Loading summary
Transcript60 lines
- [00:07]
A
Today, AI, Iran, and energy. One goal of this show is to cover the biggest stories in the world. And so what are the biggest stories in the world? Judging by newspaper headlines, you'll probably say that at this moment, it's two Number one, the war in Iran, and number two, artificial intelligence. But if you dig just underneath that headline, if you poke and prod until you get to what those stories are really about, I would argue that at a fundamental level, they are both about the same thing. And that thing is energy. I am recording this at 3pm Eastern Time on Monday, April 13. The top story in the New York Times is, quote, U.S. begins blockade of Strait of Hormuz. And why are we blockading the strait? A strait that theoretically we want to open on account of its original closure. Its original blockade drove up the cost of oil and natural gas and created all these problems throughout the world. Mr. Trump wants to prevent Iran from profiting from oil exports and force its leaders to accept American conditions for ending more than a month of war, end quote. So this is actually a bit stranger than it might initially seem. What happened is we attacked Iran. Iran blockaded the Strait of Hormuz to hurt us, and now we're blockading the Strait of Hormuz to hurt them. Putting aside the logic here, both sides are using their militaries to wage war on the other side's energy economies. This is an energy war. And then there's artificial intelligence. This morning, the top story in the Wall Street Journal homepage was, quote, AI is using so much energy that computing firepower is running out. In the last few months, demand for AI agents has exploded upward so quickly that companies like Anthropic are struggling to keep up supply, namely computer chips. But computer chips are energy intensive to make, and their top manufacturers in Taiwan and Korea face energy shocks from this war. The data centers that computer chips live in also require tremendous amounts of energy to run and keep cool. I think it's important to remember that not every tech wave is like this. Like the consumer tech revolution of the 2010s, with Facebook and Instagram and DoorDash and Airbnb. Those companies were not in the energy business. They were not fixated on scaling software to build a global consumer base. Today's tech story is fundamentally different. Yes, the demand side is important. Getting people to use Claude or ChatGPT matters to the revenue picture for Anthropic and OpenAI. But more than the tech waves of the past, AI is about supply. It's about making chips, buying chips, building data centers, powering them with electricity. And all of this is a story of making stuff in the physical world that requires molecules from oil and gas and electrons from solar and wind. Which is to say AI too is a story about energy. The author Vaclav Smil once wrote that energy is the only truly universal currency and nothing from galactic rotations to ephemeral insect lives can take place without its transformation. End quote. It has never been more true that every story in the world is fundamentally a story about energy. So I felt the time was right to bring back our plain English energy consigliere, Nat Ballard, the Singapore based energy analyst. Today we talk about Iran, AI and why energy continues to be the story that explains the entire world. I'm Derek Thompson. This is plain English, Nat Billard. Welcome back to the show, Derek.
- [04:38]
B
It is great to be back. I think this is our third time having this conversation.
- [04:42]
A
Number three. Yeah, this is the Balar trilogy and I'm excited to complete it.
- [04:46]
B
Awesome.
- [04:47]
A
So I've been meaning to have you on for a while because it seems to me like the two biggest global stories in the world right now, the Iran war and the Artificial intelligence Build out, are not just. It's not that they're exclusively energy stories, it's just that they are unusually energy stories. Because not every war is is an energy war. Not every technology boom is about the supply of energy that goes into that technology. But this is an energy war and this is a supply side tech boom. And so I figured, you know, we haven't done a single episode using energy as the lens to explain the most important stories in the world. So I thought, who better to do it than with you? Let's start with the war. As an energy guy, what do you think is the most important story of the Iran war?
- [05:35]
B
So we are in a position where there's this physical gate check on the flow of hydrocarbons around the world. And it's been there since the time that hydrocarbons started flowing like 90 years ago out of the Persian Gulf. This is stuff that I studied as a student and was the kind of thing that any war college army lieutenant colonel would have looked at. Anybody in the Navy would have known about. Anybody in hydrocarbons would have known about that. At the moment, about a quarter of all liquefied natural gas and about a quarter of all seaborne oil flow through this one tiny checkpoint. And they have done so generally peacefully for a couple of decades. There's a whole bunch of associated stuff that goes with that, energy derived products. So fuels that are refined from oil and Also chemicals that are created with oil and gas and precursors to all kinds of other industrial products, they've all flowed through there and now boom. All of a sudden that flow has been stopped completely. And it's a supply shock that we haven't had honestly ever. It's like a bigger shock to supply for these flows of things that people I think especially in the western world kind of take for granted and view as a sort of luxury and privilege to not think about are suddenly right in the frame for everybody. So that's kind of like the global thing about this. Honestly Derek, this is the kind of thing that when I was in grad school you might try to war game this scenario and people be like this is ridiculous. Okay, this is never going to actually happen. We don't need to talk about it. Then there's the sort of more to the ground way that this instantiates itself in different markets. I'm coming from San Francisco today, but I live in Singapore like your erstwhile guest Alex Turnbull. And when we talk about this conversation amongst each other over beer, it's incredibly acute in terms of what this means for the country where we live and all of our neighboring countries. And I want to sort of reemphasize something that Alex talked about a lot is that this is not any kind of an abstraction for countries that import a lot of oil, a lot of refined products, a lot of liquefied natural gas, and then a lot of the chemicals that go into things such as fertile and into industrial outputs and things like that. It's just big. It's profound in the sense of it has depth to it and it is also open ended at the moment. I mean as we have this conversation, There is apparently 15 warship flotilla from the US Navy that is preparing to or maybe is blockading the Strait of Hormuz. Even if you were to snap your fingers and make all of this stuff go away instantly, we're not really in a position where the flows are going to instantly come back. There's going to be a lot of gap in the system that needs to be filled. And then there's a longer term question about how much supply is going to ever come out of a region like this because of the way that people configure their demand. Is the flow in and out of the Persian Gulf considered as reliable as it used to be? Is a blue water US Navy considered to be the guarantor of free and safe passage? There's all sorts of things to be considered right now and I guess it's really important to emphasize how early we are in kind of determining the contours of the future based on this.
- [09:09]
A
One thing I always struggle with when I'm analyzing something that Donald Trump has done. Is this a sort of phenomenon that he's going to essentially delete and it will be remembered in history as a two week, two month, six months phenomenon that had no ripple effect on the world? Or is this something more like the tariff policy which might live on in American economic policy for decades, or the shift in Europe toward getting that continent to spend more on, on the military, which also seems to be a kind of secular trend that could absolutely outlive the Trump administration in thinking about the long term implications of the closure of the Strait of Hormuz. I know this is a, this is difficult to say right now because by definition we can't see the long term, but you're the energy expert. I am not. What's the sort of thing that strikes you as a very plausible long term, multi year consequence of this war?
- [10:12]
B
The first, most, so to speak, would be to reduce dependence insofar as possible on the fossil fuel imports. But you're going to run into structural limits of how little you can do. There are things that we currently do not make without hydrocarbons, and specifically liquid hydrocarbons provided at low cost. So there's as much substitution as you will do as quickly as possible in places like the power sector. So you want to do as much renewables as you can so that you're not dealing with important fuel. You will also, and we might talk about this later, amplify your consumption of coal, which can often be a domestic resource or is not blocked up by the Schneider Hormuz. You can electrify light duty transportation. It's a fancy way of saying get a bunch of electric vehicles into the hands of consumers, but you're going to run into limits of how many there are to buy and then also what kind of infrastructure is going to be built to integrate them and energize them at various places? So that's like the easy stuff. The long term question is where do you then shift supply? Because you cannot automatically delete this much energetic demand in the global economy without it being really, I think, catastrophic. So where would that flow then go elsewhere? So that's domestic sourcing insofar as possible in other countries. It's sourcing from other countries for seaborne trade, as Alex Turnbull talked about, you know, countries that are not able to be blockaded by any one navy. So that would be Places like Brazil and Argentina, potentially Mexico, definitely Canada. And it's also a big question for other countries that want to import from the United States. So the US Is a exporter on a sort of a flow basis, not on a certain, for various compositional reasons of where a refining stock is and what we produce and what we consume. We do export oil and we export oil products, we export loads of natural gas. We are the world's biggest exporter. The question will be how much do you want to rely on the US as a reliable exporter in this situation? Can you disentangle the corporate interests of a non state run energy sector in the US from the foreign policy implications of the current presidential administration? And your point about tariffs? I think it's great. It's funny, relatively speaking. Tariffs are a fast moving free radical in the system to my mind, compared to thinking about energy. And so they can be undone at the stroke of a pen. And then it's a question about how much industry will feel the confidence to come back in and do something. I mean, what I think we could say with a slightly wider aperture here is that we've seen countries, and in particular the US kind of whipsawing back and forth in policy with sort of an impedance mismatch so the policy mechanisms can be agreed and then disagreed in the course of a couple of years while trying to inspire or create manifest effects in infrastructure and in industry that are meant to last for decades. And I think that that's the biggest question for everybody is like, how much long term decision making can you make in a situation like that? Even though this might be inspiring really big long term change, how well are you able to make long term decisions that you feel like are both, you're confident in them, you're comfortable about them and you feel like they're within your control.
- [13:50]
A
We're going to return to the subject of renewables in just a second. But I wonder if 10, 20 years from now you think it's plausible that this war is seen as an inflection point. There's all sorts of things that have been tailwinds for the renewable revolution and headwinds for the new revolution. Tailwinds, the declining cost of solar, the declining cost of batteries, headwinds. You've got the Trump administration coming in and just being at the very least chaotic when it comes to the balance of pro renewable versus anti renewable policy and at worst actively against offshore wind specifically. But maybe 10, 20 years from now we say, you know what policy under one administration can be short term, this War scared so many different countries away from their over reliance on ocean bound hydrocarbons that you saw this global acceleration of solar and wind and whatever else that could essentially be in sourced right. The solar panels are right there in Brazil or Australia or the US wherever they're powering households and air conditioning. So there is a possibility, it seems to me like 10, 20 years from now this is seen as an inflection point. Is that Pollyannish? Is that overly hopeful? Or do you think this war really could re accelerate the global renewables revolution?
- [15:12]
B
So I actually think it's fairly realistic with the caveat that you're creating a different kind of potential trade dependency as long as you're not providing that capability domestically. And there's two ways to view that. One is that a big country like the United States and big industrial countries, Japan, Korea, Germany have the capability to basically manufacture anything that they would choose to do on their own. It is a policy decision to do so. But you could return to what Germany had, which is a fully domestic solar supply chain. And the United States, Korea and Japan are capable of doing this on their own. Clearly China is capable of doing it on their own. But what's the sort of the lower bound on that? Brazil has had lots of local content requirements, so has India to be doing a lot of renewable resources in inside the country. But is Australia going to do that? Is Argentina going to do that, Is Mexico going to do that? And then as you go sort of further out, you're getting to countries that have essentially no industrial capability and the ability to import whatever they need. But they were trading one kind of import dependency on a flow for another import dependency on a stock like a thing that they need to buy and build and own. And so I feel like there's going to be some parlay between that in certain countries. That comes right back around to my policy question. That's only going to happen if you've got policy support to do so. Do you feel like there's enough of a long term window for that? But for the tradable goods, I would say this is probably a positive for a very long time. The tradable good, that is the thing that you sell to provide energy. That's solar, wind and then the batteries that back them up, the EVs that charge off of them. I feel like this is probably a positive. And I also think there's an opportunity here where it kind of jumps past the policy blood brain barrier. And that would be part of my hope. I think if we want this to Be durable in the long term. It's a decision that people are making at the country level, at the company level, at the individual level, to say there's an opportunity here to minimize this dependency insofar as possible, with technology, to trade resource for technology. And again, with a caveat that you can only do that for so many things. You can't magically wish away ammonia production. Right. Like that. Like, that's not something I can do with an electric vehicle.
- [17:31]
A
Yeah. I was thinking, you know, as a way to sort of round out this section on the effects of the Iran war on artificial intelligence. I'm thinking about the ways that it might durably destroy demand and the ways that it can create demand. So where it's destroying demand is, you know, especially across Asia and Australia, you've got political leaders that are essentially telling their populations like, use less, there's less to use, don't turn on the air conditioning, don't drive, work from home, take Wednesdays off, cancel school once a week. I mean, that. That is a catastrophe. Like in parts of the world, maybe like the US it's merely a crisis in parts of the world where you've got governments canceling school because of the lack of inflow of hydrocarbons from the Strait of Hormuz. I mean, that is reaching the level of national catastrophe. But that's just where demand is being destroyed. And what I'm taking from you is that there's also ways in which the Iran war might induce demand for different sources of energy. Right. If there's this elevated premium on ocean BO oil and natural gas for the foreseeable future because we're just so scared of another crisis breaking the ACL of the Strait of Hormuz again, then you have more countries essentially saying, okay, well, if we can't rely on that as cheaply or as automatically as we did before, well, we have to find some other way to get energy. And that some other way is going to be more local. So it's going to be buying solar panels or buying batteries, potentially from China. We'll get to that in a second. Maybe turning on the nuclear power plants. And there's all sorts of ways in which I see new kinds of or kinds of energy being demanded more because of this war, which might accelerate some kind of transition before we move on to the connection between the Iran war and artificial intelligence. Maybe any last ribbon to tie on this section as you see it.
- [19:18]
B
So I'm broadly in agreement with all of this. I think my only caveat is that we need to remember that there are some things that we can solve through physics. And that's basically anything that's already electrified that can then be further electrified or electrified anew with renewables. And then there's chemistry, and that's the world of molecules. Alex Turnbull talked about this a lot more eloquently than I can. And your audience therefore has probably already heard a little bit about it. But you just can't wish away things that require breaking molecular bonds that come from a fundamental energy source and require different types of energy inputs to make them. You can't just wish them away.
- [20:04]
A
What's an example of this?
- [20:05]
B
An example of that would be making fertilizer, for instance. Like this is a process that does require quite a bit of electricity, but what it really requires up front is natural gas. Making urea and ammonia, for instance, are the result of petrochemical processes. And those things then flow back into becoming part of fertilizer and making food. And there are technologies that do this, by the way. I know of companies that will use renewable technology to fix nitrogen and therefore be able to start making fertilizers, but not in a kind of volume that feeds 8 and then 9 billion people. And so I want us to remember that concerted efforts need to happen to solve chemistry problems, not just physics problems. We get very easily attached to for good reason. Solving the physics problems because they're solvable in many ways. But the chemistry problems are really, really tricky. And they're the thing that requires either complete reimagination and a root level science way, or they require us to think very differently about what we do and what we use. And it's tricky.
- [21:14]
A
Yeah, I think as excited as anybody can be about the acceleration of solar and wind and geothermal and all of this, there's just some stuff that we don't know how to turn into an electrification process. So people need to eat. Eating requires fertilizer. That requires natural gas. There's a lot of plastic in the world that's also a hydrocarbons business. Stuff is made with concrete. We don't know how to make concrete in a low carbon way. It's an incredibly carbon intensive thing to produce and it is necessary in order to put up buildings. I'm sure people are familiar with if they're looking at vacations for this summer or, you know, later in the winter of 2026. What's going on with jet fue? Planes are amazing. We don't know how to power them at scale with electricity. We just don't like. They just, they require hydrocarbons, and they probably will for the next few years, if not decades, if not maybe century. And so there's all of these enormous systems where, notwithstanding all the incredible things that are happening with renewables, we just don't know how to move the dial there when it comes to pulling back our dependence on hydrocarbons for now.
- [22:21]
B
Right.
- [22:22]
A
Maybe what we're gonna get at the end of this interview too, the sparkling potential future of electrifying the world. But I guess to round out your point, there's just some stuff that we don't know how to make do without. Oil and natural gas and sometimes coal, to build a bridge from the consequences of the Iran war on the energy picture to artificial intelligence. At what point do you think this war starts to meaningfully impact the economies of Taiwan, Korea, countries that are critical for international supply chains of necessary chips, memory chips, for the construction of artificial intelligence in the us? At what point do you think it will be factual to say this war is clearly hurting the AI economy?
- [23:14]
B
I mean, this is a really big fundamental question, and I think it's important that you phrase it this way in terms of thinking about it as these critical path inputs that themselves are easily bottlenecked just by the fact that not that many companies make these things and also by the fact that they themselves are very dependent upon both energy and chemical inputs that are part of the flow that comes out of the Persian Gulf so soon. Potentially, I don't have a great answer to that because it's a prioritization question within countries and sort of analogous to what I think we'll end up talking about regarding AI and the great American consumer in a minute. But at what point do countries decide it's air conditioning in the summer or it's running this TSMC fabric? At what point do you make this decision that you have to do one critical thing for the global economy or one critical thing for your own economy? And there's always going to be a high willingness to pay for AI's inputs, certainly, which is partly what complicates this conversation, because these things are so critical to AI buildout and training. There's every reason to pay, there's every impetus to bid prices up. You have the world's biggest and most cashed up corporations capable and interested in doing so. But you've also just got limits in countries where they say there's only so many people that work at this fab and there's only so much GDP that comes from it versus suffering through a hot summer. So look, we sort of haven't tested this yet. It is mid April, it is not yet July, it's not yet August. I can tell you that if we're already in the process of wearing shirt sleeves in government offices and reducing school hours so that you can reduce energy inputs, nothing will be immune from this pressure, but it will be highly concentrated, which is I think, part of the point. This is specific pressure points on specific companies in specific places.
- [25:30]
A
It's interesting to think of artificial intelligence being akin to a mouth to feed in a place like Taiwan in the near future. Like, do we feed the people, which is to say, give them air conditioning, or do we feed artificial intelligence again with the same electrons that would otherwise go toward air conditioning for human beings? There's a way in which that trade off is already being felt or observed in the U.S. if you look at the last 15 years of private fixed investment in the U.S. we're seeing a world where private fixed investment in the US that's not artificial intelligence is declining while investment into AI is absolutely surging. This is a chart in your presentation for folks who are watching on YouTube or Spotify. You can now see this chart. Before we talk about it, tell me a little bit about what we're looking at here or for those who aren't watching, what we're talking about here. Because this is an incredibly unique moment where it almost seems to me, and this is where I want you to edit my language. It almost seems to me like the AI boom is stealing resources from the rest of the economy.
- [26:42]
B
So the AI boom is an unusual thing in the big booms in capital expenditure in the US and we've got like half a dozen of these things over the course of the last couple of centuries. Railroads, interstate highways, the Apollo program, building out broadband and fiber networks. The one that always gets memory hold. That was actually, until now, probably the biggest in nominal dollar terms, which was upstream oil and gas exploration and production during the shale boom. What's interesting about all of those things is that some of them are providing capacity. Like building broadband is clearly about providing new capacity to an economy. Oil and gas is clearly about, about providing new energy to the system. And the highway system is about creating connections. Obviously you're creating connections through AI. But the biggest thing that I can think of about this in terms of way to measure it is that you measure it based on the energy input, the dollars that go into it, and the energy that goes into it. So the economies themselves, it's fascinating they look at this from a perspective of energy, not just obviously a Function of compute the amount of computational power that's being provided. But when you spec out a data center, it is specked out in terms of the amount of energy that it requires. And that's unusual. It's this boom that really does draw fundamentally from everything else as a first order requirement for what it is. And in terms of what's being built. Yeah, what you're seeing is basically declining fixed asset investment. So that's like anything long dated. That's power, that's energy, that's infrastructure. That stuff is sort of coming off the boil. With the exception of AI related stuff, which is basically the driver of a lot of capital expenditure in the us. The Capex boom, this is something I've talked about elsewhere in the deck and in other places is right now the second largest capital expenditure as a share of GDP in the US ever, bigger than the railroads in the 19th century and second only to the Louisiana Purchase, which I would say is debatably a capital expenditure, by the way, in nominal dollars that was only $25 million. So we have to adjust that for GDP. But it's just unusual. And I think it's really important to think about it this way is that it's a thing that first and foremost requires all of this input in order to happen. It requires the money, which any capital boom does, but then it requires energy in order for it to work. As opposed to money creating energy. It's money creating demand for energy. I guess we could say I want
- [29:22]
A
to understand how we are powering these data centers right now and whether the popular understanding that artificial intelligence is gobbling up electricity that might otherwise go to consumers, thereby raising the cost of energy for Americans is a narrative that you think matches up with the evidence as you see it. So how are we powering these things and is it indeed a fact that powering them is driving up energy costs for everybody?
- [29:52]
B
You can hear the sound of a thousand power analyst knuckles cracking as I sort of lean in here and try to do my best. So I think it's important to step back. The demand for electricity in the US had a 15 year plateau. Like almost my entire career was at a time when power demand was not growing across the country. The annual total consumption was pretty much flat. And it's important to lay the groundwork here. Part of that was a lot of efficiency. Weirdly, a lot of it was fluorescent and then LED lighting. A lot of it was a more efficient economy. A lot of it was, to be honest, an underperforming economy. I think for a lot of that time, a sort of quasi austere economy in the 2010s. And in that context, today's hyperscalers were able to get whatever power they needed from a system that had ample supply and their demand was relatively low compared to the total demand that's there. Then you have a situation where demand is growing in a secular way. It's not just data centers. We did build a lot of new manufacturing capacity, continued moving to the Sun Belt, drives up demand for power, for air conditioning and things like that. And you have a little bit of growing baseline, and then you've got a lot of this growth on top coming from AI. And it's like potentially, let's say 4% of US power demand right now coming from that. So it's big and it's non zero. And it's growing in a system that's also growing. So it's growing on a growing base. And that means that it's inevitably going to have some interaction with the rest of the system. And this is where things get a little more nuanced, I think, which is that the costs that ratepayers are paying is a function of the fuel inputs or the energy inputs that are providing them electricity. And then the cost of the system itself, the infrastructure and the network. It is those costs primarily that are going up. In fact, if you look at the biggest grid in the US over time, with the exception of massive storm events and real energy disruptions, it's like a monotonic function of the declining cost of the energy that goes into the system. But the cost of the network is what keeps going up. And we have more specific.
- [32:10]
A
When you say networks, what are you talking about? What machines are you talking about?
- [32:14]
B
Just keep going. The transmission, so the long distance stuff and then distribution, which is what delivers power to the end user, those costs are going up and those costs are rising faster than the rate of inflation. And also that's what you need to build to provide access to a data center, or for that matter, a hospital, or for that matter a factory, or for that matter a stadium, whatever it might be. But those are the prices that are going up that are kind of inescapable. And the question is, how much of that is intensely localized in an area where people feel that, either directly or semantically, however you want to say it, in terms of what their rate is going to be. I see a $10 billion data center, I see a billion dollars worth of power infrastructure that needs to get built for it, and I see my rates going up. This is to an extent, there's an element of this that is just sort of unavoidable math. There's a lot of nuance underneath it, such as utilities essentially asking or forcing data center operators to pay their own way. They pay for their own power, they pay for their own infrastructure. That really does remove that impact on ratepayers, but it doesn't remove the physical substantiation of this stuff. People still know that it's there, they still feel it. And they also feel the fact that new things that are Manhattan sized infrastructure projects are being built in places that previously had very little or any of that kind of infrastructure.
- [33:49]
A
I'm going to recapitulate your answer.
- [33:50]
B
So complicated, right?
- [33:52]
A
Let me try to make it simple. We named this show for a reason. Let me try to make this a little bit more simple.
- [33:59]
B
Let's go.
- [34:00]
A
I think a lot of people believe that energy prices are going up for the quote unquote exciting reason, which is that AI is the demon and one of the demon's faculties is making everybody to pay more for energy. You are telling me. No. Energy is not getting more expensive for the exciting reason. It's getting more expensive for the boring reason, which is all of those boring little machines that go into an energy grid. The transformers, the wires, this, the cost of these machines, the cost of making them, sighting them, getting them integrated into the system. That is what is principally driving up the cost of energy. Am I right so far about the disagreement? You are right.
- [34:43]
B
So here it raises an important question,
- [34:44]
A
which is why is this happening? Why is the cost of the wires, the transformer machines, why is that going up? Is that more about it's more expensive to build these things because we can't get enough of the materials necessary to build them, or we can't make enough of the plants that are necessary to build them? Or is this a function of siting? It's harder to actually put these machines, which we can build, where they need to go. Whether it's because of FERC or NIMBYism or something else that has to do with the geographical distribution of building out the grid.
- [35:21]
B
Okay, so let's talk about transformers first. The critical thing to get energy between big system, small system, those are way up in fundamental cost because tight supply chains, tariffs, low capacity to actually build, and a lot of demand. There's no wishing that away. Not a great satisfactory answer, I think for a lot of people, because it seems so far away and abstract, but that's true. Cost of wires goes up because of the material that goes into Them copper, steel, also both things that are subject to tariffs. The cost of labor has gone up, up. That's a post Covid effect on a lot of these things. Where I am right now at the moment in California, costs go up because there's a need to secure the system against wildfire impacts. There's like tens of billions of dollars of spend that has nothing whatsoever to do with generation of power or with the consumption of it, but entirely about making sure the system is as fireproof as possible. And then there's also the fact that a lot of this was effectively deferred. If you don't grow for a long time, it's hard to justify a lot of investment and so been some deferred investment that needs to happen now. That therefore is more urgent and is therefore more expensive.
- [36:34]
A
I think this conversation connects to politics right now. To be an energy populist, the easiest card to play is to say artificial intelligence is raising energy prices. We need a moratorium on data center construction. We need to stop the construction of the machines that are powering artificial intelligence. What I'm hearing you say, and you're the expert here, I'm not. I'm just trying to recapitulate accurately. You're saying, that's exciting. That's an exciting answer for why energy is going up. The boring and more true answer for why energy is going up isn't that we can't make the machines that make artificial intelligence. It's that we can't make the machines that make energy for consumers and artificial intelligence. And what a kind of wise populist might try to do here is say, hey, my fellow Americans, you're right about the cost of energy. You're right to be upset. But the enemy here is not the data centers. The enemy is the materials and policies that go into the construction of the transformers and the wires. I mean, I'm not sure that's a very. I'm not encouraging anyone at the political level to pick up this language. I'm just interested in understanding the landscape as you see it. Is it fair to say that the politics, the data center moratorium, are going after the wrong culprit for the rise of final energy cost to consumers?
- [37:58]
B
They're going after the most visible, the most easily rhetorically addressed challenge that's out there. This thing didn't exist and now it does. It requires a billion dollars worth of infrastructure in my community that will be distributed across everybody who pays the power rates. Therefore, I don't like it. I would argue that the truest possible Strongest possible response is, I believe there was a book about this. There's something about it being abundant. What you really need to do is be building more of everything. You actually need to be providing more capacity faster for everybody in a way that will ensure that costs come down. Build more renewable power that will lower the actual energy costs of your energy bills. Try to do as much coordinated buying and building of infrastructure as quickly as you can so that the costs are more easily distributed across everybody. Who's paying for power? Do what we're starting to see within the politics of the data center, which is work with the providers that themselves are sourcing their own power and are paying for their own infrastructure so that it is not pushed down on the ratepayer. But these are all very kind of tactical kinds of solutions. But are they easily expressed in a political context? And I appreciate you wanting to have them expressed that way because I think it's really important. Everything is politics when it comes to this. And when I talk to data center operators, the biggest and best of them work not just hand in hand with their energy provider, but integral to their process. They don't exist without each other and they're all looking to move as fast as they can while minimizing or completely eliminating any of the costs that could be passed on to the wider community and trying insofar as possible, to get in front of that message where it helps. But it's a challenging one, I think, to make for a number of reasons.
- [39:55]
A
You've said two things that I want to put side by side. One thing you said is that a lot of these data centers are trying to go behind the grid. They're trying to provide their own power, or they are agreeing to provide their own power to avoid the reality or the perception that they are absorbing electrons that would otherwise go to cooling the bedroom of a sleeping five year old child. We also have on the table the fact that there are are more and more politicians from the left and the right becoming AI populists and arguing for a moratorium on data center construction. Is there a way in which the solution to the data center moratorium moment is for all of the hyperscalers to essentially say we now promise to provide all of the energy inputs to our data centers to guarantee that none of the local consumers are going to be harmed or affected by the energy that we're supplying? Is that an answer to the AI populism of the moment, as I'm calling it? Or is there a way in which that creates its own problems?
- [41:14]
B
So I could Certainly see this kind of approach taking hold at a kind of concerted and consolidated level. It definitely happens at local levels all the time. These are the kinds of things that people do and to go along with them. What other actual positive investment that is outside of this particular domain can you make? Are you supporting local emergency services? Are you supporting local public goods and things like that? This has long been a part of the tradition, by the way, of building infrastructure everywhere. If you're making a $10 billion capital commitment, how else can you extend that capability into supporting a community in other ways? So certainly the politics of that, I think are possible to be extended, but it is something that seems to be, and to my mind, a little bit weirdly, both right and left AI populism at the moment in the US
- [42:14]
A
Want to make the most of your tax refund file with TurboTax on intuit credit Karma? They help you get your biggest refund and then we help you do more with it with a personalized plan designed to help you hit your money goals. Start filing today in the Credit Karma app. Can we make a prediction here? I want to understand what you're seeing sort of at the edge of the present. So no need to go five years down the road. I mean, who the hell even knows when what artificial intelligence is going to look like in five years? But where is this going to go in the next, say, 18 months? I mean, we're going into a midterm, we're going into a period where both parties are going to have presidential primaries where I cannot imagine artificial intelligence not being absolutely core to every single televised debate. Where do you think this is going to go with regard to data centers? The energy question, I think it will
- [43:10]
B
be sort of what you've already suggested. I think you will see every bit of effort to get in front of this as a potential issue and to point as operators to as much positive impact as can be done. Like, how are you, A reducing cost completely to the consumer and B, how are you looking to provide something else to the community as well? So that's what I would imagine we see in that next 18 month window. Beyond that, there's a lot of more technical stuff that I see coming. But I think in that window this is probably, probably where you would see it settle in and shake out.
- [43:45]
A
Moving to the U.S. i want to talk a little bit about the vibe shift of the last few years. I feel like we've gone from a period of maximal ESG investment, tons of money flowing into renewables and renewable projects and pro renewables venture capital firms, a lot of that is reversed. Far fewer mentions of ESG in corporate earnings statements, outflows of investment income into this general space. And yet at the same time, solar and batteries seem to be being built. I mean, if you look at, I think it was Texas that just recently set an all time record in terms of the share of electricity that's being generated by renewables. Famously climate change obsessed with Texas. So something is happening here that's kind of interesting. Where, where on the one hand, renewable energy as a vibe seems like it is in significant decline, but renewable energy as a material phenomenon seems like it's continuing to move forward. How are you seeing this picture?
- [44:50]
B
100%. So if you asked any of my friends, associates who develop energy for a business, renewable clean energy, batteries, working electric vehicles are like, please don't make this an ESG thing for me. I build stuff. I'm a chemical engineer. I do trillions of dollars annually, collectively and annually worth of hard infrastructure for energizing the world. And I do it in a way that happens to be low or zero emissions. ESG is like 50th on the set of concerns, considerations and inputs for that process. That's really a thing for certain elements of, of financial positioning in rhetoric more than anything else. And it's weird in a good way that the kind of decline of ESG as this investment focus, a particular driver has had little to no impact on the deployment of capital and what I would consider to be environmentally positive assets. There's a whole bunch of.
- [45:53]
A
Before, I don't want to. Again, given the name of the show, I don't want to say ESG back and forth without defining it. This, these are investments or this category of capital I.e. pro environment, pro social, pro governance. And what does social and governance mean?
- [46:07]
B
It's up to you.
- [46:08]
A
But from my perspective, thought of as like it is sometimes thought of as goody two shoes liberal capital. Is that a little bit unfair?
- [46:20]
B
What I would say is fair. Unfair. What I would say is, is an inevitably fraught marriage of things that have nothing to do with each other. I'll be more blunt about it. You can have companies with a great environmental record, an okay social record, and a very bad corporate governance record because they have Golden Share shareholders who control the board. Okay, why have these things been allied together? I honestly don't know if I could go back in time and get the foundational document or consulting PowerPoint that annealed these three things into one. I would love to find it, but they have nothing to do with each other, to be very clear. And this is why to my point, the people that I know who build a lot of these things are like, please don't put me in this bucket. I don't raise capital based on that. I don't deploy capital based on that. And I'm not interested in being part of a conversation that has anything to do with any of that. And so, yeah, I appreciate you making me step back in to kind of disambiguate this stuff and make it into these real terms. But in so doing, I actually find it really, really helpful. Like one of the things very quickly, as an aside that I used to do in my previous life when I would do big conferences about this stuff is I would make all the participants define every acronym and if they couldn't immediately do it, then they weren't allowed to talk about it. We have ways of.
- [47:46]
A
That would definitely explain why I never talk about esg, because I have to confess, I never question quite understood what it meant based on the particular audience I was hearing it from. But I don't want to move this conversation too far away from the underlying question, which is this schism between the way the vibes recession around renewables, which seems to be happening at the same time that solar and batteries are still booming in parts of this country. So I wonder sort of how you see this schism.
- [48:14]
B
Well, these are the lowest cost, quickest to build new ways to add either generation or capacity for power to the grid. That's it. Right? And that's the simple reason that even in the current administration, the U.S. annual energy outlook from the government, the Energy Administration Information Administration says we're going to have declining fossil fuel burn for power, declining fossil fuel consumption and transport in every scenario that we can think of in the future. It's just the math thereof. And it's funny because it's funny. The vibes are really kind of political while the actual flows and numbers are financial.
- [49:04]
A
I wonder what you think is the biggest energy challenge of the US in 2026 versus in the next decade. I mean, there's all sorts of things that are happening at the same time here. So you've got a wide buffet to choose from. There's the Iran war. There's the fact that. And again this is from your report. Global gas turbine manufacturing capacity is well below current demand. So this incredible source of energy that we've had for the last few decades, which is natural gas, is having some of its own challenges. There's the continued challenge to add renewables. If Donald Trump has decided that he absolutely despises wind and from time to time hates solar as well. I wonder how you would answer this though, double barreled question. The biggest energy challenge for America in 2026 versus the biggest energy challenge for the next decade.
- [49:50]
B
2026 is very simple and it's one word. It's cost. In the abstract we talk about rising costs, but there are double digit percentages of U.S. households that report difficulty paying for their power bill right now. And that's with an economy that is not, not in a recession. It's easy to look askance at people who bought an F250 as their daily drive and might be now paying $5 a gallon for gasoline, but they have no flexibility in the meantime in terms of what they can do. These costs are passed through instantly and they are the most important things for people to pay, perhaps besides food. You pay your power bill and you pay your bill for gasoline because you need these for transportation and you need them to actually live. And so we have an exorbitant privilege in the US relatively speaking, which is we're not going to have deep molecular shortages based on changes of flows in and out of the Straits of Hormuz. But prices for many of these things are set globally. And then there's all the other factors that I mentioned about why costs are going to go up for things like electricity. And they're very real. And I think there are oftentimes people in more analytical roles that tend to easily dismiss this because you can math the heck out of it to find ways to say that we don't spend a lot on this or that or the other. It's not that big of a deal. But if you earn lower income, you spend a much higher percentage of your income on energy and on food and on transportation, and you cannot wish those away completely as either income changes or as costs go up. So that's the biggest one, I think, in the next year is just costs for everybody. Thinking further ahead, the biggest challenge is deciding what we want to be when we grow up. What does the system want to be? Does it want to be big? Do we want a power system and an energy system in the US that's 50% bigger than it is right now? I would say yes. Low cost, abundant and ideally very clean energy is the input to so many of the things that we want to do in the future. The other question is like, how much of that do we want to be within our own domain? Like we manufacture that ourselves, we keep critical stocks and supplies of all of the inputs that we need. We try, insofar as possible, decarbonize as much as we can because that's better for reducing risks and potential shortfalls. But these are all big questions that are kind of higher level than just what engineers and planners can do. They're almost like societal questions about how we decide what we're going to be. And so that's our big challenge, I think. Looking at 10 years, I want to
- [52:48]
A
close by asking you a couple questions about the automotive industry. This is a fact from your deck that hopefully people on Spotify and YouTube can look at. Just before COVID China sold hardly any electric vehicles to the world. Like maybe 1 or 2 million to the world. In 2026, China is expected to sell more electric vehicles to the world than the US buys vehicles entire. So like the Chinese export market will be larger or manufacturing industry will be larger than the market of American demand for cars. This is extraordinary. And you have all sorts of other charts showing that byd, the largest EV manufacturer in China, has now clearly surpassed Tesla when it comes to global production. The US auto industry just looks like it's in a ton of trouble. Is the American car industry kind of doomed?
- [53:53]
B
So actually, I think the US auto industry is really challenged by new technology and by the whipsaw of policy back and forth. First being told to go all in on EVs and then being told not to do any of that. We're also like a weird island in terms of the type of cars that people buy in the US I don't see a lot of people in Vietnam driving F250s. I just don't. And while I do see all kinds of pickup trucks, they're not quite the same thing in Thailand as they are in Texas. And weirdly, I think the US is somewhat more insulated than others, but it's islanded. It's a smaller and smaller potential share of the global market. And that's been going on for decades by the way. The bigger challenge is for everybody else facing down the Chinese auto making and Chinese auto exports. Europe, Southeast Asia, Latin America as well. And then the related question is, does the US get the chance to take advantage from a customer and consumer perspective of low cost, high quality, high performance EVs made in China. In the future, I see Chinese EVs everywhere in Singapore. BYD is the biggest seller of cars full stop in Singapore. And they are delightful as a driver, as a consumer, they are more economical than anything else. They obviously have to have a lot of infrastructure built around them to go but they're an experience for drivers that I think is, is self evident in so many different places and in so many different ways that the US is just sort of walled off from at the moment. So weirdly, it's the rest of the world that's feeling this impact first. But it has major implications for the future of US automaking as an export market. Are we just making things for this particular island of ours, so to speak, or are we making things that are fit for the rest of the world?
- [56:07]
A
I think I remember the last time that we spoke, you said these cars byd and what's the name of, how do you pronounce the name of the other manufacturer that begins with an Xiaomi? Xiaomi. These cars are just absolutely extraordinary. And I almost thought when you told me this a year ago, it's almost like a reverse Gorbachev that rather than like Gorbachev come to the American grocery stores and say, oh my God, you have 27,000 different kinds of cheese and 29 different kinds of Heinz ketchup. It's like that for electric vehicles where for an American to go to Singapore or maybe some other Pacific island, I don't know, China itself, Vietnam, and be able to experience these BYDs and Xiaomi cars. It's a bounty of EV choice and diversity that maybe they just don't have in the American market. And so it might be one of the rare consumer markets in the world where Americans clearly don't have first class access because first class is just barred by export import law to be available to American consumers.
- [57:16]
B
I agree 100%. This is my lived experience every day in Southeast Asia that the best new cars on the road by far are all electric and they're almost entirely Chinese and they're not coming here anytime soon. Soon. And for American consumers that are trying to decide what new car they want to buy at what price points, actually there are plenty of low cost EVs right now that are quite good. But the efflorescence, the huge volume of these and the styles and the fit and finish of what you can get. Go to a shopping mall in China and you will find a Xiaomi dealership where you customize everything as you do at a Tesla dealership. But more colorways, more custom treatments, more wheel types, all being done there in house, all basically on demand and in a vehicle that you can have specced out to Ferrari level driving specifications if you want, or at low cost. Still better driving specifications than almost anything you can get elsewhere.
- [58:22]
A
Do you think it's possible that in the next few years, sometimes the American auto market and the Chinese auto market might have to converge in a way. Because if America is making expensive trucks for Americans, so high profit margin trucks for Americans in order to stay alive as a company because they can't export to the rest of the world, no one else wants or is buying these F150s. Whereas in China you've got these companies, BYD Xiaomi, who are barely making any profit, as I understand it. They're working at the slimmest of all possible margins in order to eke out advantage in a really, really competitive industry. Is there a world in which we just might need some conversions? These Chinese companies are going to have to raise the price of their cars in order to actually be profitable enterprises that don't exist because of a circumstance that's drawn up by the Chinese government. While at the same time, eventually Ford and GM have to discover that there is a benefit to making cars for consumers that don't live in the United States. Is there a possibility that though right now it seems like the gap between America and China could not be any further, there might have to be a little bit of convergence in the near future.
- [59:31]
B
It really is going to be consumer led, to my mind. What's the consumer demand? And this is to your Gorbachev point. Well, how do we know I get to see Chinese EVs all the time. If you were to have a Chinese electric vehicle barnstorming tour of the United States, let's imagine that this is the early 20th century all over again and you're basically doing whiz bang demos of new stuff.
- [59:55]
A
It's like a World's Fair. Yeah, exactly.
- [59:58]
B
Imagine doing that.
- [59:59]
A
Look, an electric toothbrush.
- [60:00]
B
No, precisely. Come on, look, it's an electric car. Sit in, sit in and do the acceleration test and have it peel the enamel off your teeth. But then also see how well it does as an entertainment system. Check out how easy it is to clip in a car seat. How many groceries can you fit in the back? I think people would be. I think people would be sort of surprised. The perception in many ways is lagging the reality in terms of what's being built. And to your point, on the Chinese companies, they're basically competing in an auto industry at tech hardware manufacturing margins, at smartphone margins, ex Apple and ex Samsung, and that's a tough spot to be in. And there are dozens of Chinese EV manufacturers, which is also probably not the steady state in the long run. We had hundreds of auto manufacturers in the US a century ago. Europe had a similar number. There's a reason that we sort of coalesce around a smaller number of bigger companies over time. So to an extent, I kind of view this as an inevitable shaking out without having to propound one company or another as a likely winner at any given moment. Some of them seem great, but like the Xiaomi that you mentioned, those cars, they didn't even exist a couple of years ago. Right. This is like a smartphone in home goods company that has made what is a really lovely car and driving experience.
- [61:26]
A
Well, netballard, I always appreciate talking to you about the world of energy and always appreciate you making me feel unbelievably jealous about the automotive abundance that you live with in Singapore. It sounds tremendous and one day we might have it good. Thank you very much.
- [61:42]
B
All right. Thank you, Sam.