AI, Bottlenecks, and the Great Infrastructure Buildout: A Conversation with Daniel Dreyfus

In this episode of Forthlane Features, Vanessa Hui sits down with Daniel Dreyfus, Founder and CIO of Bornite Capital, a global equity long/short fund focused on cyclical equities, specifically in commodity-related sectors.

Drawing on more than two decades of experience at Goldman Sachs and 3G Capital, Daniel outlines his approach to investing across supply chains, with a focus on identifying bottlenecks.

He discusses the fragility of existing supply chains and the implications of what he sees as the largest infrastructure buildout in human history, spanning energy, materials, and labour. 

The conversation also explores the scale of the buildout required to support AI, particularly in power generation, data centres, and the materials needed to build them. Whether it's energy transition, deglobalization, remilitarization, or the AI buildout, Daniel shares his perspective on the significant alpha opportunities emerging from companies building the infrastructure to support where the world is going. 

WHAT TO LISTEN FOR

0:49 What led Daniel to launch Bornite Capital

5:02 How Daniel's investment process focuses on supply chains and bottlenecks

13:48 The multiple forces driving the largest infrastructure buildout in human history

30:25 The implications of rebuilding supply chains for inflation

34:17 A current high-conviction position in power generation

GUEST: DANIEL DREYFUS, FOUNDER AND CHIEF INVESTMENT OFFICER AT BORNITE CAPITAL

Website | Email | LinkedIn

CONNECT WITH VANESSA HUI, SENIOR CLIENT ADVISOR, FORTHLANE PARTNERS

Website | LinkedIn 

Transcript

Vanessa Hui (00:06):

Welcome to another episode of Forthlane Features, a podcast where we dive into the world of global wealth and asset management. I'm Vanessa Hui, senior client advisor at Forthlane Partners. And today I'm joined by Dan Dreyfus, founder and CIO of Bornite Capital, a fundamental global equity long short fund with a core focus in cyclical equities, specifically in commodity related sectors. Bornite is also one of the specialist managers in Forthlane's absolute return strategy, so we've had the privilege of following Bornite closely for over five years now. Dan, thank you so much for joining me today.

Daniel Dreyfus (00:47):

Vanessa, it's great to be here.

Vanessa Hui (00:49):

So before we dive in, I'd love to start with you. Dan, you spent over a decade as a portfolio manager at Goldman Sachs, then ran a commodity fund at 3G Capital, and then decided to venture out on your own to start Bornite with 60 million of your own capital. What was the catalyst for this decision and what do you think drove you to say, "You know what? This is the opportunity I need to go for.

Daniel Dreyfus (01:17):

" So my game plan was after spending 13 years at Goldman in their flagship long short equity hedge fund. That was when I was going to go out and start my own firm. However, on the way I got diverted because my friends and at the time, people who I just held in immensely high regard as both people and as investors. These are the folks from 3G, Alex Bering, George Paolo Leman. We were talking and they proposed instead of starting my own firm, why don't I come and join their firm and start a fund under the 3G umbrella? And to me, it was just a really easy decision. There is no other group of people that I would've said yes to, but for them, it was a no-brainer because in my mind, after watching what they had done over their careers and the value they created and the value that I believed they were going to create, I wanted to watch how they were going to do it.

(02:23):

I wanted a front row seat to say, how do these guys do what they do and create the value that they create? How do they operate? I wanted to see under the hood. And so in many ways, it would be like, think of leaving your 13-year career at Goldman to start your own firm, but then along the way, Warren Buffett calls you up and says, "Hey, Vanessa,

(02:46):

Why don't you come and sit beside me at Berkshire Hathaway and run some money and do it under the Berkshire Hathaway umbrella and sit neck to neck with me? "That's a decision that doesn't even require any thought.

Vanessa Hui (02:59):

And

Daniel Dreyfus (02:59):

Probably the best professional decision I ever made because working with them exceeded every expectation that I had in terms of what I could learn from them, how to run a business, how to analyze a business. They look at things through obviously a private equity lens. I'm in the public markets. I was just a sponge for six years. And after six years plus the 13 years that I did at Goldman, I was knocking on year 20 of my career. And I said to myself, "If at year 20 I don't go and start my own firm, then I never will." And so I made the difficult decision to approach them and ask them if I could start my own firm. And they were just so incredibly supportive. They loved the idea. A number of those partners were day one investors and continued to be investors today. We compounded their capital at a really great rate of return and they're thrilled and I'm thrilled and everything just worked out so well.

(04:11):

I really count my lucky stars that this was the path that I took.

Vanessa Hui (04:16):

That's incredible, Dan. And you know what? I've known you for many years now. And the fact that a question about your journey, you just spent the most of your time talking about how privileged you were to work with others really says a lot about you. And I don't think it's a coincidence. As we said earlier, good people are magnets for other good people. So congratulations. And I'm so glad that you made that decision. So Bornite's investment philosophy is described succinctly as understanding where the world is going and investing in what we're going to need to get there. I mean, that sounds intuitive, but in practice, how do you turn a big picture view of the world into a concentrated portfolio of just 10 to 20 companies?

Daniel Dreyfus (05:02):

Well, first of all, I have a very strong view that if you don't get the big picture right, nothing else matters. You have to fish where the fish are in the lake.

Vanessa Hui (05:12):

You

Daniel Dreyfus (05:12):

Be the greatest fishermen on planet earth, but if you're fishing in an area that has no fish, then it doesn't matter. You have to get the big picture right. And what we mean when we say we try to understand where the world is going and figure out what we're going to need to get there. And then step three of that process is really try to understand which companies are best positioned to get us there. What I really mean by that is we're infrastructure and commodity investors. And my whole life, I've been studying supply chains my whole life from where things come out of the ground to where we consume things, the entire chain, womb to tome, cradle to grave, germ to sperm, as some would say. And what I've learned is that if you can identify where the pinch point is in any supply chain, that area of shortage that slows down that production process, that is where all the excess alpha lies, where all the excess returns can be generated.

(06:22):

And so we look at big supply chains that are really going through demand shocks as our starting point. And let me give you a few examples. So for example, the aerospace supply chain, if you just add up Boeing and Airbus's backlog, it's well over a trillion dollars over the next 10 years. Now in that supply chain, there's all sorts of bottlenecks. And if you can buy the companies that are going to help alleviate these bottlenecks, and if these companies have big barriers to entry, because as you know, it gets very challenging to get a part certified on a Boeing or an Airbus plane, so there's big boats on these companies. And if you can identify the company that is going to alleviate a bottleneck, there's huge excess returns to be made. I mean, other areas, the transmission and distribution infrastructure, the electric grid, we're going to be spending well over a trillion dollars modernizing, upgrading, hardening the US electrical grid over the next 10 years.

(07:21):

I mean, we haven't seen any investment in the electrical grid since post World War II. It's a complete mess. You know that part of the grid where Paradise, California caught on fire and burned down half of Napa a while back. When they did the forensics on that fire, they found out that that part of the grid was over 106 years old. Did you know that there's parts of the grid that are over 106 years old and are causing these wildfires in California?

(07:50):

So all of this needs to be modernized. Forget about AI, which is going to be turbocharging that, but there's going to be huge bottlenecks in that part of the supply chain and we identify those, the generation infrastructure, power generation. Just in one little part of the market of the US called the PJM region, which is Pennsylvania, Jersey, Maryland region, the regulator is forecasting that we're going to need over a hundred gigawatts of net new generation, power generation supply over the next 10 years. Now to put that in context, that's more power generation than all of Japan has today,

(08:26):

And I bet they're lowballing it. Where are we going to do that? If Thomas Alva Edison was alive today, he'd be rolling over in his grave because he's going to say, "How are we going to do this? " And there is going to be bottlenecks in that supply chain and we can identify the companies that can alleviate those bottlenecks. Then there's the remilitarization. Did you know that in the Ukraine conflict, which has been going on now for four or five years, we have used more explosives than we've done than we've used in all of World War II. And we've depleted our ammunition stockpiles globally in Europe and in the US. And that's why the military budget for rearmament is going from a trillion to a trillion and a half dollars per year. And there's going to be bottlenecks in that supply chain. And then the big one in my mind, actually, this is the big one, is the critical materials supply chain.

(09:23):

Back in April of 2025, at this time last year, China cut off its exports of rare earths to the US. And our Czechs told us that we were within days, literally days of the Ford Motor Company's productions lines shutting down. The entire Ford Motor Company, McDonald Douglas, same thing. And what that made the US government realize and the Department of Defense realize is that China has so much leverage on us now because they control all of the critical minerals. While we were busy during the last 20 years de- industrializing and tearing down all our factories and moving into China, they were busy building up control of the entire critical minerals supply chain. And now, whether you want the world to go green with electric cars and solar, whether you want rearmament, whether you want data centers, whether you want AI, whether you want a modern grid, all of this is going to require these critical metals that China has an absolute grip on.

(10:35):

And so there's just tons of opportunities in that supply chain. And then the final supply chain that we love to talk about is obviously AI and data centers. That's the big one. I mean, if T&D infrastructure is a trillion dollars over the next 10 years, if transmission is a trillion dollars over the next 10 years, if aerospace, Boeing and Airbus backlog is a trillion dollars over the next 10 years, well, AI data centers are 700 billion this year, but easily going to a trillion dollars per year,

(11:04):

Per year over the next few years. And these numbers are just mind-boggling. This is all hard assets, infrastructure, critical minerals that go into building this out, and it's all happening at the same time. So look, figure out where the world is going. It's kind of easy, right? Data centers, electricity. I mean, if you believe that AI is going to be the core driver of productivity, the core driver of scientific breakthrough, the core driver of military advancement, then I think it's fair to say that we're going to be measuring human progress going forward by how much power a country generates because AI tokens are just a derivative of energy, right? You put oil through a refinery and you get jet fuel on the other side. With AI, you put electricity through a refinery, which is a data center, and you get these intelligence tokens on the other side.

(12:01):

And in order to meet America's technological ambitions, we have just an incredible amount of work to do. 20 years ago in 2005, we used to produce twice as much electricity as China did. Fast forward to today, China produces three times the power that we produce. So the amount of catch-up that we have to do, we have to start getting to work today, we have to go at hyperspeed, and this is a cycle across the board that's going to last a minimum of 10 years, a base case of 20 years with an upside case of three decades in duration, and we're just at the beginning of this cycle. So bigger picture, try to figure out where the world is going, figure out what we need to get there, and then find out the best companies that own the pinch point in the supply chain to invest in on a micro basis.

(12:57):

That's our process.

Vanessa Hui (12:59):

Wow. There is so much to unpack there. And honestly, one of the reasons I was so excited to speak to you is you have such a powerful way of distilling incredibly complex multilayer themes into something that we all understand, which is the supply chain. So thank you for that. And honestly, one of the things I find most fascinating about your thesis is it's not just one theme. It's AI, it's energy transition, deglobalization, countries reshoring, manufacturing, all happening at the same time. So is that convergence of all those themes, what makes this opportunity so compelling or does it make anything fragile if one of those drivers slow down?

Daniel Dreyfus (13:48):

Our supply chain is so fragile right now. If you think about it and you go back and look what happened over the last 20 years, really from the period of starting in the year 2000, all the way really through 2025, so the last 25 years, if you look at the US economy, it was almost this too good to be true, miraculous two decade plus period where the US created so much value in comprehensible amounts of market cap, huge amounts of growth without spending a dime of capital. And that is very, very unusual, right? The holy grail for creating value is to grow revenues, to grow free cash flow, but not have to spend any capital to do it. Just by definition, if you have two businesses both growing top line at 20% with the same margins and the same cost structure, the business that grows 20% that doesn't have to spend a dime of capital to grow 20% is obviously more valuable than the same business that grows 20% where they have to spend all their capital to grow that 20%.

(14:54):

And the difference is the one that doesn't have to spend the capital can either give that capital back to you while growing or they can reinvest that capital to grow even faster than 20%. And if you think about what the US did over the past 25 years, the value that was created from search engines like Google, social media platforms like Meta, software as a service, enterprise software, streaming platforms, ride sharing platforms, food delivery platforms, all of these businesses created obscene amounts of market cap, obscene amounts of value. And the commonality amongst each of these companies was they did it with basically spending no capital. I still remember the day that Facebook bought WhatsApp for $30 billion. I think they had 13 people in the whole company. And that's the definition of capital light growth. You don't even need people. You just create a network effect and create billions or hundreds of billions of dollars of value from a network effect.

(15:59):

Now, at the same time, we were creating all this value in what I call the internet economy. We were deindustrializing in the US and tearing down all our industrial factories and moving into China because the argument was if you take a glass half full and optimistic view of it that they could do things faster, cheaper and better. But I think the more pessimistic view is it was just too difficult to get permits and the regulation in the US was just too difficult because the environmental concerns in China, it was just much easier to get things done. So you had a much easier regulatory backdrop and you had a much lower labor cost. And we absolutely destroyed our industrial base because our manufacturing base, because everybody was just pressured to take whatever they had here, tear it down, pay out the severance to the people, move it to China.

(16:57):

This created a lot of problems in the economy, right? The middle of the country, the manufacturing base, the salt of the earth, the core of the country. These people lost their jobs. It was difficult to get retrained. They created this gigantic wealth disparity between the coastal states where these tech companies and finance companies that were financing them were creating all the value and the middle of the country was getting hollowed out and it created fentanyl, it created divisiveness. It was a really difficult time for that part of the country. And now what we realized, our economy was tested first time was in COVID. Then we got tested again in the Russia-Ukraine conflict. Then we got tested again when the tariffs hit. Now we're getting tested again with the Iranian conflict. And what we're seeing is each of these moments, which seem to happen every February now, there's some big geopolitical flareup.

(17:57):

Each of these moments, prices for everything that we buy start to spike and they don't come down. And why do they spike? They spike because our supply chain is so damn fragile. We can't get things. We can't make things here. We can't alleviate these price hikes. In some cases, we have outright shortages. I mean, it's wild to me that we still source 60 or 70% of our penicillin from China. I'm not saying China's bad. I love China, but that doesn't make any sense to me. It makes no sense to me that we source all our critical raw materials from China. That makes no sense to me. And so the supply chain here is so fragile that there is absolutely a panic right now at every layer of the government when you say the world words critical raw materials or life sciences facilities or semiconductor fabs.

(18:51):

I mean, to me, it makes no sense that we have the nucleus of semiconductor fabrication in Taiwan. Why don't we have it here? And the answer is we're going to build it, but that takes 20 years. And so look, this is all coming together at the same time. It just so happened that we have to rebuild this completely broken supply chain and rebuild it here in the US. And it just so happened that this is coming at the exact same time that AI has been introduced to center stage and this is the most infrastructure intensive build out that the world has ever seen.

(19:26):

And this just so happens that it's coming at the same time that we're fighting a few wars and we have to go and remilitarize and restock all our ammunition. And so to me, this is a moment in time where I believe that if you look out over the next 20 years, if the last 20 years, all the alpha came from big tech because they could create so much value without spending any capital. I think the next 20 years, all the alpha is going to come from the companies that are going to be effectively helping build this infrastructure because if you think about it, they're going to be taking in all the cash from the big tech companies. They're going to be taking in all the cash from the government military spending and reshoring. They're going to be taking in all the cash from private companies that are reshoring, and the cash is going to flow into their treasuries and out of the big tech companies treasuries.

(20:19):

And so if you follow the cash, my belief is that the infrastructure related companies and the critical minerals related companies are going to take over the market leadership. They already have, by the way, and that's going to continue for 10 or 20 years.

Vanessa Hui (20:34):

It's so interesting how you frame the AI story as a physical infrastructure story. I think so many investors think of AI as purely kind of a software story. So can you just elaborate on what actually needs to be built out to support AI at scale and where are the constraint pain points?

Daniel Dreyfus (20:54):

So there's just a tremendous amount of constraint in the entire supply chain in AI. And this is why this just feeds right into the core thesis of what we do, figure out where the world is going, what we're going to need to get there and where the supply constraints are within the supply chain. I mean, you can start it right out with the minerals that we need to build AI. AI requires a lot of electricity. The cheapest and most efficient way to conduct electricity is over copper wiring. When we study the various architectures of data centers as they mature, we started out with the hopper data centers, now we're in the Blackwell architecture, then we're going to move to the Vera Rubin architecture. Each next generation architecture of data centers becomes more efficient at getting more power effectively per unit of data center capacity. And what that means is just the amount of electricity and power that has to go into the data center just explodes.

(22:02):

And copper is really simple. If you put a little bit of power into a data center, you can use a small amount of copper wiring. If you have to put a lot, you have to use a lot of copper wiring. It's literally that simple.

(22:14):

And our analysis shows that with the Blackwell architecture, we need 50,000 tons of copper per gigawatt of data center capacity, and we're building 15 gigawatts globally per year, and that number is growing every day. And at 50,000 tons, just to contextualize that on 15 gigawatts, that's 750,000 tons of copper that we need. Before AI, the whole world was consuming 750,000 tons of copper, and we just added another 750,000 tons of copper. And this is coming at a moment where we haven't commissioned a major new greenfield copper mine that is bigger than 250,000 tons in 10 years. The only thing we're seeing the copper industry doing is M&A. And the reason is, is because the copper price isn't high enough. It's nowhere close to the level that would incentivize them to go and build new mines. So they just say, "Okay, we'll buy back stock and we'll do M&A." So the copper price has to double, and that's no big deal.

(23:16):

I've seen molybdenum go from a dollar a pound to $30 a pound. So the copper price doubling, I mean, that's no big deal. So there's things like copper. There's things like silver. We're going to have to use a lot of solar to build out all of this power generation that we're going to need because there's not enough slots for gas turbines. They're all sold out. Vernova and Siemens energy is all sold out into the 2030s. And so solar's going to have to be a bandaid and solar consumes copious amounts of silver. And China stopped exporting silver, by the way. That's a big deal. If we want to build semifabs here and build out solar capacity, we're going to have a silver crunch. And so you're going to need a lot of silver. Obviously, a lot of uranium, you need a lot of natural gas. So that's the first layer, the natural resources layer.

(24:04):

Then you go into the middle of the supply chain and what goes into the data center. So that would be memory, HBM, NAND, the fiber optics switches, fiber optic cable itself. There's just all of these industries which haven't really been great industries for the last 25 years are facing this overnight demand shock. And what's really interesting when obviously power is a huge, huge bottleneck in the AI ecosystem, we have to build a lot of power generation. And what's really interesting is the power companies, the power producers, just remember the days when they got so badly burned in the post-deregulation era, in the Enron era where the power companies went and built huge amounts of supply on speculation that demand was going to come and then the demand never came and the whole industry went bankrupt. Those memories are fresh. They remember the carcasses of the CEOs that were sprawled out on the highway that got fired from doing that overbuild.

(25:13):

And they're saying, "I'm not going to do that. I see this AI demand, but I'm not going to build into it. The only way I'm going to build into it is if the data centers enter into a 20-year contract with me to buy my power at a premium to where power prices are on the forward curve." And the most famous example of this was there's a famous nuclear reactor that we all have heard of as Three Mile Island, where in 1978, there was a nuclear meltdown and Three Mile Island, everybody had to evacuate. And that was the nuclear meltdown that gave Nuclear a very bad name and shut down the industry for a few decades effectively.

(25:52):

And about 18 months ago, Microsoft knocked on the door of the company that owned Three Mile Island and said, "Hey, can you guys restart this nuclear reactor?" Microsoft, which prides himself on their ESG credentials, "Can you restart this nuclear reactor?" And the company says, "Well, we got to enter into a long-term take or pay. We're not going to build this on spec and we're not going to restart this on spec." And Microsoft said, "Okay." They agreed to doing a 20 year taker pay contract at $100 a megawatt hour to put that number in context. At the time, the forward curve was trading at $50 a megawatt hour. So they paid two X the forward curve. That would be like Delta Airlines or United going to Exxon and saying, "The price of oil today is $90 a barrel, but we're so worried about future supply that we want to lock up the next 20 years of your supply for us and we're willing to pay $180 a barrel for oil when it's trading at 90." I mean, that's how desperate the data centers are to get hold of this power.

(26:57):

So the shortages are everywhere. We're drinking out of a fire hose in terms of figuring out where to deploy capital into shortages. And it's like whack-a-mole. I mean, we almost find a new shortage every month that's investible. This is the golden age.

Vanessa Hui (27:13):

Wow. So on the topic of big tech companies, you mentioned earlier that they are heading towards a trillion dollars of spend on building out infrastructure per year. At what point does this massive investment move from a growth driver to a margin headwind for these companies?

Daniel Dreyfus (27:34):

I think it's going to continue to be a growth driver for a really long time. We still haven't seen the physical AI layer in action yet. This is the robotics layer where there's just going to be a huge amount of productivity unlock. That's going to be the big one. Chatbots I think are really threatening entry level white collar labor. It's ironic that the blue collar or craft labor, I call it, craft labor that really saw their livelihoods turned upside down when we deindustrialized into 2000s. And while that was happening, the white collar labor was crushing it. Right now, it's the opposite. This craft labor is in such high demand and they name their price. They go on business trips like I used to when I was at Goldman. They say, "Oh, you're an electrician. Can we fly you to this part of the country? We'll put you up in a hotel and you can expense all your meals." And same way we used to do it when we were in banking.

(28:34):

And now they're getting this while they're getting put to work and doing really well while at the same time they're creating the infrastructure that is going to effectively make a huge swath of white collar labor completely obsolete. So the tables have turned. Instead of tearing down factories and making craft labor obsolete, we're building factories to make white collar labor obsolete. So I think it's going to be margin enhancing if you own capital, if you have a company, because you're going to benefit hugely from that productivity. And I think the more you invest in AI, the more productive you're going to be and the higher your margins are going to be. And we haven't seen anything yet because we haven't seen the robots, we haven't seen the physical layer.

(29:16):

So that's going to continue. And so if you believe that not everything is a win-win-win, I think really the part that I would be most concerned about, much more so than margins contracting, because I think margins are going to expand, is what's going to happen to more entry level white collar labor that can easily be replaced or enhanced by agents. Maybe you still hire them, but you don't need to hire as many because now they have these copilots and assistants that effectively give them the productivity of three people. So I think we're going to have to reinvent ourselves. We always do, but it's always a very choppy process. And so I think that is where the concern is if there's something that you have to watch out for.

Vanessa Hui (29:59):

Right. And so when we think about deglobalization, we think about countries reshoring manufacturing, the fact that we're in one of the largest infrastructure buildouts in human history through AI and just the fact that we've been neglecting building out so many key infrastructure supply chains, what's the readthrough for inflation over the long run?

Daniel Dreyfus (30:25):

So in the very near and medium term, balkanizing our supply chains and building everything onshore here, I mean, the whole reason we de- industrialized was to make things cheaper, faster. And so by definition, if we're rebuilding here, then things are going to be more expensive. I mean, the reason we don't have a rare earth industry here is because our resources of rare earths are sort of marginal. They're not great. And building out the rare earth's processing capacity is just cheaper to do it in China, but we have to do it here. And so when you look at these deals that the rare earths companies are doing with the government, similar to what the data centers are doing with the power producers, the government's knocking ... I've never seen this, by the way, in 25 years of investing in mining stocks where the government goes to some junior resource company, knocks on their door and says, "Hey, I know you've been waiting on a permit for the last 25 years.

(31:21):

Happy birthday. Here's your congratulations. Here's your permit." And they say, thanks. And they say, "What's this attached to the permit?" And they look at it and they say, "Oh, this is a check." The gross government is, "Yes, this is a check to help you get the mine up and running. And the US government now is going to be an equity shareholder in your company." And then they see another piece of paper attached to the permit and say, "What's this? " And then the government says, "This is a take or pay agreement where we guarantee you a minimum price that delivers a really attractive IRR that we're helping you fund for the build out and giving you the permit, and you can keep all the upside if the commodity trades over the minimum price." This is a Vuja day moment, being the overwhelming feeling that this has never happened before.

(31:59):

And the reality is that when you do it like this, the trade off is we build a stronger and more resilient supply chain, but the trade off, the trade off is everything is going to cost more because it just costs more to build things here. But that's okay. Who said that inflation has to grow at 2%? Who said that was the right number? Wouldn't you rather have inflation grow at 3% and have a really resilient supply chain where every time we have one of these geopolitical flareups, it doesn't feel like the world is coming to an end because we have our own supply chain and the trade off is have inflation grow by another point. Why is 2% the right number? Why isn't 3% the right number, but with a really resilient supply chain? So yes, we're absolutely going to have inflation that is going to grow at a higher rate than it did in the past.

(32:51):

I mean, don't forget, as we spoke about, we're coming out of a historically unprecedented period where we created so much growth of no capital. And if you don't invest capital, you're not creating inflation. So we're going from one extreme to another. So get used to high inflation, but that's fine. That's okay. Nobody should be upset about that and you have to adapt because this is something that is absolutely mandatory that we have to do and that's the trade off. Now, I think over the longer term, as we build the physical AI layer, that is going to be a natural suppressant to inflation because we're going to have an explosion of productivity and that's going to help keep inflation from really running off on us. So I'm very optimistic when I look out over the next three years, five years, 10 years. I think it's okay to have a little bit higher inflation.

(33:45):

And by the way, when you have higher inflation, guess what? Asset class performs the best. Infrastructure, hard assets, commodities, and that's what we do. So I'm okay with that cheap.

Vanessa Hui (33:57):

And that's what we do at Forthlane as well. So let's zoom into Bornite's holdings. I could honestly spend a whole day speaking about the macro with you, Dan, but I'd love for you to walk us through one of your highest conviction positions today, and specifically, what do you think the market is getting wrong about it?

Daniel Dreyfus (34:17):

So talent energy is our biggest position now. We've held the stock since $40 a share. Today it's about $340 a share.

Vanessa Hui (34:27):

Wow.

Daniel Dreyfus (34:28):

But we think the upside here is just incredibly substantial. Talent Energy is a power producer. They've got about 12 gigawatts of capacity, two gigawatts of nuclear capacity and 10 gigawatts of natural gas capacity. And of their two gigawatts, they signed a deal with Amazon to sell that power. Again, 20-year taker pay agreement, like we spoke about at a big premium price to where the futures curve is trading. And as that Amazon data center ramps up, they're going to be selling power at a much higher power price into that data center. So there's a natural earnings growth progression just from that. You just have to wait till they build this Amazon data center and they start selling the power into that data center at the premium price and you just get a natural path of earnings growth. And so when we look at Talend at $340 a share, if they just sit there and literally do nothing, the management team just says, "We're just going to watch our assets and make sure they run properly, and really that's all we do.

(35:36):

" Then just arithmetically, the earnings power of this company by 2029, which is when the Amazon data center should be largely built, is $50 a share. And we think a business like this, high quality assets, good secular growth potential, which we'll get into in a second, should trade at about 15 times that $50. So that gets you to $750 over, call it, three and change years versus 340 today. So a little more than a double. So that's pretty good for a status quo case, but they're not going to sit still. They're going to continue entering into above market power contracts to sell power under 20 year taker pay agreements to hyperscalers and data centers, and that's going to be earnings accretive for them over and above the $50 a share that I mentioned in the status quo case. But also, I think the power prices have to go up.

(36:33):

Right now, the reason you're not seeing a huge rush to build power generation in this country is that on the futures curve, power prices are still too low to generate a good rate of return to build any kind of merchant power on spec, on speculation, where you just build it and say, "I hope demand comes." That you can't justify it with the cost inflation that we've seen to build a power plant and the labor and the materials at $75 a megawatt hour in the PJM, which is where Talon operates. PJM again is Pennsylvania, Jersey, Maryland area. The power price is just too low. We think it probably has to go to a 90 bucks or a hundred bucks from 75. And that won't be such a huge deal for the electricity bills, the utility bills for people because power generation is only 20% of the cost of what the end consumer pays.

(37:23):

So if we go from 75 to 90, let's say that's a 20% rise in the power price. Well, 20% rise in the power price, but power's only 20% of the utility bill. It's a 4% rise over the next four years. That's very digestible. And so at $90 a megawatt hour, the earnings power goes from 50 up to $70 a share. Now, you put the 15 multiple that I spoke about on the 70 and you have well over a thousand dollars a share in three and change years. And the stock today is at 340, as I said, so you 3X your money. But then I think it gets even better than that because there is such a need to build power. As I said earlier, in the PJM, the regulator themselves say that we need to build 100 gigawatts over the next 10 years, and Talon is a core player, one of the biggest power producers in the PJM.

(38:22):

And so out of that hundred gigawatts, now some of it's going to be solar, some of it's going to be nuclear, it's not necessarily all gas, but we think that Tallon's going to probably be able to build at least five gigawatts of gas capacity to help fill that 100 gigawatt gap. And if they do that over five years, they can self-fund it with the cash they generate from their existing generation capacity, so they don't need to take on any debt, they don't need to issue any equity. Those five gigawatts would add another $35 a share of earnings power to the $70 that we had. So now you're looking at 105, 110 bucks a share of earnings power. Again, the stock's at 340 today. Now to build that out, it's going to take a little bit of time, so that's not in three or four years.

(39:08):

Maybe that's in six or seven years. But my view is this is a classic textbook set and forget. Buy it today at 340, take it off your screen, wake up early next decade, 2033, call it. You'll see it at 15 or $1,600 a share, and then you can take your profits and buy something nice. But that's probably our most exciting idea at the moment.

Vanessa Hui (39:37):

Amazing. So Dan, it's 2033. We've woken up. It's early next decade, and the entire infrastructure thesis didn't play out the way you expect. What's the most likely reason you were wrong?

Daniel Dreyfus (39:51):

Well, early next decade, six or seven years, even 20 years, in geological time, that's like a nanosecond from now. Like you blink and we're going to be there in geological time. I mean, we're so used to everything being done in internet time. But what I mean by that is you press your button on your phone and an Uber shows up to take you to the airport or you press a button on your phone and your Chipotle gets delivered to you and it happens instantaneously. That's internet time. Infrastructure, like we're talking about, gets built in geological time and 2033 is literally like tomorrow morning.

(40:26):

So my fear is we just don't have the supply chain to build this and we're going to fall behind. I think the good news for the US is we are by far and away the leader in GPUs and semiconductors, but we really lack in power and infrastructure. And China is excelling in power and infrastructure, but lacks in the GPUs. And so it's an interesting race that we have because they're running with one hand tied behind their back because of the chip export restrictions, and we're running with one hand tied behind our back because we don't have enough infrastructure and power to get there. So it'll be interesting to see how it plays out. But when I look at the landscape today, I really scratch my head. I think probably the biggest shortage that we're going to start to see emerging by far, and it's a structural shortage.

(41:27):

When we have a memory shortage in HBM, we just go and build another memory plant. It's that easy. When we have a shortage in fiber optic cable, Corning can just fire up another plant. Now it takes time, but they can do it. When we start having shortages of civil engineers, mining engineers, carpenters, electricians, plumbers, you just can't fire that up. You have to train a whole new generation of young people to replace, not even to net grow, but to replace the older people that are retiring from the industry. And so you have to net new hire, meaning we just have to change the compensation structure for this industry. I think the first year analysts in banking, that salary is going to have to be flipped and we're going to have to pay those to the first year analysts in civil engineering school or master electricians and pay them what the bankers used to make.

(42:21):

And then the bankers are going to make what the electricians used to make. I mean, the world's flipping because we're at an infrastructure deficit. So I think that my biggest fear is twofold. One, we don't have the people, and that's going to be really hard. Two, we don't have the critical raw materials like copper and silver. That's going to be really hard. And then three, I think that the social part of the AI buildout is just going to continue facing lots of opposition. I mean, who is going to want one of these AI data centers in their backyard? Because this is sort of the trade, right? The trade, at least in their perception. And I don't think this perception is right, but this is the perception that's sort of permeating out there right now. This is the narrative that is gaining momentum is that, okay, they build a data center.

(43:08):

What that means for you is your electricity bills are going up, your water bills are going up, and you're probably going to lose your job. That's the narrative. And so you're starting to see a lot of societal pushback from building these things. And so that's another element that can derail the momentum that we have.

Vanessa Hui (43:29):

Right. Okay, Dan, one last question. It's one that I ask every guest on Forthlane Features. What is one lesson you learned early on in your career that still guides you today?

Daniel Dreyfus (43:41):

It's a good one. Well, I was a summer intern. My first job in finance was a summer internship at BMO in Toronto. It was the summer of 1997, and I went to go and work for Gizio Bianchini, Don McLean. They were the gold analyst equity research at BMO. And I had just an incredible summer there. And I was young, I was really ambitious and excited. I couldn't believe I had that job. It was a dream come true. I lived at the office. And one day, I think after spending probably three days straight at the office, meaning I didn't go home because it wasn't because they were making the work, it was because I wanted to be there. The director of research, it's a great guy, Mike Miller, took me aside and he said, "I love the ambition. It's great to see." But he says, "You got to live by the three to five year rule." I said, "What's the three to five year rule, Mike?" And he said, "If you wanted to become an Olympic swimmer," he said, "Do you swim?" I said, "Yeah, a little bit." I know how to swim.

(44:50):

He said, "If you wanted to become an Olympic swimmer and you said to me, I want to do that in three months, obviously it's impossible." He said, "If you want to become a Luke consumer and say I'm going to give three to five years and I'm going to pace myself and I'm going to get there in a marathon style." He said, "You could become an Olympic swimmer in three to five years. You can do anything you want in three to five years." And that really stuck with me because it actually formed the way I invest. Now when I speak to a management team, management teams love speaking to me because I don't even ask them what's happening about the quarter or the next quarter. That drives management teams nuts. The whole industry has gone to this short termism, play for the catalyst, play for the quarter, try to capitalize on that.

(45:32):

That's gotten so crowded, so competitive, and I realized that that's not an arena that I want to compete in because it's got the smartest people and the best people and a lot of people competing in that. The arena that's less competitive is the arena that says to the CEO, "Hey, we've got a three to five year canvas here. Let's paint the picture. What are you going to do with your excess capital? How are you going to grow? How are you going to reinvest that capital? What are you going to do to optimize your assets? What adjacencies are you going to get into in your businesses? And let's try to paint a picture of what your company is going to look like in five years. Maybe this is how you look today, but you can become that Olympic swimmer in five years and the market's not realizing it.

(46:08):

" And it was really that comment that actually framed how I invest. And so when you ask the single greatest piece of advice I got, I give a shout out to a fellow Canadian in Toronto. I'm sure he's still there. Mike Miller, former director of research at the Monesbit Burns.

Vanessa Hui (46:27):

What a great way to end. Dan, this has been such a rich, insightful conversation. I'm so grateful for you. Thank you for breaking down what is an incredibly complex macro landscape and such an easy way to understand. I appreciate you and thanks for your time.

Daniel Dreyfus (46:55):

Thanks, and the best is yet to come.