then I'll pull back up to the broader question. The country and its leading manufacturers have been compounding on their ability to automate their manufacturing lines, their homegrown capabilities to write software, design hardware, design jigs that like fully automate to the extent possible what it looks like to manufacture everything from an iPhone to [00:25:00] a car.

Right? And as a function of that, you know, they've been investing in new innovations. Like how do you manufacture batteries? What electrochemistry should you use? You know, they really brought LFP to the table in a meaningful way.

Now they've innovated into sodium ion, which is even cheaper and safer and more cost-effective and less dense, but, you know, incredibly good for certain use cases like grid-scale storage. And so what really did shock me was the progress of companies like BYD.

So you've got real global brands at this point that are emerging from China's investment over the last 20, 30 years in not only being a manufacturing base, but observing and educating its people overseas and having, you know, its citizens become important parts of large global corporations. Understanding the marketing game, understanding people management, understanding everything in between, such that [00:26:00] you end up with companies like Xiaomi, like BYD, you know, these really interesting companies that have actually become global leading innovators, right? It's not the me-too company of the past. It's the leading edge. And they've focused — I would say they've leveraged and they've gotten the most acceleration in areas where being able to tie hardware design and manufacturing capability at scale can be a critical edge.

So it's a lot of what you would just determine to be like physical world innovations.

And so that's what was different. You want to take a look at, you know, where we are right now — the climate in the United States, the capabilities in China, what's going on in Europe, what's going on in the Middle East. I mean, there's so many factors to talk about. On the one hand, we could be encouraged by how intertwined — I mean, it was only 20 years ago that the United States was at war with Iraq, [00:27:00] and the region felt war-torn and distant. And today, while once again, we're seeing military activity, sadly, in the region, we also just have these incredible ties. The relationship between the UAE and Saudi Arabia and their financial markets, as well as their commodity markets, to the rest of the developed world is astounding.

And the level of communication and interaction between that part of the world, Europe, the United States, Canada — you know, it's really interesting to see just how much more connected we all are. Right?

In some ways, globalization is the greatest deterrence to war. I think we're seeing that in the interactions between the United States and China. On the one hand, you've got increasing adversarial conversation, cyber espionage issues, all sorts of stuff, you know. And yet, we're taking a much more docile [00:28:00] tone relative to our relationship with Europe. Um, or perhaps, you know, recent activity that we've seen in other parts of the world.

I think because, you know, we saw this with critical minerals last year, the United States doesn't have the cards to really push because our economy would be so devastated if we truly cut ourselves off from certain other players. So yeah, I think right now we're standing at the precipice of a pretty insane shift to work and our overall capabilities.

'Cause not just AI, it's quantum. Potentially nuclear fusion. We have a few — and this isn't to be like a VC techno-optimist — we have a few things that are right at our fingertips. AI, which is hitting now. Physical AI, which is on its way. And quantum and sort of new ways of producing energy that are maybe five to seven years out, that are all radical in their impact on [00:29:00] the human condition.

And yet we're also sitting here geopolitically and sort of geo-socially in a world where we both are enjoying being interconnected on things like TikTok and other, you know, sort of social cultural forums, and yet also feeling, um, maybe further apart than we ever have. So, yeah, it's a really interesting moment in time, perhaps a flashpoint.

Alfred Johnson: Yeah. So let's talk about how AI enters the picture. Uh, and I'll frame the question this way. So, Eclipse has been at the forefront of investing in industrial hard companies, energy systems and software, and the connection of all of those things. You all, a couple days ago, announced another $1.3 billion that you're going to be investing in physical AI companies. Talk to me about how all of that perspective finds its way through into your [00:30:00] investment thesis on where the leading companies of the next era will be formed. And particularly, you've made this call that there will be a trillion-dollar manufacturing company. So talk about the interaction of AI and the physical world.

Aidan Madigan-Curtis: So, the vast majority of global GDP is still focused in the industries of making, empowering, and moving things. And the United States has really set itself back on being — I couldn't even say that — anywhere close to leading. You know, we are lagging horrifically in our ability to play effectively in those sectors as innovation leaders.

I mentioned earlier, companies like BYD and Xiaomi — like, they're really impressive companies that are vertically integrated and certainly compound on their ability to manipulate both atoms and bits. And the [00:31:00] United States' economy for the most part is almost wholly propped up by services.

And in order for us to become globally competitive once again, I think, you know, Eclipse's belief, my belief, is that we need to — you heard the term like leapfrog, in reference to countries that skip a generation and need to find a way to innovate radically to not only catch up but become leading.

We really need to leapfrog, uh, when it comes to how we make and power and move things. And that the companies that do that — companies like Tesla, companies like SpaceX, to name a few Elon Musk companies, um, companies like NextEra, you know, companies to some extent like Apple, right? Historically, even Nvidia, right?

These companies that transact in the atoms and the bits and do it in a way that is deeply innovative, that these will be the trillion-dollar companies. These are the trillion-dollar companies. Because the moat, when you innovate in physical sectors and you [00:32:00] go through the J-curves associated with them, which means standing up manufacturing plants, which means figuring out new ways to cut and create and scale real-world things, uh, which always costs money and takes time, more so than vibe coding something does —

Or the historical kind of fail fast and break things and reship them in the software world, you know, the moats that come from building and iterating and innovating in physical technology are just so large that you can really run away with an industry. And as a VC, you know, we're playing to own parts of generational businesses. The way we deliver value to our LPs is not to hit a double or a triple. It's to have companies that are worth a hundred times more or more than they were when we invested in them at an early stage. And it's not to say this is the only way of doing it, but really our [00:33:00] core thesis is that those best companies will be created in the physical space. And also that the people you want with you on that journey, making the capital allocation decisions, acting as board members, taking you through the toughest moments of your career, trying to make hard decisions about people and resourcing and dilution and all the pain that comes from innovating in the physical world — that you want people there who've been there, done that, who aren't gonna have their stomach turn at the first factory failure.

Because, yeah, those are the people who are gonna get you through the toughest times and also are gonna have the networks, whether it's fiscal or talent or partnerships that accelerate, you know, that can really radically accelerate these types of businesses. So —

Alfred Johnson: Yeah, so you've talked about moving atoms and creating bits. Both of those things require massive amounts of energy if you do them at scale. Talk to me about your perspective on [00:34:00] how the landscape around energy needs to shift given the present economic, political, geostrategic alignments.

Aidan Madigan-Curtis: It is really interesting to watch the effect of the closing of trade on energy prices. You know, you can run any amount of analysis and simulation to watch, real time, what it means to cut off access to energy, how dependent our economy and everyone's economy is on the cost of energy, and how directly that translates to people's pockets in a very short amount of time. It was just a painful, good reminder and interesting case study.

We're also watching this real time with the explosion of data center development. You know, the LLM-based transformers, while just radically transformational in their capabilities — I love Claude. I use the technology every day, um, to do what I would've [00:35:00] previously, even just months ago, thought to be completely impossible.

Um, they're not energy efficient. The way that both training and to some extent inference works, it's sort of heavy-handed relative to — the human brain works on like 20 watts, right? And not every human brain can do what Claude does these days. And — But the, just on an operation-to-operation basis, what takes us like a measure of just a few watts, can take AI kilowatts based on the model infrastructure and the silicon infrastructure. So number one would be like, we have to find a better way to do AI. This is like certainly a seven- to 10-year-out kind of investment thesis that I'm working on, but this is a field called neuromorphic chips and neuromorphic models, and there's different approaches to how we do AI, and I think that we're gonna need to adapt to make the energy equation make more sense. That said, those technologies are not anywhere [00:36:00] close to manifesting themselves in a real way in our daily lives. And we are seeing gigawatt-scale data centers pop up everywhere.

We have radical limitations to — you know, the energy infrastructure in the United States is incredibly ossified. You don't need to hear it from me, you see it everywhere. And therefore we're having a significant shock to the system when it comes to energy prices. I think the — you know, again, to the role of regulation, this concept that if a hyperscaler wants to build a data center, they can, permitting-wise — not in every municipality, that's a municipality-by-municipality thing — but the federal kind of mandate down is just make sure their prices don't change.

So you're gonna have to create more energy generation infrastructure, right? If you wanna actually go build the data center to create those AI outcomes. So I think that the level of demand that we have for today's version of AI, something to the tune of like — by orders of magnitude, radically outstrips our [00:37:00] energy generation and transmission capabilities.

We've got minimum three, sometimes five, sometimes longer lead times for the transformers — the sort of physical infrastructure it takes to build the substations, to create the infrastructure to allow for data centers to merely exist. There is a great way to power data centers with off-grid, behind-the-meter renewable energy that we've, you know, had a really exciting — with the Infrastructure Bill and the Inflation Reduction Act, you know, there's a lot of activity spurred in that space. We've really pulled back from that now as a country, which is just, I think, very tragic because, you know, solar and wind are the cheapest forms of energy generation. And if you co-locate large solar farms with battery and some smart controls to data centers — so, okay, so throw some gas turbines in there too — but, you know, you can actually create a durable firm load for these [00:38:00] data centers that does not drive up costs and does not create climate impacts like a massive series of gas turbines. So, you know, I'm hopeful that just the dollars and cents of that equation create more activity in the behind-the-meter space.

We've needed some regulatory adjustments so that that type of infrastructure doesn't get regulated and blocked as a utility, which FERC just did. But so I think that there's — it's like a very interesting moment and pinch point in the energy-meets-the-demand-of-AI equation where we're gonna see a radical crunch, and that crunch play itself out in the cost of energy unless we build a lot more generation and storage capabilities.

Alfred Johnson: So Aidan, let's finish where we started. So, frontier woman in frontier tech. Somebody who has always been good at seeing where things were going ahead of when we got there. What is something that you think people misunderstand [00:39:00] about the present reality? So you gave this example of people thinking you were crazy to go into industrial systems when you did, to go to Apple to change the way that the manufacturing processes worked there. And now you're in this reality — you're a VC with a significant amount of capital under management. Where do you think the next frontier really is here, and how should we collectively be orienting around it?

Aidan Madigan-Curtis: Yeah, I think that many themes in the world feel like vectors that shift from one side of the pendulum to the other. And knowing where you are in that curve — you know, are you on the way up, on the way out? Are you on the way down, on the way back in? — is where great investments are made. It's where good bets are to be had, and it's where you sound like a lunatic, when everyone else is like, what are you talking about?

And one of those [00:40:00] inflections that we might be living through right now is — so everyone's been talking about deglobalization, decoupling, and I just had this thought of how, you know, radically inefficient and unwanted it is for every country to be moving more and more towards protectionism.

Aidan Madigan-Curtis: And I wonder if we're actually, not unbeknownst to many of us, about to be on our way back in. And by that I mean, like I said, I think that in some ways we've never felt more far apart. But at the same time, there are so many ties that bind us, economically and culturally. I really wonder — I had this crazy thought about whether, if there were an entity that could safely assure a foreign country that the technology of another country could be safely operated, wouldn't have spyware in it, could be a functioning [00:41:00] part of their country too — how much more productive would we be?

'Cause it would create, for example, better competition between a Pony AI and a Waymo. You know, are we ever gonna see Pony AI on US shores? Probably not. It feels like in the environment that we live in right now, the concept of having autonomous, fast-moving machines running down the street with software determined by someone from another country — like, not gonna happen, right?

But if we knew for sure that it was just pure tech — it was functioning in a way that was observable and auditable — how great would it be to actually have a competitor to Waymo that was really giving them a run for their money, and how much better would that be for the consumer, and how much better would that even be for the technologists at Waymo, who would be pushed harder, right?

To keep innovating to stay at the top of their game. So I just wonder, if I were to make a [00:42:00] contrarian bet, that contrarian bet right now would actually be that we're about to head back in on the pendulum of swinging out to deglobalization. I think that there will be, over the course of the next few years, a major resurgence of desire to collaborate, of desire to experience and win from the benefits of, you know, truly globally competitive, positively competitive interactions.

How we get there, I'm not exactly sure. It's one of those situations where if I could explain to you exactly the, you know, thousand data points that give me this pinprick of, oh, maybe we're gonna go the other direction — well, that would be the black box, right? But that's where my head is and it's making me feel both optimistic and also really think through what those opportunities look like.

Alfred Johnson: Aidan, I think in a world that is awash in pessimistic takes, that was a [00:43:00] rare, rarely optimistic perspective that was also sober and realistic. So I appreciate your perspective. As always, it's fun to have you here. Uh, thanks for coming on Critical Capital.

Aidan Madigan-Curtis: Thanks for having me.

[THEME MUSIC]

Alfred Johnson: Aidan Madigan-Curtis is a partner at Eclipse Ventures. We covered a lot today — industrial history, China, supply chains, AI, energy, and the future of global competition. And this is just a preview of the cross-sectional conversations we've got planned for you in upcoming episodes.

So make sure to subscribe on Apple, Spotify, or any other podcast app. Critical Capital is a co-production of Crux and Latitude Studios. Our production team includes John Sheen, Jenna Herzog and Bailey, Steven Lacey, and Sean Marwan. It also includes Emily Hughes and the excellent Crux team. Crux is the capital platform for the clean economy.

I'm Alfred Johnson. Thanks for listening.

‍

‍Critical Capital is a co-production of Crux and Latitude Studios. Learn more about how Crux is financing the future of energy.