December 6, 2024 – Financial Sense Newshour's Jim Puplava interviews Mark Mills, Executive Director of the National Center for Energy Analytics, on the rapidly growing demand for electricity in the US, driven by AI, data centers, EVs, reshoring of manufacturing, and the increasing need for reliable baseload power. Mills explains how data centers alone consume around 50 times more electricity per square foot than office buildings, and AI infrastructure is intensifying demand. Mills emphasizes that natural gas will remain the primary source of new electricity due to its affordability, scalability, and reliability compared to renewables. He critiques the unrealistic reliance on intermittent power sources like wind and solar, highlighting the urgent need for smarter energy planning to prevent future grid instability and blackouts. Want to know more about what's likely to happen under a Trump administration and with energy policy the next four years? Listen in and find out!
Bio: Mark is the Executive Director of the National Center for Energy Analytics, a distinguished senior fellow at the Texas Public Policy Foundation, a contributing editor at City Journal, a faculty fellow at Northwestern University’s school of engineering, and co-founding partner in Montrose Lane. Mark served in President Reagan’s White House Science Office and, earlier, was an experimental physicist and development engineer in microprocessors and fiber optics, earning several patents. He earned his physics degree from Queen’s University, Canada.
To speak with any of our advisors or wealth managers, feel free to Contact Us online or give us a call at (888) 486-3939.
Stay ahead of the market! Subscribe to our premium weekday podcast
Mentioned in today's show:
- Grid Strategies report: U.S. power demand dramatically increases yet again
- Power demand to soar 15% in next five years thanks to AI and manufacturing boom
- Green Energy Needs Minerals, Yet America Blocks New Mines (video)
- The Inflation Reduction Act: A Looming Political Earthquake | City Journal
- MIT Free Speech Alliance Debate: Is Decarbonization Worth the Cost? (video)
- America’s LNG Future
- Meta to build $10 billion AI data center in Louisiana as Elon Musk expands his Tennessee AI facility
- Elon Musk powers new 'World's Fastest AI Data Center" with gargantuan portable power generators
Transcript
Jim Puplava:
AI could increase electricity consumption by 28% by 2030. Data centers consume 10 to 50 times the energy per floor space of a typical commercial office building. EVs will require more electricity to charge the electric batteries. The cloud requires enormous data centers, which will consume 8% of U.S. power by 2030. Reshoring will require more power to run new factories. Renewables will also require more baseload power to fill the gap left by intermittent renewable power. Where will all this electricity come from? That’s the topic of today’s show—the biggest story of this decade and the next—and very few are talking about it. Joining us on the program is Mark Mills. He’s executive director of the National Center for Energy Analytics. He’s a distinguished senior fellow at the Texas Public Policy Foundation and a faculty member at Northwestern University’s School of Engineering and Applied Sciences. Mark, we are seeing power-demanding hotspots. We’ve seen them in California, Texas, Georgia, Virginia, and North Dakota. The demand that’s going to materialize in the next few years vastly exceeds our current plans to build generating capacity. Where are we going to get this power—certainly not from renewables?
Mark Mills:
Well, there’s an easy answer. It’s a two-word form of energy: natural gas. That’s going to be the old Pareto law. 80% of the net new electricity for the kinds of demand you’re talking about will come from burning natural gas.
The other thing you said was that data centers use 10 to 50 times more electricity per square foot than a skyscraper. For commercial buildings, the number is closer to 100 times more. They use a lot of electricity per square foot. And data centers—the so-called enterprise-scale data centers—in total square footage are the size of skyscrapers. They’re not tall, but in square footage, they’re skyscraper in scale. The world has, depending on how you define a skyscraper, on the order of 500 to 1,000 skyscrapers, and we have 5,000 data centers in the world. We’re building data centers faster than we’re building skyscrapers.
So, to your point, you can do the math. We’re building out commercial square footage at two to three times the rate for data centers as we are for skyscrapers. And the power per square foot is 50 to 100 times more. That begs the question: Where are you going to get the power? The answer is from sources of electricity that we can build quickly and that are affordable. The things you can build quickly and affordably in most of the world are still coal, by the way. Massive buildouts of coal-fired power plants in Asia are still underway—not slowed down. If anything, they’ve slightly accelerated.
In the United States, the answer is natural gas. That’s what’s going to be used. It’s what’s being used. By the way, according to the latest EIA (Energy Information Administration) data, the order book for natural gas turbines and natural gas-fueled diesel engines—i.e., diesel-type engines or generators that burn natural gas—is off the charts because they can be built quickly, far faster than you can build windmills and solar farms, far faster than you can build coal plants or nuclear plants. And we have lots of gas—natural gas. In America, we are gushing with natural gas.
So that’s the short answer, but it’s going to be a fight because a lot of environmental groups really hate natural gas.
Jim Puplava:
That’s amazing. They don’t like nuclear; they don’t like natural gas—both of them are clean burning. I’m jumping the gun here, but isn’t this what Elon Musk did in his recent data center? I think he brought in two natural gas generators, and he was able to do it within three months.
Mark Mills:
Exactly right. So, what he did is he built a GPU array of Nvidia AI chips—the biggest one built so far—that consumes 50 megawatts on its way to 100 megawatts. They built it in record time. Musk petitioned the utility, of course—remember, they’re in Tennessee, under the Tennessee Valley Authority domain. They had enough power at the head end, so to speak, to supply an incremental 50-megawatt load, but not in the timeframe he needed.
So, it was very clever. The utility said, “We can upgrade the substation to serve that site, but it’ll take us two years or three years.” Musk apparently offered to buy the substation, upgrade it himself, and then give it back or sell it back to the utility—I think he may actually be giving it back, but I’m not sure of the details. He said he could get it done in one year, which he will.
During that year, to your point, they brought into the parking lot—this is a refurbished manufacturing site—a clutch of natural gas turbines and, I think, some gas engines to produce the 50 megawatts. And you can do that in months, not years. Boom, Bob’s your uncle, as they say in England. You’ve got 50 megawatts on the way to 100 megawatts in months.
I’ve got to say—this is not me patting myself on the back—but I’ve been on the circuit preaching the gospel of electric demand from data centers. Not EVs, but data centers. Data centers use about 10 times more electricity than all the world’s EVs combined. So this is a big vector for electric demand because it’s a massive infrastructure. It’s an unprecedented infrastructure at scale.
It’s not like we just started building it. We started building it 20 years ago. My book, The Cloud Revolution, maps this out, and I made the obvious forecast that a lot more of it’s coming—and it is, because data centers are the core engine of the cloud.
Inference—not calculation, but inference—is different than calculation. We infer a best route to take a trip in your car, and that information is all mediated and moderated by networks connected to data centers.
But this was predictable. What’s going on is that we weren’t building enough electric capacity to meet growth in demand as we repatriate factories, as we build data centers, and as we add EVs. And there’ll be lots more EVs. So what the heck went wrong? Utilities—this is not me being some soothsayer genius who could see this coming. It was obvious. Anybody who studied it—others, not just me—were studying this and pointing out that this was going to happen.
Yet the utilities have been living in la-la land, pretending they can supply this magnitude of baseload demand with episodic power—wind and solar. Episodic power. And it’s expensive. This trope that they’re cheaper is just silly. Yes, when a wind turbine is operating, it produces cheap electricity. Obviously, it doesn’t operate all the time. So the actual costs of the turbine are its system costs to allow the electricity to be delivered when the market needs it. And with data centers, that’s 24/7.
Roughly speaking, the real cost of electricity from a wind turbine or solar is triple its nameplate cost. In other words, to make the electricity available when the markets need it. We know that’s true, by the way, because Germany did the experiment for us. They built a massive increase in their grid—they doubled their grid size, primarily with wind and solar—kept the original grid (about 80% of it), and they doubled the grid to meet a 10% increase in electric demand. That’s what they got in Germany, because they grew more slowly than us. And they tripled their electric rates.
It’s just really borderline insane. It was predictable. It’s what happened—it’s what we’ll end up doing. We’ll follow Germany if utilities continue to build out just wind and solar to meet baseload demand. And they’re not going to do that, by the way—that bubble’s already burst.
So, man, it’s good. It’s good for the economy—it’s good for markets.
Jim Puplava:
I’m glad you brought up the cloud, which is used for storing data of all sorts and accounts for well over 90% of the forecasted demand for power, which comes from the things that we do since the internet. You know, we’re on the internet. I don’t think individuals understand that when they have a smartphone in their hand. It doesn’t—you know, it doesn’t take much, as Apple says, to charge that phone. It’s what the phone does and what’s required on that phone.
Mark Mills:
Right.
Jim Puplava:
And here’s the thing: it has to come from power that is consistent, 24 hours, seven days a week.
Mark Mills:
Correct. Yep. Well, it’s easy to think about what’s really going on. You’ve got this magic device in your hand that doesn’t take much electricity itself—it’s like a nightlight’s worth of electricity in a child’s room. It’s trivial. But obviously, everybody knows that it has networks that connect it to other phones, data centers, and businesses.
The data centers don’t just store data; they manipulate data. They do all the artificial and non-artificial intelligence computing that goes on to answer a query. You go online to do e-commerce, to shop online—the website is stored in the data center. But it’s not just that the data is there; it’s that when you do a query, you’re asking a question, which requires computation. Increasingly, this involves AI computation to curate your query based on who you are, what you’ve bought, and what you’re used to buying. It high-grades the responses in a way that’s more useful, which is why AI is so fascinating—because it’s a semantic interface, better than just pure typing.
But that takes a lot of computational horsepower. And then it has to send that information back to you at the other end, which requires a network. The networks themselves collectively use as much electricity as the data centers—they’re roughly equal. So, if I prorate the amount of electricity that your phone uses, based on your pro-rata share of the data centers and networks, your smartphone is using more electricity than the refrigerator in your kitchen.
Jim Puplava:
You know, another wild card is reshoring—where all that excess power is going to be needed because we want to bring factories back. A good example is the CHIPS Act, which is subsidizing chipmakers to build factories here, whether it’s Intel or someone else. So here’s another aspect that’s coming in that’s going to demand more power. All of this seems to be converging almost at the same time.
Mark Mills:
Yeah, well, it is. Who doesn’t want to repatriate factories? I mean, we’re not going to get all of them here, but let’s at least get some of them back. I don’t agree with the CHIPS Act methodology, but I agree with the philosophy—we should have more factories in America. I’m bullish on that.
The way you get more factories in America, by the way, is not by giving money to the factory builders. It’s by stopping punishing them for building factories here. If the beatings stop, you’ll get more factories. By beatings, I mean the regulatory beatings.
The National Association of Manufacturers publishes a study of the per-employee cost of regulatory compliance in the United States. It works out to over $30,000 per year per employee—just for regulatory compliance. Think about this: if we follow Elon Musk, Vivek Ramaswamy, and others and clean up regulatory burdens—this is what President Trump wants to do—let’s just say, for the sake of discussion, we cut that in half. Well, that’s free money to the economy. We don’t have to subsidize something. We’ve now saved $15,000 to $20,000 per employee in a manufacturing firm if the firm reshoring here. That’s better than a gift of money, in my opinion.
And also, don’t put your finger on the scale of which kind of factories come back. Let the market—the businesses—figure out what kind of factories we want here. But make it cheaper and easier to come back, and it’ll happen.
But to your point—plus EVs, plus data centers, and plus just organic growth—the economy gets bigger because we’re more economically productive again thanks to AI and information. You get growth. Growth drives electric demand—it always has, always will—and we need to plan for it. Basically, the planners have failed because instead of planning, most of them have been engaged in aspirational, wishful thinking about what kind of demand they’d like, as opposed to what kind of demand the market would like.
Jim Puplava:
You wrote an article equating these new data centers to "digital cathedrals." There are just thousands of these warehouse-scale computers. And the other thing, Mark, they keep getting bigger and bigger, like the skyscrapers that we see that dot the skyline of our major cities. One of the popular tropes of our time is that the digital revolution has dematerialized our economy, delinking economic growth from hardware and energy use. I believe there are, what, a little over 5,000 enterprise-class data centers? How long before we’re going to start to see more power outages?
As we started this conversation, we think of power outages in California. But now it’s spreading across the country.
Mark Mills:
Well, we certainly are stretching the grid’s capacity to be reliable—not only by adding more demand but by making the supply episodic and unpredictable. In technical jargon, it’s called "non-dispatchable." You want power plants on grids that you can turn on—dispatch—when you need them.
You can make up for some of it by convincing people to shut things off. This is called demand-side management. But that’s a very limited tool because a lot of people can’t turn things off. Many manufacturing processes can’t be turned off. Data centers don’t want to be turned off.
So, here’s what I think will happen: some electric utility regions are already facing higher prospects of blackouts, and they probably will see some rolling blackouts in the coming year or two. It takes a while to fix these systemic problems—they have a lot of inertia in them.
But the net new demands won’t strain the grids in those regions because the demand won’t be built there. That’s what will happen. Growth—data centers, factories—will go where there’s power. That’s what will happen.
So, this will benefit states and regions that already have surplus power and have the capacity and willingness to add it quickly. There are a few of those. They tend to be in the so-called red states and southern states. This will accelerate the deindustrialization of the blue states and the economic expansion of the red states. I think that’s net-net a bad thing for the country, but that’s essentially what we’re doing.
Jim Puplava:
You know, that’s a big issue for us in California. California shut down a nuclear power plant for Southern California, and now we’re buying excess power from Arizona. Arizona has two nuclear plants there. But Mark, there are three factories going up—Taiwan Semiconductor is building one of the largest plants in the world. You’ve got Intel, and you’ve got Apple. When those three plants come online, I don’t think Arizona is going to have excess power to be sending it to California.
Mark Mills:
That’s exactly right. Good for Arizona. They’ve got the excess power, so they’ll keep it for themselves and let California figure out how to build power plants again.
Mark Mills:
That makes sense. I mean, they’re building lots of wind and solar in California, along with lots of batteries to mediate the handoffs. But the batteries are not big enough, sufficient enough, or affordable enough to fill the gap when there are wind droughts or overnight demand. They can do an hour or two of transition handoff, but what you really need is half a day or a day’s worth of electricity storage or backup power, which you can’t do with batteries. That would cost trillions of dollars.
Your point about building the chip fabs in Arizona is a really good one. I did the calculation this way just for fun: if you build a billion dollars’ worth of electric vehicles, that will induce an electric demand of about $200 million worth of electricity over a decade.
If you build a billion dollars of chip fabs, you’ll induce $300 million worth of electric demand over a decade. So, if you look at the numbers for those chip fabs, you can do the math here, right? Those are multibillion-dollar fabs. There are billions of dollars of electric demand being created by building those chip fabs.
By the way, data centers—if you do the same metric—if you build a billion dollars of data centers, you induce about $600 million of electric demand over a decade. You can see why the numbers are so heavily skewed for electric demand coming from data centers and chip fabs compared to EVs.
Jim Puplava:
When it comes to energy, Mark, we think of the gasoline that powers our cars, trucks, and planes or heats and cools our homes. But we seldom give thought to the enormous power, as I mentioned earlier, required by the various electronic devices we use every day. Our phones, desktops, iPads, smartwatches—all of these have become part of modern life. But what they all have in common with the things we’ve been talking about is electricity.
Mark Mills:
Yeah, well, electrification of our economy has been going on for over a century, as everybody knows, since Thomas Edison famously built the first electric power plant in New York City. And that trend is continuing. The world continues to electrify. That’s the most important marginal source of energy for the world because it fuels things like data centers and all the other new technologies we rely on.
For example, lasers are not possible without electricity. MRI machines are not possible without electricity. CAT scans are not possible without electricity. Electrification has enabled entirely new classes of technology. And much of the world still doesn’t have access to those technologies. So, they want a lot of electricity too.
This brings us back to the underlying energy demands. They want electricity to be cheap, and the cheapest forms of electricity globally remain coal and gas—or hydropower, if you happen to live near big rivers that haven’t already been dammed. But by and large, the world has exhausted its appetite for damming up big rivers. There might be a few left that countries will dam up.
So, the electrification story is not over—especially globally, nor is it domestically, as we’re already seeing with the so-called surprising demand from AI and data centers.
Jim Puplava:
I want to talk about another issue you’ve written about, which you’re calling a political earthquake: the energy transition. The Inflation Reduction Act—which is misleadingly labeled—was really the Green New Deal. It’s turning out to be a boondoggle, anywhere from $6 to $7 trillion in grants, subsidies, mandates, and rules. It’s strewn with waste, abuse, fraud, and corruption. I think of Solyndra during the Obama administration, and I think you had a figure that it’s costing between $23,000 and $32,000 per incremental EV sold. And then you talk about all these NGOs that are receiving billions of dollars. This is probably the most expensive effort ever made in the nation’s history to try to restructure the entire U.S. economy.
Mark Mills:
Yeah, I’d say it’s probably the most expensive effort. I’m looking for someone to tell me that I’m wrong. It is the single most expensive piece of legislation—not just industrial legislation—ever passed by Congress.
The direct costs entailed by the Inflation Reduction Act are somewhere between $2 and $3 trillion. That’s been properly costed out by a number of organizations. These are subsidies that go on forever—they’re not limited in time or magnitude. It’s larded with all kinds of grants and giveaways. It’s full of all kinds of social programs.
In addition to the direct $2 to $3 trillion, it induces the need to spend another $3 trillion. That is, it forces electric utilities to build more infrastructure to deliver electricity to EVs specifically. If this actually happens—which I don’t think it will—that would cost another $3 trillion.
So, that’s $6 trillion. For context, World War II, in inflation-adjusted terms, cost the United States $4 trillion to prosecute. So, this is a crazy amount of money to replace stuff that’s already working, stuff that’s cheaper—namely the hydrocarbon, or so-called fossil-fueled, infrastructure of the U.S. economy.
So, what will happen? Well, they’re spending some of it now at epic levels. Government spending, when it’s bloated, is full of waste. We know that. Some of it is unavoidable; some of it is avoidable. When you add this much spending, you also add waste. It’s also full of fraud—that’s human nature.
I’ve said—and I’m not the only one—the obvious thing is, if you add this much spending, you’re creating an unprecedented attack surface, let’s call it in cybersecurity terms, for both waste and fraud. This is an epic opportunity for abuse and fraud given the magnitude of money allocated by the administration.
It’s going to be really hard to stop this juggernaut of wasteful spending because, sadly, this is the moral hazard problem: everybody’s benefiting from it, in both parties. Republican states and Democrat states are both getting this money from the Inflation Reduction Act gusher.
And you said it correctly: this is not an Inflation Reduction Act. This is an Inflation Creation Act entirely aimed—almost entirely aimed—not at climate. Nobody’s spending money on climate research. This is money being spent on energy, on machines, and on people to advocate for those machines. It is a staggering, unprecedented, colossal misdirection of capital. And it won’t end well if it keeps going.
Jim Puplava:
Where are we seeing this in our nation's national debt? We've gone—I think we’re about $36.4 trillion in debt, and it’s growing by the minute. You know, one possible solution would be nuclear power. But the current regulatory agency seems more bent, Mark, on shutting them down than building new ones. I’m looking at China—they have 55 plants, 23 under construction. India has 23 plants, 11 under construction. We have 92 with one under construction. If the U.S. wants to maintain its lead in AI and technology, how is that possible without nuclear power?
Mark Mills:
Well, let me stipulate: I’m a nuclear bull, and we need to revitalize the U.S. nuclear industry. We created the industry, in a sense. I mean, to be fair, the British were collaborators—they actually had the first commercial nuclear plant operating before the U.S. did by about a year or two back in the ’50s. But the U.S. was the one that created the basic infrastructure and the technologies that are widely used by the world—the light-water reactors.
We’ve got to revitalize that industry. It’s critical to have for the future. We need lots of options—especially nuclear—because it’s the most interesting phenomenology in energy fields ever discovered by humanity. It has the smallest physical footprint, smallest material footprint, smallest air emissions footprint of any form of producing electricity, including solar and wind, by huge, huge margins.
But to be clear, we can build lots of data centers with AI without nukes in America because we’ve got lots of gas. We’re blessed with massive amounts of natural gas—we’re the biggest gas producer in the world. So, we’re not hobbled in that sense.
However, we are hobbled in the strategic sense. You never want to have any energy infrastructure overly dependent on one form of primary energy. You want options. Options make sense. Different kinds of power plants make sense in different parts of the country for different reasons.
And we’ve taken the nuclear option off the table. To your point, we have one small reactor being built. We’ve only built two reactors in the last several decades while the world rushes ahead of us. We gave away our nuclear enrichment capacity to, of all people, Russia. Isn’t that crazy?
We’ve abandoned our nuclear fuel reprocessing capacity decades ago. You don’t have to mine more uranium, by the way, to get more nuclear fuel. You can get it through a process called reprocessing. Environmentalists like to recycle, I thought. And the best way to get more nuclear fuel for future reactors is to recycle the used fuel from existing reactors for a century.
So, we need to do something about this. I’m optimistic that Trump will.
Jim Puplava:
Yeah, and especially if they follow the example of Elon Musk, where he built that large data center and then powered it within three months.
Mark Mills:
Exactly.
Jim Puplava:
Mark, I want to move on to my own state of California, which is home to fruits, nuts, and flakes. California has the waiver that we talked about—a special exemption from the EPA and the California Air Resources Board. Under the guise of fuel efficiency and things like that, they’ve come up with just a whole host of regulations that other states are adopting. Unfortunately, the CARB (California Air Resources Board) is made up of probably the worst potpourri of environmental wingnuts and banana greens who make policy with no appreciable science, reason, or logic to any of their mandates, regulations, or edicts.
Mark Mills:
Well, the idea that we can be all-electric cars and all-electric trucks is just prima facie insane. There’s no other word for it. There’s no possibility—in timeframes that are imaginatively useful, that is, a couple of decades—that we can switch transportation over to electric.
There’ll be a lot more electric cars and electric trucks in specialty applications, but the physical infrastructure required—getting the power there, grids—will cost trillions of dollars and take decades to build. The batteries themselves require an order of magnitude—10 to 50 times more mining in the world—to produce the metals and minerals needed to build that many batteries.
So, it’s not going to happen in the timeframes people imagine. We can mine this stuff eventually, but not in a decade or two decades. The world’s mining industry is not expanding that way.
And maybe fundamentally, my biggest objection to what CARB is doing—the California Air Resources Board—is that they are fundamentally just moving the tailpipe. Let’s be clear: electric cars don’t have a tailpipe; as the kids say, “no duh.” But there is a tailpipe to electric vehicles.
The emissions associated with making the electric vehicle are just moved elsewhere. The massive mining machines and all the grinding and crushing of metals and rock—metal-bearing rock—emit carbon dioxide, nitrogen oxide, sulfur dioxide, and particulates. It’s just happening elsewhere—in Africa and in China.
Environmentalists apparently don’t care if it happens elsewhere. It also makes lots of CO2—elsewhere. In fact, there’s enough CO2 emitted in the construction of an electric vehicle elsewhere that there’s clearly only a tiny reduction in net CO2 emissions by not burning gasoline. And there might actually be an increase in CO2 emissions in many circumstances, given the amount of materials and energy needed to make the EV in the first place.
So, the whole thing is an upside-down and borderline insane policy from an economic and physics perspective.
Jim Puplava:
You know, this gets back to the legislative mandates, and I’d like to get your thoughts on this. What is so sacred about the years 2030, 2035, and 2050? It’s like they just pull these dates out of their hat, and all these things are going to take place by then.
Mark Mills:
I’m glad you said they pulled them out of their hat—a good, polite way to express it. They’re made-up deadlines. Everyone knows that.
The idea that there’s some date certain by which we must accomplish something with energy or climate is obviously driven by the climate narrative—that we have some hard deadline against which we must act.
Let me just invoke Bill Gates for a moment. In an interview he gave after a Davos meeting earlier this year, he talked about these deadlines and objectives. He said something pretty interesting that I agree with and that is very true: whatever we do globally—building windmills, solar panels, EVs, all the rest—it will make no difference at all in terms of what the models say the world’s temperature will be in 2050 or how the climate will be altered because of that.
In terms of the models, it will make no difference. No measurable difference.
What Gates is pointing out is that the planetary system has so much inertia in it that these kinds of changes we’re proposing to make won’t make a difference. This is Bill Gates talking, by the way. Scientists know this—it’s not a disputable fact.
So then you have to ask yourself: if that’s true, and doing all these things costs trillions and trillions—hundreds of trillions of dollars over the next few decades—why do it at all? That would be a fair question to ask. Gates did say something else I agree with, which is that we should spend a lot more money on adaptation and resilience. If you believe the world’s climate is going to become more hostile to humans, the most important thing we can do is to protect ourselves from whatever predations are going to occur—through adaptation and resilience.
That takes money. And since we’re squandering money on machines and solutions that won’t change anything, we should repurpose that money. That would be the moral and sensible thing to do.
So, the 2030 thing—when you talk about it in the broader context of what’s trying to be accomplished—makes no sense whatsoever.
Jim Puplava:
I want to talk about President-elect Trump’s appointment of Chris Wright as Energy Secretary. What do you think that will do? Will he get rid of a lot of these regulations and things that have made it so difficult to produce, or even, for example, to do mining in this country?
Mark Mills:
Well, first I have to disclose a bias—I know Chris. I’ve known him for some time. I like him. He’s a great guy. So, I’m pleased and surprised that he wants to take this on. And good for him. I expect and hope he gets confirmed by the Senate. He’s a very smart guy, very rational—he went to MIT, built a great business himself. I think he’d be a terrific addition to the Cabinet and a terrific Energy Secretary.
There’s a lot of things he can’t do, though—things he doesn’t control. Mining, for example, is under the Department of the Interior, where another great nominee, Governor Burgum, will have jurisdiction. I also happen to know Governor Burgum personally. By way of full disclosure, I like him too. He’s very smart, a great operator, and has been a great governor.
So, the mining challenge will largely fall under Burgum’s territory.
But look—every time there’s a change of administration, the new team gets to evaluate what the other team did. Some of what the other team did is useful, and that was true when Trump was president the first time. And you keep those programs, or even if you want to get rid of them, Congress has a lot of say in the matter.
It’s a very interesting and challenging job. Maybe the most important thing to keep in mind about the Department of Energy is something most people don’t know: nearly two-thirds of the Department of Energy’s budget has nothing to do with energy.
It’s the defense nuclear materials production operations to make weapons-grade materials for the U.S. military and the cleanup operations for World War II and post-WWII-era defense nuclear production. That all falls under the Department of Energy, and it utterly dominates the budget.
It’s never made sense to me that it’s in the Department of Energy. Reagan tried to end that, to my earlier point that you can do some things as a president, but Congress doesn’t always support you. Reagan wanted to abolish the Department of Energy when it was just a four-year-old agency, and he failed. He wanted to at least push the defense nuclear stuff into its own agency, and he failed at that too.
I was a young man working in the White House science office at the time, and that was one of our assignments—we failed. We tried to get congressional support but failed.
Perhaps Chris will try to do that. It would be a good thing to do—usefully giving some clarity to where the money’s going. Secondly, it would usefully, in my view, somewhat defang the Department of Energy’s power because it would have less money as a department. The two-thirds of the money would still go to where it’s already going, but it wouldn’t have the same political salience.
Not many secretaries want to give up power, but I think Chris is a different kettle of fish. He might be interested in doing something like that. We’ll see.
Jim Puplava:
Well, let’s talk about something that’s also going to be needed with the things we’ve been discussing—whether it’s data centers, EVs, or reshoring—and that’s raw materials. You want to talk electricity, you want to talk copper, you want to talk EVs, you want to talk lithium, cobalt.
And we have several potential large mines in this country that have really been put on hold by the administration. I think of the copper mine in Arizona, there’s a big mine in Minnesota, and one in Alaska that have just been tabled. It doesn’t make sense if we’re going to do all this stuff that we want to buy all this stuff from China.
Mark Mills:
Yeah, no, it’s a really good point. The United States has not significantly—or really at all—expanded the size of its mining industry in decades. Many decades.
We still mine, but we haven’t expanded mining here. Mining expansion has gone on globally, and to your point, it’s not just that EVs need copper and lithium—obviously for lithium batteries—but also cobalt, manganese, and nickel. Everything does.
Every smartphone, every building, everything we build starts with mining. Even wooden buildings start with mining because of the nails, the metals, and the copper.
So, mining is critical. It’s more than just a little bit critical—it’s strategically critical for the military and for our economy.
We depend on imports for about 100% of 17 critical metals and minerals, and for another dozen, we import more than half.
This is not a good situation strategically, economically, or geopolitically. There will always be a lot of mining around the world because you chase high-grade ores like you do with everything else. Some countries have greater or lower-cost, higher-grade endowments than the United States.
But we have lots of minerals on the continental U.S.
This administration—I hope the incoming administration—will continue to do something about this. This administration has talked about solving the critical minerals strategic problem and economic problem, but they haven’t done anything about it.
The difference between what the Biden administration has done and what the Trump administration says they will do comes down to the regulatory environment. The Biden administration has quietly made it harder to mine and has overtly canceled mine permits. The Trump administration has said they want to make it easier to mine and to grant the permits.
But we need Congress to step in and codify this as a matter of national policy—not just something that presidents can whipsaw back and forth between approving and not approving.
Jim Puplava:
So, let’s bring this down to investments. We’ve been invested in AI, but we’re very bullish on natural gas, uranium, and nuclear power. If you were looking at this and seeing the demand that we see coming from reshoring, reindustrialization, AI, the cloud, or EVs, what do you think are the investment opportunities here? You know, one could look at the markets right now—utility stocks have been performing like technology stocks recently.
Mark Mills:
The trick with investing is—you’re the expert here—but I play one on TV episodically.
I’ve been an investment advisor, and I wrote an investment newsletter, as you know, for a number of years and functioned as a stock picker as well, with some success—a decent track record.
The problem is, things I’m very bullish on—like AI—are already massively hyped. My book, The Cloud Revolution, focuses heavily on AI. I think we’re at the end of the beginning, not the beginning of the end, of the AI buildout. There’s a massive buildout happening.
The markets are wildly enthusiastic, but stocks in that space are probably going to go through the classic cycle of enthusiasm getting ahead of reality. Enthusiasm is often correct in substance but premature in valuation. So, I’m nervous about AI stocks—not because I don’t think AI is a big deal, but because the enthusiasm may have outrun the velocity of adoption.
So, I’m very selective. Personally, I’m interested in the choke-point infrastructure stuff underneath AI. Things that make it possible to power AI itself—both the electrical equipment and the electrical conversion components in the power semiconductor space, for example.
EVs? I’m very bearish on EVs, for all the reasons you can guess. They’re overhyped. The hype is already coming off. The Wall Street Journal recently had a piece showing the epic stock price collapse of all the EV startups—from Rivian to Lucid to—you name them. They’re just not doing well.
Tesla is the outlier because it’s a great car company and it’s more than a car company. They actually make a profit. It’s a profit inflated by selling emissions credits, but they still make a profit. None of the others do, including GM, Ford, VW—all the rest. Elon started much earlier in the game and is way ahead of everybody else. He still is, I think.
So, where are the opportunities? The opportunities are always in what’s undervalued. So, I would say it distills to this:
Electricity is obviously needed. People have figured that out. Nuclear is a tough one—it’s premature on investment opportunities, only because of the timelines it takes to build stuff. But from a stock picker’s perspective, the enthusiasm for nuclear is now over.
Whether it’s AI or nuclear, we’re going to stay very enthusiastic for quite a while, and that may well inflate stock prices. But we’re not going to be able to build many nuclear plants for a long time—many, many years—because we have to rebuild the infrastructure to build them.
We have to first build the new kinds of reactors—the small reactors, so-called small modular reactors—to see how they work, get the supply chains going, modify the designs, and then start building them at scale. That’ll take a decade.
So, where’s the underappreciated, low-hanging fruit? It’s obviously in oil and gas—and frankly, global coal. Coal plants are being built everywhere in the world. People may hate coal, but coal’s not going anywhere.
Global demand for oil and gas—bullish on both. Gas is going to grow faster than oil, but oil has an advantage: the velocity of building the infrastructure for oil is cheaper and faster. So, they balance each other out.
There’s more net marginal demand for gas than oil in terms of growth rates. But the velocity of supply and demand is faster with oil. Oil runs the world’s transportation system.
So, if you think there’s going to be any economic growth ever again—if, say, the world renormalizes in a couple of years and we have growth again, and China starts behaving and grows again—that drags transportation.
And 97% of all the world’s movement of goods and materials is fueled by burning oil. And 100%—not some small percent, but 100%—of all materials, goods, and services use oil somewhere in the supply chain.
Oil is the biggest traded commodity in the world. It’s the largest single source of energy for the world. It’s not going away. So, to the extent that high-value oil companies are undervalued, I’m bullish on oil. And for the same reason, I’m bullish on natural gas.
Jim Puplava:
I couldn’t agree with you more because that’s exactly where we’re invested. And if you look at a market that is, in many ways, overinflated—as you mentioned, especially some of these AI stocks—you look at oil and gas stocks, and they’re selling for a low fraction of what they should be.
Mark Mills:
Exactly.
Jim Puplava:
Mark, as we close, if our listeners would like to follow your work, tell our listeners about your Energy Institute and the things that you’re doing there.
Mark Mills:
Sure. I founded this new national think tank called the National Center for Energy Analytics. It’s easy to find through Google—just type in either my name, Mark P. Mills, or National Center for Energy Analytics.
I also have a website where I maintain a canon of all my writings, speeches, and reports. It’s called TechPundit—techpundit.com. It has my books, what I’ve written, and my most recent book. You know, the usual stuff—testimonials and podcasts I’ve done and TV shows.
Pretty much everything I’ve published is there. Some of it is paywalled because I write in the Wall Street Journal frequently, and other places that are paywalled. But I’d say 80% of what I’ve published is not paywalled—well, not by me, but by the publications.
I have a number of things I’ve written recently about AI that might be of interest to your listeners, and they’re on my website. It’s all about the magnitude of demand that’s still to come.
I also do the Twitter and LinkedIn stuff to comment on what’s going on in the world. It’s easy to find me by typing in my name—Mark P. Mills—on either Twitter or LinkedIn.
Jim Puplava:
And one more thing—would you tell our listeners about your book on the cloud, which I thought was very prescient when you wrote it?
Mark Mills:
Yeah, well, it’s only a couple of years old now, which is nice. It’s called The Cloud Revolution. It’s still available on Amazon.
The subtitle—I won’t give you the long version—the bottom line is that it’s about “the roaring 2000s.” I think the 2000s are starting to roar again, and they’re going to roar.
I lay out why. I cover more than just the cloud, although it’s the central thread of the book. I write about the contemporaneous revolutions in material science and basic research and their implications for everything from education and healthcare to entertainment and industrial operations.
It’s fundamentally a technology forecasting book. It got to number one in technology business forecasts a number of times in the Amazon ranking system. I think it’s still very relevant because it’s not about short-term forecasts. It’s about a decadal cycle.
Jim Puplava:
I couldn’t agree more. Well, listen, Mark, it’s always a pleasure speaking with you. I hope we can do this again in the future. And I want to wish you happy holidays and a healthy new year.
To learn more about Financial Sense® Wealth Management, give us a call at (888) 486-3939 or click here to contact us.
Advisory services offered through Financial Sense® Advisors, Inc., a registered investment adviser. Securities offered through Financial Sense® Securities, Inc., Member FINRA/SIPC. DBA Financial Sense® Wealth Management. Content is for informational purposes only and does not constitute financial, investment, legal, or other advice.
