全球最伟大的能源交易员谈市场、中国与 AI

My guest today is John Arnold. John is probably the most famous energy trader of all time and certainly the most successful. One of the things John says is that he wanted to cultivate and build the best seat in his industry—the seat with the best perspective, the most information, and the best systems.

What's most interesting about John is that, after being the most successful energy trader of all time, you could argue that he's gone on to be the most innovative philanthropist as well. John has applied this idea of philanthropy to all different sectors, and what's so exciting about this conversation is that it feels like you're talking to a talented entrepreneur or a talented operator in all of these different fields who's willing to share exactly what he and his team have learned about what makes certain problems manifest across our country.

1. China’s Rapid Transformation

He has an incredible perspective not just on the things he's worked on, but on his travels, on companies, and on technology. This conversation is a reminder to me that cultivating the seat is such a powerful concept that we probably underinvest in. Please enjoy this wide-ranging discussion with John Arnold.

Everyone listening, I think, is somewhat interested in what's going on in China. Your lessons that you shared were so interesting, and I'd love you to share some of the highlights from that trip and what you learned traipsing around, studying robotics, studying AI, and everything else that you did. What were your major takeaways from that trip?

The origination for the trip was this realization that China has gone through a transformation unlike any other that's happened in the world, both economically and culturally, in the course of 30 years. This is a country that has gone from really trying to replicate the West to, in many ways, leapfrogging it.

I was fortunate enough to spend a week running around, meeting with 4 or 5 companies a day that were all very open, and getting to tour factories. It was really striking. I came away incredibly impressed and also with a lot of questions about what happens to the rest of the world as this rise of China is happening.

The 1 big takeaway was just the speed and scale at which they can do things is unlike anything in the world. They have this highly educated population. It's a very entrepreneurial culture. They've figured out how to get capital to these companies, and they have this deep domestic market.

That allows them to build up these supply chains and agglomeration effects to create something that I don't think exists in the rest of the world. I was talking to a battery company and asking them if they were looking at replicating factories elsewhere in the world, and 1 of their responses was, “Every 1 of my suppliers is within 200 miles of here, and I can call them and meet with them the same day.” You can never get that.

Then there's the scale of labor that they have and the flexibility. If you need 1,000 workers tomorrow, you can get that. Again, it is a skilled workforce that's still very hungry, that's oftentimes coming from poverty. Having this factory job, which in the West might not look very appealing, is highly appealing for many people in China. It's just the first step up.

2. Lessons from the Chinese EV Market

This combination of things has allowed them to create this competitive force that the world is suddenly reckoning with over the past decade. It's really interesting to think about what's the right response from Western countries that don't want to cede their markets to China for obvious reasons. There's a lot that China can offer, so what's the right symbiotic relationship that the West should be having with China? I think this is a massive question that policymakers are grappling with, and I wanted to try to dig into it.

Which company visit was the most illustrative of these big lessons? Looking back on the trip, what company visit taught you the most?

I've been fascinated with the EV market in China, which about 10 years ago was deemed to be a strategic market. The way that the story is told is that China realized it was behind on internal-combustion-engine cars and would always be playing catch-up, and that if it could leapfrog to the next technology, then it could have an advantage over the rest of the world.

Right now, there are apparently over 100 different manufacturers of EVs in China. Many of these have a contract manufacturer, so you can send in your design and do your branding independently of the manufacturing. But there are many manufacturers there that actually have their own plants.

We went to 1 for NIO, which is kind of the upscale Chinese car in the $40,000 to $80,000 range, although they've recently released 1 that's below $10,000. Going to their factory and seeing, number 1, how quickly that factory was built—from the first shovel in the ground to the first car coming off the line in 17 months—was phenomenal.

Second, there was the factory automation that they had. There were certainly people on the assembly line, but much of the process was done via robotics. You start looking at the auto plants we have in the United States. We still have a plant that was originally built 100 years ago operating outside of Chicago. The average age of U.S. plants is a little hard to get, but we're roughly at 40 years.

3. Robotics

They've been upgraded over time, and there are certainly robotics in American car manufacturing, but I don't think they replicate what's happening with robotics in China. They have this combination of being able to do things really quickly with a skilled but low-cost labor force, and then adding on the robotics has just allowed China to create a quality product at a price that nobody in the rest of the world has really been able to figure out yet.

Do you know anything about robotics or AI that just surprised you?

The number of robotics companies—there are over 100 now in China. As I understand the process, in each 5-year plan, China specifies certain industries that are deemed strategic. Then the head of the province gets evaluated on a number of factors.

The head of the province is selected and not voted on, so the evaluation is based on things like employment and GDP growth, but also whether the industries being created there are aligned with the 5-year plan. There are also some subsidies coming down from the top on those favored industries.

Robotics is certainly 1 of them. Each province takes a couple of companies that it favors and gives them subsidies and support to try to get the winner, or 1 of the winners, to be in their province. Then it gets the supply chain to develop around them, along with all the associated jobs and GDP.

You get this massive competition. One of the end results is that most of these companies aren't very profitable today because there's such intense competition and, a lot of times, overcapacity because of the province-level subsidies and support. But this intense competition, which is termed “involution,” also creates better technology, right? If you're faced with that type of competition, to be 1 of the winners, you have to be fantastic.

I think the question that China has is what you do with those who are not the winners. Do you have a process where the losers stay in the industry and keep everybody unprofitable with the overcapacity, or are they closed?

China has started this new process of anti-involution—trying to support the winners and make sure that they can build up to be healthy, strong companies and be really global competitors, rather than just being brought down by this overcapacity in the domestic market.

Obviously, there's tons of misunderstanding about what the state of China wants and is trying to accomplish. But if you add up all your conversations with all the people that you met across the week, how would you sum up what just the people that you met and interacted with wanted? What did it feel like they wanted and were trying to accomplish, and what was their attitude toward us?

It seems increasingly adversarial from our direction toward them—skeptical, worried about it, and so on. But just adding up your conversations, what was your felt sense of what the people wanted, were trying to accomplish, and felt about the U.S.?

I was struck by just how much the 2 countries have separated since 2019. The number of flights between the 2 countries is down 70%. I talked to a couple of expats in Shanghai who said that the number of Western expats was down 50% to 75%. The number of American students studying there is down 90%.

I started to push on this a little bit about what was happening. Part of it was that when China was starting to develop and trying to copy the West, 1 of the ways to do that was to bring over Western expats and teach them Western business practices, along with all the aspects of how to run businesses and capital formation and allocation.

Those roles and those learnings have now been domesticated. Instead of firms paying multiples of the cost to bring over a Western professional to do it, those learnings are now domestic, so it's cheaper and they have those skills. They don't need the West anymore.

That was 1 of the big senses I got: this confidence that's building there. They used to try to just copy the West. Now they're world leaders in many of these things. They don't need the West coming to teach them things. They're going to teach the West.

I'd love to rewind the clock a little bit. I mostly talk to investors on this show. I have rarely talked to some of the world's great traders through history. At a point in your career, you were probably the greatest active trader, or one of them—certainly the best in your market. I'd love you to describe what it's like—what it takes to be truly excellent at that specific discipline.

If you think back on the time when you thought you were sort of at the peak of your powers, I'd love you to take us behind the scenes a little bit of what it took for you personally to get to that stage, since I talk to so few people who have done this.

I might step back and think about my kids and what I think—or what I want my kids to do. One of the big components is: do something that you're really passionate about, that you want to do, that is not a job that you do for income. But if you can have a profession that you love, that you have this real, deep passion for, that's what I found with trading. I just loved the battle, the puzzle, the game of it.

I would sit there from kind of 6:00 in the morning to 6:00 at night at the desk, either staring at the computer screen or doing some analysis. I would then go out with people from the industry that night and dream about the industry. In the shower in the morning, I'd be thinking about it. There were negatives associated with just being so locked in. I'm not sure I was a great person, a great friend, or a great partner during those times, but I think just dedicating your life to this craft for a time period was one of the things that separated me.

There are consequences to doing so. It's not the healthiest lifestyle. It's not healthy from a relationship standpoint or a physical standpoint, and I think it can be mentally exhausting. After doing that for 17 years, with this one thing and being so deeply intertwined with it, at some point I just had to step back.

If you think about the second- or third-best natural gas trader who was active when you were active, I'm so curious about the difference between you and them. Another way of asking—if you turn it into advice for your kids, which I like that frame on it—what does it take to be number 1 in something versus number 2 or 3?

I don't know if I thought I was the best. I felt I was among the best. I had also managed to create probably the best seat in the industry.

Say more about that.

4. Leveraging Scale and Proprietary Data

My first job out of college was at Enron. I started there when I was 21 years old in 1995. It went bankrupt in late 2001. So then I'm coming out and I'm trying to figure out what to do. I had a bunch of options. One is I could have stayed with the Enron trading floor, which ended up getting moved to UBS. I decided I wanted to do something that was a little bit more entrepreneurial and ended up deciding to start my own hedge fund.

I had the classic hedge fund 2-and-20 structure. Now there wasn't an intermediary between me and the 20%; I was getting the 20%. So I had the best, or as good of economics as almost anybody in the business. Then we had very good financial returns early on. So that did 2 things. Number 1, there was a lot of retained earnings and new investors came in, and so we had a lot of risk capital. Number 2, we had a very good investor base that trusted the team I had built.

Whenever we had a down month or a down time period, which we had, they weren't calling to redeem, but they would call up and say, “Do you need more capital?” because we had earned their trust and we had done forced distributions along the way. So I'd set up this really powerful seat where I had good economics, so I could hire the best people in the business that I knew of. We had a lot of risk capital and a very solid and stable investor base.

5. Lessons from the Baseball Cards

I like this concept of developing the best seat in your industry. Are there any other interesting or useful components to that for someone who's listening and is an investor, or even just a normal founder or business person, who's thinking about setting up the best structural advantage in a seat? Is it information flow? Is it other stuff? I love this concept.

We started at 2 and 20. We did very well and had a lot more demand from people to invest than we had capacity. We raised the fees over time and ended up at 3 and 35 by the end. That allows investment in the business.

You can build up a fundamentals team that's the best in the business because you can just pay them more. You can go do side projects, like trying to get proprietary data sources that give an advantage. You can pay the trading team and the middle and back office more money. So everything in the business starts to work, and you have a sense of excellence around the firm. That makes everybody better.

We were able to develop a proprietary trade-entry system and a position-management system. These are things that you don't necessarily think about but are extraordinarily useful: being able to buy any and all data, come up with proprietary data sources, have the best people trying to translate that raw data into something useful, and build the best fundamental models in the business. I think all of that becomes part of the flywheel.

It's really about scale. It's the redeployment of your scale that let you do all that and get into all these fine-grained ways that you could be better than your competitors.

Yeah, I think that's right.

Can you tell me about the baseball-card business that you built in high school? I'm curious about it because I always wonder where the initial spark comes from. You mentioned the importance of passion and that you just fell in love with energy markets. I'm curious to trace why you were entrepreneurial in the first place, what the formative things were, and maybe where a chip on your shoulder came from, to the extent there was one. In hearing about that business, I'd also love to hear the story between that and the discovery of passion for the energy market specifically.

I was always entrepreneurial. I always wanted to make a dollar. As a 13-, 14-, or 15-year-old kid, you kind of have limited options, and I didn't really want to work retail because you weren't going to make very much money doing that. I was in high school in the late '80s and early '90s, and this baseball-card boom was happening.

I remember, even in middle school, first getting started and being introduced to baseball cards. It was when they were really starting to take off, around '87. It became clear to me—I guess in retrospect—that this was this really interesting kind of financial instrument: very volatile, with information on pricing that was not very uniform across the market, and a lot of geographic price differences happening.

6. Trading Natural Gas and Market Dynamics

I managed to talk myself onto this bulletin board of baseball-card dealers, which was when internet commerce was just getting going, and they had this wholesale system that was kind of real-time pricing, in effect. You'd have people in New York who were trying to buy hockey cards and people in Texas who were trying to sell their hockey cards because there wasn't much demand for them. There were these arbitrage opportunities that got created.

I'm like, I kind of know this from this bulletin board: people in Montreal, New York, or Buffalo are willing to pay X, and I can buy these at Y and make that money. The sports-card business kind of really started to boom, and so there was more and more money coming into it, more and more cards. I was living in Dallas at the time but would travel to Houston, Louisiana, and Oklahoma.

I started getting on planes when I was 16 and going to big card shows around the nation. I always had this sense of the value of what something was worth at the time. It turns out there were a lot of similarities between doing that and trading.

I was, in a way, doing market making in sports cards. I was taking advantage of arbitrage opportunities in sports cards and knowing what every product was worth at every moment. That kind of stuck with me, and that was my mantra when I was trading gas: I knew what every month was worth better than I think anybody else did. I knew that every moment of the day.

To do that, it took an intense focus—just sitting there, staring at the screen, and listening to every trade that happened in the market all day long, every day. It takes that work and intensity.

Can you describe the actual instruments that you were trading? When you were trying to put positions on and take them off, what were those positions, and what were they in, on average? I want to get into some of the nitty-gritty details just to give people a sense of how far you had to go to get an edge in this world, as an analogy for how far you have to go to get an edge in anything.

You enjoyed this, but what were you literally trading in those days?

It was primarily futures and swaps that were kind of a look-alike to the future. I traded natural gas. The hub for natural gas trading was Henry Hub in Louisiana.

There would be price differentials between what gas was worth in Louisiana versus gas in Pennsylvania, West Texas, or Colorado, for instance. That's basis. I did that for a small part of my career—those price differentials between one point and the hub in Louisiana.

But most of my career in trading was trading what we call the fixed price of natural gas. If you're watching CNBC and on the ticker it has a gas price, that's what I was trading.

In a given day, how much volume and activity would there be? How many key decisions would you make in a given day?

It changed over time. For a long time, I was the largest market maker in the business, which I did both because it was profitable and, maybe more importantly, because it allowed me to put on and take off positions with lower slippage and with fewer people knowing what my position was.

It also gave me insight into who was doing what in the market, so I could start to build the psychology of the market. I could see certain traders positioning themselves in certain ways and try to reverse-engineer what their thinking was, which helped me figure out how I wanted to position myself.

Maybe for those who are less initiated in the act of market making, just describe what it does—how you're provisioning liquidity for the market, but also why you were able to cover some of your tracks by being that player in the market.

7. Innovation in the Modern Energy Sector

A healthy financial market exists for a reason. That reason is that there are commercial players—in this case, producers or end users—who have exposure to the commodity and are willing to pay something to have somebody else reduce their risk to the commodity price. They will pay something to the market for that risk management.

If you think about a producer of natural gas whose revenue is almost entirely based on the price of natural gas for a given month, or even day by day, there is this huge boom-and-bust cycle that can happen because commodities naturally have this boom-and-bust cycle. Businesses that are exposed just to the spot price, or today's price, also go through a boom-and-bust cycle.

Many businesses will choose to hedge out, or fix, some of their forward revenue and reduce their exposure to any short-term swings in the market. That's the demand for hedging. If there's demand, there has to be a supply. There has to be somebody who's willing to take the other side of that trade. Somebody has to warehouse that risk, price the risk, and manage the risk. In aggregate, that's what the traders as a whole have to do.

8. High-Level Goals of the U.S. Energy System

There are also speculators that come in. The people who are warehousing that risk have to price that risk, so they become experts in pricing it and figuring out how to hedge it. There is often speculation that happens on top of that, as well as speculators who aren't actively market making but just have a view on natural gas, oil, gold, or anything. They come in and want to put on a position, and again, they need liquidity. Somebody gets paid to provide that liquidity.

It strikes me, as someone who looks at a lot of young companies, that energy is exciting. Maybe it's just because of data centers and AI and the new demand that's coming on stream for power. Does it feel that way to you?

I know you've looked at tons of companies, and I'm sure you're incredibly on top of energy markets today. Do you feel like it's a good time, if you're interested in this field as you were, to go into it relative to 10 years ago, when it seemed a little bit quieter?

I think it's more on the asset side today, but I think there's an enormous amount of innovation happening in energy assets, whether that's in new technologies like batteries, geothermal, or advanced nuclear.

I also think about all the data center development and the amount of money and capital that's in those fields, and the need for innovation. If you can make things even a little bit more efficient, there's enormous value created that a new company or a new individual can take advantage of.

One of the characteristics of the energy industry broadly is that it is enormous. If you can have a little edge in a niche of the energy industry, there can be tremendous gains to be had. Certainly, this mad scramble to build data centers and power those data centers is creating enormous opportunities for new entrants to the market.

If you're describing the industry today to someone who knew nothing about how the U.S. energy system works, what would be the high-level way that you would approach explaining it to somebody new? Here's the state of things today.

Maybe stepping back, what are the goals of the energy system? I think about systems a lot these days—what the incentives are, what the rules are, and what the goals of a system are.

In energy, you can think about having a product that's affordable to the consumer and reliable, so that whenever you want it, you can access it. You can think about times when there have been gas lines, like in the 1970s, or blackouts, when reliability hasn't been there.

You can think about having reduced emissions and a cleaner system, as well as energy security. Maybe fifth, I'd say that it creates good jobs for whatever country or locale you're in.

If you have those as goals, then how do you meet them? America is blessed with tremendous energy resources. It has a lot of oil, gas, and coal, as well as a lot of wind and solar resources. We're also a very innovative country, with the ability to deploy solar and wind and to think about both traditional nuclear and advanced nuclear.

We have all these inputs, and we're trying to get to a system that meets those goals. The challenge is figuring out how to devise policy and rules for the system to do so. That gets tricky because the priority or ranking of those goals changes from administration to administration.

It's an industry that is slow. It takes a long time to build the infrastructure, and you need stable supply chains in order to meet the needs. You have these changing goals, so every 4 or 8 years, the industry gets sent a different set of priorities or price signals about what we want as a society. Then the industry has to scramble.

Then you throw data centers in here, which is just load growth, with data centers being less concerned about price and more concerned about speed than any consumer of energy we've seen in this country, maybe ever. Put it all in the mixing bowl and try to get a system that works.

If you could imagine a U.S. energy future, I'd love to paint the poles and have you imagine the best possible scenario you can for us 10 years from now. What would the components of that system be that are different from how it is today? And maybe the worst possible version of the system that you can imagine in 10 years, just to get a sense—having studied the system so carefully, I love your way of thinking about the competing goals and this weird data center variable as well.

Maybe starting—well, let's start bad. What would be the worst scenario that you can imagine that would get you concerned about the state of the U.S. energy system in 10 years?

The worst scenario is that the energy system becomes the bottleneck for both U.S. innovation and individual flourishing in this country.

9. Overcoming NIMBYism

On U.S. innovation, especially in technology and AI data centers, it's largely about whether we can build and create the supply that industry is demanding. On individual flourishing, it's whether we can do so and maintain the affordability and reliability components that Americans have come not only to desire but to demand.

There’s this politics around energy, food, and housing because we need all of those, right? Those aren’t optional for people. But there’s this politics around it where, if you start to lose affordability of any of those, there are real, tremendous political ramifications.

And so you can start to see that if an energy system goes bad, it will have a huge impact on the politics of this country.

If you think about that bad-case scenario, both supply- and demand-wise, I’m curious, starting with demand, how certain you feel that this increase—this big increase in demand, really for the first time in decades in the US—will be lasting. How lasting does it feel to you?

Does it feel pretty certain that we’re just going to get a ton more data centers almost no matter what, or demand for them anyway? And then on the supply side, what would be the reasons that we just failed to meet that demand? Why would we stumble and just not succeed?

Visibility through 2030 is pretty clear. The investments are being made. The actors that are making these investments are the largest, most profitable companies that have ever existed on this planet. They’re also, many of them, growing at very healthy clips, and so their free cash flow is increasing. They have the financial capability to execute on these plans, and they’re making the investments today.

I think the 2030s onward—who knows? The assumptions that you have to go into any model on this have such wide error bars, and you extrapolate out that many years, and you just create a garbage-in, garbage-out model.

And what about our ability to meet that demand? What would screw us up on the supply side most?

I think it’s policy. It’s gotten harder and harder to build in this country over the years. And there’s always this tension between what the needs of society are and what the preferences of a local community are.

Everybody realizes we need transportation, housing, and energy, but people also want it not in their neighborhood. This is the whole NIMBY movement: We need affordable housing, or workforce housing; we need transportation; we need energy assets. But don’t put them by me, and if you try to, I’m going to go fight you in the courts.

For developers, time is money. The opponents to projects have gotten very clever about how to use and abuse the existing regulatory laws to delay and delay projects, and in doing so can often kill them.

10. The Challenges of U.S. Transmission Lines

I’ve been very focused, through our foundation, on how to do permitting reform to make sure that the projects that are good for society, as well as not unnecessarily painful to any community, will get built while maintaining the parameters so that bad projects don’t get built.

Just NIMBYism and this difficulty in building are what would lead to the problem of not being able to supply energy, or energy becoming the choke point for development in the United States. And particularly, bringing it back to the first part of the conversation, China and just the speed and scale at which they can build—they don’t have this problem. I’m confident of that.

It is one of the biggest differences between the 2 countries. And I really worry about, if energy becomes a choke point in the US, the real ramifications for the strategic ranking of this country.

If you’re a politician, would you use the China story as fuel to the fire to try to communicate and market this concept that we’re sort of getting lapped by someone that’s sort of our, if not our adversary, our competitor?

100%. Not only is it a good narrative, but it’s true. If energy becomes the constraint, we will become less competitive vis-à-vis China.

If I was an energy asset investor and I really wanted to fund stuff that would help us solve this problem, where you’ve got the joint growth in demand and this rise of NIMBYism, where does this whole thing break down? Is it in transmission? Is that the problem?

11. The Future of Nuclear, Fusion, and SMRs

Because presumably, if we could just produce lots of energy in remote places where nobody lived or cared about it and transport it efficiently, the problem would be solved. Is that the right way to think about it? Do you think about it differently?

I want to get towards the best version of things you could imagine, in 5 or 10 years’ time, that would solve this problem and make us energy abundant and not slow.

Transmission is certainly a component of it. It’s not the only component. We can talk about supply chains; we can talk about land use. But transmission is a huge component of it.

I actually started a company about 5 years ago trying to build interregional transmission lines because it was an industry that had become so difficult to permit and build that private capital had largely given up trying to do new projects. And it’s one of the solutions that can be a win-win-win for everybody.

We’re talking about the goals of the system, right? It can reduce costs, increase reliability, reduce emissions, make the country more secure, and create more jobs. But it had gotten so hard to build these lines—it is still very hard to build these lines—that most of the developers who had gotten started with projects in the 2000s kind of gave up.

They started things thinking these were going to be 5-year projects, and they were 10-plus years in, and many of them hadn’t broken ground yet.

Again, this is part of the challenge of America: It just takes so long. And part of this is, there’s not one entity that can just say, “Okay, do it.” There are multiple veto points along the way, and you have to convince a lot of different actors of this project.

I talked to a lot of politicians, a lot of federal politicians, on both sides of the aisle. One of the things I hear from almost all of them is this need to figure out how to permit and build faster and more efficiently in the United States.

There’s very broad agreement at the top level. You can get disagreements about the details, about how to go about that. I think I remain reasonably optimistic that we can get federal permitting reform done this year.

It’ll probably be the only bipartisan legislation that happens besides maybe a budget. But there is this will that I sense, unlike any other issue in DC right now. There is this realization that this can’t set the country back.

People will look at it for different reasons. People will have prioritized goals in different ways. So how can you do a bill that helps all of those goals? That’s what we’re shooting for.

What do you think about nuclear, both new forms of fission, but also the panacea potential of fusion?

Promising, but very, very difficult. The innovation hasn’t been proven. We can build the AP1000. The AP1000 is kind of the latest of the traditional nuclear power plants. We know it’s a very, very costly electron.

It was the last nuclear plant to be built in the United States, finished, I think, in 2024. Vogal Units 3 and 4 are AP1000s, with a very advanced safety design for a nuclear power plant.

The Vogal plants were enormously expensive. And if anything, they’ve gotten harder to build since then because they require a very significant amount of labor in generally rural parts of this country.

At its peak, Vogtle had 9,000 laborers on site, many of whom were highly skilled craftsmen. Trying to do that today in the United States is very difficult. But we’ve proven we can build that. We know it’s a high cost.

There’s a lot of effort to do advanced nuclear. That’s either the SMR, or small modular reactor—that’s fission—or through fusion. And I think the question is, we don’t know what the economics of either of those are.

For one, we don’t know whether we can build a commercial fusion unit. I think there’s a lot of optimism in the industry. It remains to be seen whether we can get there or not.

I think the greater question is, what are the all-in economics of that? And that’s a big unknown. Because at the end of the day, you’re competing in a commodity market.

We’ve seen that we’re willing to pay a little bit more for a cleaner electron, but we’re not willing to pay a lot more for a cleaner electron. So it has to be cost-competitive.

Optimally, the cheapest electron is one that minimizes emissions. And so I think that’s the promise of nuclear.

These aren’t coming anytime soon. There are some pilot plants being drawn up today. They may or may not have started construction on some of this. I think most of these announcements are kind of for the PR.

By the end of this decade, I think we’ll have taken some small steps there. But we’re probably looking at 10 to 15 years to really have advanced nuclear be at any scale in the United States, and that’s best case.

What do you think about the basket of startups that have been launched and, in some cases, raised a lot of money? What are the prospects for good returns on the dollars that are going into these companies and projects, either today or historically over the last 5 years?

It surprised me. In some ways, I thought that some of these technologies were almost uninvestable from the venture-capital side because of the long duration required to get to a truly free-cash-flow business. I'm surprised at how much money has gone into them. I think it's fantastic.

I think this has to be a public-private partnership where the government is putting real capital into these companies, because I don't think they make sense as stand-alone investments without public support. The time frame to commercial deployment and the amount of money that has to be raised are significant, and you don't really know early on whether you're making progress or not. There are a lot of benchmarks to be overcome.

Until you actually build one, you don't really know what your economics are going to be or what the efficiency of the units is going to be. I worry that there's going to be a falling out in the industry. There's probably too many SMR companies, and it would be better if the industry could coalesce around 3 or 4 different technologies and focus the resources there.

It's a scary circumstance to imagine as an investor when funding dries up and free cash flow is still over the horizon—not even within the visible horizon.

I'm curious if there are other areas where you think that's less true. Solar and batteries seem like a really interesting way of solving some of the problems in the energy markets. The cost curves for both look really cool and interesting. I'm curious what you've learned about those 2 technologies and, generally, what you think about that style of energy generation and storage.

Solar has 2 trends. One is that technology and manufacturing costs keep going down. The main technology is improving, and manufacturing costs are going down, so the end product—a solar panel, mostly made in China—keeps getting cheaper.

A solar panel doesn't get you what you want. You want electrons at the location where the load is. There are a lot of steps to go from a solar panel to electrons in the right location. Some of those are inflationary. The panel itself is deflationary, but it requires land, labor, access to transmission, and access to capital. The cost of capital was falling for a long time and has now been increasing, really over the past 5 years.

12. Data Center Demand

People love to talk about the cost of a solar panel, and usually whenever you see these graphs, they show the cost of the panel. It just goes from the top left to the bottom right. Meanwhile, I know the cost of a PPA, a power-purchase agreement for solar, is well off the lows. The lows happened around 2020, and the cost of delivered electrons from solar is probably more than 50% higher than it was at the cheapest point in 2020.

It's because the panel itself becomes a smaller and smaller percentage of the total cost of the system. Any advantage from continued technological or manufacturing improvements is affecting a smaller and smaller percentage of the total cost. Meanwhile, the inflationary aspects of the system become a bigger and bigger percentage of the total cost.

That being said, there are a lot of ideas about how to bring load closer to where the solar is being generated. Can you use some type of automation, factory automation, or robotics to build the plant—the solar field—in a better and cheaper way that requires less labor? Advancements are being made.

It's good to see that a lot of data centers are now being built with developers or end users signing new PPAs with solar developers. We've seen a tremendous amount of solar come onto the system.

I think it's also true that, generally, in a stable-demand environment, each megawatt of solar that comes on is worth a little bit less. The most valuable solar that came on was the first megawatt. Then you get to a point where the supply of solar has exhausted the demand during the sunny parts of the day in a certain region.

As you bring on more solar, you either have to also bring on batteries or bring on transmission to move it to a different place. That has a cost.

13. Housing Reform

Battery costs have been declining as well. I wonder if some of the same dynamics that happened with solar will start happening with batteries soon. As the technology gets better and manufacturing gets optimized, you start running out of the advancements that have significant cost ramifications for the battery system. It becomes more and more about what the input costs are.

We've seen lithium prices, for instance, rise more than 50% just in the past few months. That's going to translate into higher battery costs, all things being equal.

It's fascinating that it all comes back to the same problem: labor. Can we build these things? Can we build them quickly? Permits. A lot of our attention as a country maybe should just be there. We could take for granted these falling cost curves, which are great, but they're not sufficient to have the end system that we want.

If I was forcing you to invest in something related to the proliferation of data centers today, I'm curious how you would attack that problem. There's so much excitement around these things, and lots of companies are predicated on the rise and buildout of data centers. What would it take for you to get excited about investing in something tied to that trend?

There are a number of companies that are trying to make the buildout of data centers less labor-intensive through robotics. That's an interesting source of opportunity. It could be a very crowded space, though, so it remains to be seen how good an investment robotics companies will be.

From an energy standpoint, I think advanced geothermal is one of the most interesting components of the system today. It's a baseload energy source that is friendly to the environment. It's coming down the cost curve, but it's still very early in the industry, so it's unclear what costs are going to look like in a few years.

There is a skilled labor force that already exists in this country, and it comes largely from the oil and gas sector. The technology has now been proven out. There's a lot of work to do to scale this. It's kind of like the early days of the shale gas or shale oil revolutions. It starts out small because these companies have had to prove out the geology, prove out the techniques, and prove out the management team.

You also have to get to the project-finance market in order to get a very low cost of capital to make these things really work. It takes time to get from a startup in the industry—even getting all these steps proven—to a point where banks are willing to loan you money because they know it's going to work and they know they'll get repaid.

I can see that, in 5 years, the geothermal industry could be the most exciting in the United States.

We've talked a lot about energy infrastructure specifically. Are there other parts of U.S. infrastructure, outside of energy, that interest you most?

14. Rethinking the Role of Philanthropic Foundations

The housing discussions happening across the nation are really fascinating. A lot of YIMBYism, or “yes in my backyard,” emerged as a counterpoint to NIMBYism, largely in California. California had some of the highest housing costs, the fastest-growing housing costs, and some of the most stringent restrictions on trying to build new housing. It had gotten so bad that there started to be this pushback against all of that.

What we've seen, similar to permitting reform, is that this has become a very bipartisan issue: make it easier to build housing. We see it not only in California, but also in Montana, places like Austin, and in the Northeast. It has superseded politics.

I think politicians are now realizing that this affordability issue is front and center for voters. That's driving a renewed focus on the cost of electricity, the cost of groceries, and housing.

People are asking their mayors, governors, and the federal government, “What about housing? What are you doing to reduce the cost of housing and make it easier for people to buy that first house?” If you don't have a response as a politician, you're just not going to win a race these days.

President Trump came out and made this comment: “Well, I want housing prices for those who have a house to go up.” He also talks about making it more affordable to buy a house. You can't do both unless there is a massive government subsidy.

I think that's one of the risks of this affordability push. It's been decades in the making, from all these regulations and restrictions on building things in this country.

There aren't easy solutions to overcome this besides just big subsidies. The real solutions take time. The problem is that the window is longer than the political cycle. If you're a politician facing reelection every 2 or 4 years, you have to have an answer today about what you're doing that you can actually see during your term.

One of the easiest things to do there, if you have access to the budget—which federal politicians do—is to start subsidizing things. That makes the problem worse in the long term, but the electorate likes it in the short term.

I love to apply your sort of systems thinking to some other areas that you've studied carefully and been really involved with through the foundation. But before we do that, maybe have you describe your goal for the foundation overall.

One of the things I've heard you say that's so interesting is that a foundation should get less powerful over time—individual foundations, that is. And as someone who's running a big one, that's an interesting thought. I'm curious why you think that.

Before getting to some of the things that you focus on, just frame why you have a foundation if you wanted it to get weaker over time.

15. Improving the Criminal Justice System

Any institution gets less effective over time. I think it's true of companies. I think it's true of countries or governments. I think it's true of foundations. You have your best, most innovative times when you're still relatively young as an organization.

I'm very cognizant of some of the downsides of having perpetual foundations. One of the roles of a foundation is to take risks that the private sector and governments aren't incentivized to take. There might be political risks or economic risks, and foundations can provide that capital and resources because there's less accountability on foundations. They can provide the capital and resources to try things that others aren't willing to do.

But those are risky. Oftentimes, those will fail. I think organizations become more bureaucratic and more risk-averse over time, and so they're not able to do one of the main functions of a foundation.

There's a difference between charity and philanthropy. Charities are often designed to meet short-term needs, and philanthropy is more about trying to create long-term solutions. We do both of those. We're mostly on the long-term-solution side, but we also give money to our local food bank and help the local hospital.

It's harder work. It requires having a bigger staff, a longer time frame, and a willingness to live with ambiguity and the likelihood of failure on many of the things that we're doing.

The easiest money to raise is for a university that's going to build a new building, because a donor has trust in that institution. They know that if they give a dollar, they can point to the brick that their dollar went to. Sometimes, they even have their name on that brick, and so they see directly where their money is going.

A lot of the work that we do, which is based on trying to improve these big systems in America—whether that is the health care system, the criminal justice system, infrastructure, our public finance system, or education—is long-term work, and progress is slow. We really act as this conduit between researchers and policymakers, trying to do experimentation, evaluations, and research about what works and what doesn't in these systems. We help people think about new ideas, how they would test them, and how we can translate them into something useful for policymakers to improve the outcomes of these systems.

What about criminal justice? Describe your work in that part of the world. What progress have you made, and what progress do you think needs to be made in that system?

We first got involved in this by passively funding the Innocence Project. The Innocence Project tries to look at cases of people who have been convicted of a crime, do new DNA testing on some of the evidence, and see whether the evidence that was used as part of the conviction actually points to them as the perpetrator of the crime or not.

The Innocence Project has been able to overturn many convictions, and there are a lot of very sad individual stories there. But I think what's most important about the Innocence Project is that they're really looking at what process led to this wrong conviction and how we improve the system to lead to fewer incorrect prosecutions and convictions. They were using the individual to try to improve the system.

With that entry point, we started looking more and more into the criminal justice system. We heard a lot of anecdotes about some of the inefficiencies or failures of the system, so we took a step back and said, “Okay, where can we try to help in the system?”

16. Privacy and Security

A lot of the criminal justice principles came from the great crime spike that happened in the 1960s, 1970s, and 1980s in this country. It culminated with a bipartisan crime bill under Bill Clinton in the early 1990s that was really about getting tough on crime. It was about increasing the penalty if you're caught doing things.

Meanwhile, researchers have known for years that what's more important than the penalty, if and when you're caught, is the probability that you get caught. A lot of people who are committing crime don't have the ability to process, “If I get caught, will I have a 5-year sentence or a 10-year sentence, and is that going to change my behavior?” A lot of it is more, “Am I going to get caught or not?”

There were lots of questions about things like how you do pretrial detention. This is the time between when you're arrested for a crime and when you're convicted in court of a crime, and what should happen to you in that time frame.

We were asked to come into New Jersey to help them think about that. Kentucky, a Republican state, was one of the first to really try to redesign the system when its existing system was deemed unconstitutional. That meant thinking about what's important in that decision process.

What many states have decided is important is whether you're a threat to others and the likelihood that you're going to come back for your trial. That should decide whether you need to be detained before you're convicted of the crime or can be released back into the community.

Meanwhile, we have this cash bail system that's largely based on whether you have money to post to get out or not. We've worked with a number of states, both Republican and Democratic, over the years, trying to think about how to align the process with what the goals of society are.

This criminal justice reform movement has had its ups and downs. We were one of the very few philanthropies that was looking at it when we got started close to 15 years ago, particularly on the adult side. Then it became a very popular movement, and a lot of the people who were coming into it were coming in from a social justice and racial equity perspective.

I think one of the things that distinguishes us is that we always try to think about what all the goals of the system are. There are the goals of the criminal justice system, and first and foremost, you can never lose sight of public safety. Any reform you do can't have a trade-off on public safety. But social justice and racial equity are certainly goals in that system.

We look at a lot of ideas about how to improve it, and we try to figure out what trade-offs exist, quantify them to the extent that we can, and figure out whether there are trade-offs. If you do one thing, it could improve social justice or racial equity, but there could be a trade-off on costs or a trade-off on public safety.

17. Education and Life Outcomes

Do you think that the best place to improve things is just that probability—that subjective probability—that you'll get caught? Is that where we should invest, using technology especially, and invest the outsized amount of our time and attention?

Because it seems all it takes is 1 case: if some guy's probability of getting caught is low, so you let him go and he commits another really bad violent crime or something, everyone freaks out and says, “Shut this down.” It's like nuclear: Three Mile Island happens, and then we get no nuclear forever.

Does that mean that we should focus our attention on making people more scared that they'll get caught as the number 1 lever in the system?

I think so. The way you do that in a society where, first, there aren't enough public funds at the city level to really significantly increase the number of police, and second, communities are hesitant to have a lot more police in their neighborhoods, is difficult.

Yeah. Right. It's kind of weird in that communities don't want to lose the police they have, but you also ask them, and they don't want a lot more police coming in. This is the status quo bias associated with it.

And so the question is, if you don't have a lot more money and communities don't want a lot more officers just walking around, then what's the solution? I think there are a lot of interesting technologies that are being developed, and there's always this trade-off between security versus surveillance, or security versus privacy.

18. The Promise and Pitfalls of EdTech and AI

I think this is one where each community has to figure out where they want to be on that spectrum, and different communities are going to choose different equilibriums there. You walk down Midtown Manhattan, which might be the most surveilled place in the United States, right? It’s also one of the wealthiest. I’ve been in the real-time crime center in Beverly Hills, with cameras everywhere and drones, and everything is being filmed.

The wealthy communities have decided, “We want this, and we’re willing to trade some of our privacy for security.” Meanwhile, a lot of the debate happens around the assumption that low-income people don’t want that, and that they want more privacy at the expense of security. I think there are real questions about that.

You see the positive response people have whenever crime goes down, and the privacy that we’re willing, as individuals, to give up on the internet in order to get something we like—the rewards of the internet. We’re willing to give up a lot of our privacy. I think there’s this broad rethink now about this trade-off, what tools are available, and how to use the tools in ways that don’t lose the trust of the public and minimize the chances that people are going to misuse any of these tools.

What about education as something upstream of crime? Have you observed or discovered in your work a useful or interesting relationship between those 2 things? How do you think about the system of education as well?

The first issue we got involved in was K-12 education because the outcomes there are associated with almost any outcome you could think about that matters, whether it’s drug dependency, economic outcomes, or criminal outcomes.

It’s strong. Better K-12 equals better outcomes.

Yes, there’s a strong correlation, right? Then you try to ask, “Okay, is there causation here?” If I can improve educational outcomes, does that change those other outcomes? I think the answer is generally yes.

The hard part is figuring out how we improve educational outcomes. That’s the question the United States has been struggling with for decades. We’ve been able to do it a little bit as a society, but we haven’t cracked that nut.

19. Identifying Market Failures in Healthcare

What’s maybe surprising is that this is true globally as well. We’ve worked with a number of charter schools and worked with Teach For America over the years. Almost every country around the world that has a significant K-12 system has come to America to try to figure out what we’re doing because they’re facing the same challenges in education that we are.

What makes you most hopeful in that area? I had Joe Leman on the show talking about Alpha School last year. That seems young but extremely promising based on some of the early results.

I always have to catch myself so as not to be too overly optimistic about technology solving all of our problems, but it sure does seem like AI could solve a lot of the problems in education, maybe. I’m curious what you think.

I think there’s a lot of promise. I think there’s promise in Alpha School. My hesitancy about getting too excited is that we’ve been hearing this promise from the edtech industry for 20 years now. We’ve adopted more and more technology in the classroom over those 20 years, and outcomes have gone down.

You talk to any of these providers of an edtech platform, and they have all this data they’re going to show you about the remarkable results they get from their students. Then you step back, and you just never see it in the actual data when it’s applied in the real world.

Is that kind of like the solar thing, where the AI system is the equivalent of the solar panel? It’s getting better and better and cheaper and cheaper, but that’s not actually the thing that matters. It’s more like the rest of the stuff in the system.

I think the question is, how do you engage with kids? Oftentimes, kids have short attention spans. It’s probably more true today than it’s ever been. We can all probably think back about some remarkable teacher we had who captivated us and kept our attention for the full 60 minutes of a class. We can also think about teachers who were just terrible, and we wouldn’t pay attention at all.

Part of the challenge has been that sitting there staring at a screen often has not been an engaging format for people. The question is, can some combination of AI plus AR or VR create some type of more engaging content, as well as a more engaging delivery of the content, that we haven’t seen before?

I think there’s promise there. I want to be optimistic, but again, I’m so frustrated with the lack of results we’ve seen from this field for decades, with a lot of promise being there for a long time and just seeing nothing.

I feel greedy asking you about all these different systems, but there are 2 more that I really want to hear about: health care and journalism, which is one that I’m personally interested in, of course.

20. The Role of Regulation Across Different Systems

Maybe starting with health care, everyone knows the headline stats. The percentage of GDP that health care represents has ticked up steadily over time, while outcomes in many cases haven’t gotten better. There are some countervailing things, like GLP-1s, which seem to be a modern medical miracle of sorts that can address some of these problems and costs in the health care system.

When you assess that overall system, where does your attention get drawn? What are the things you’ve discovered are most interesting and important?

We’ve had this kind of multidecade financialization of the health care system. That’s really driven up costs as people have figured out ways to get in, start businesses, or consolidate businesses and take advantage of some of the regulations that exist in the field to maximize firm profits.

Health care violates almost every principle of a competitive market from economics textbooks. Think about even what information the patient has, what information the payer has, and what the incentives of the provider are. The provider has this asymmetric information advantage and also gets to decide what treatments you need.

You can load up all these market failures onto the system, and that requires this huge system of regulations to try to overcome them. If there were no market failures, then you could just say, “Okay, no need for government policy here or government regulation. We step back.”

But with all the market failures, you need books and books and books, and tens of thousands of pages of regulations, about how to deal with every single thing. You have a third-party payer, whether the government or a private insurance company, for the vast majority of the system. That creates all these rules, and private industry will naturally try to seek out and find where the gap in the rules is.

21. Journalism as the Fourth Estate

This happens in every industry. Right now, one of the big things is skin substitutes. If you get a burn, you need some type of new skin on it. There are regulations about how those get reimbursed, and for the first 6 months whenever a product is released, there’s no history of how that product has been priced before. So the federal government gives a lot more latitude on how a manufacturer of that product can price it for the first 6 months.

As the market starts to react, after 6 months, the government puts more regulation on the price. What manufacturers have done is introduce a new product, let it be out there for 6 months, take that off the market, introduce a new one that’s slightly different, and the prices keep going up. There are some kickbacks.

It’s just an example of this cat-and-mouse game that constantly happens when the government or private industry writes a regulation, and then everybody tries to figure out how to beat it. A lot of our work is trying to identify these problems, trying to correct them, and making the rules and incentives of the system more aligned with the outcomes we care about as a country.

As a general matter—not just in health care, but in all of these systems that you’ve studied—is the conclusion that fewer rules are better? Do you find yourself pro-deregulation as a general matter?

Not necessarily. This is one of the reasons why I don’t feel at home with either political party, because I think different problems need different solutions. We’re going to have a health care system that has third-party payers forever, and that creates the need for very tight regulation.

If you look at K-12, for instance, it’s a very different system, with very different actors, and there are fewer market failures. You can have maybe less regulation, and you can regulate different things. You can regulate outputs rather than inputs in the system.

22. The Kindness of Hard Truths

I’ve been very favorable to having the government try to get out of being the service provider in K-12. The government doesn’t do a great job of being both the regulator and the service provider, as a traditional district is. The government needs to choose one of those roles.

It’s not a question of whether regulation is good or bad. Different systems need different types of regulation. There’s always going to be this cat-and-mouse game between regulators and private actors, and we need to make sure that regulators both do a good job of devising the rules of the system up front and are actively fixing them as they fail.

Finally, what have you learned about journalism? It's one that stands out in contrast to some of these other massive countrywide systems that you've studied, explored, and done things in and around. What about journalism have you learned, and what does it need?

We got drawn to journalism because we view it as the fourth estate—a check on both government and private actors. There is a public good in both the dissemination of information and the investigative journalism that many outlets do: trying to find the frauds, abuses, and wastes that are happening. For a long time, you had this package product of the daily city paper, and that's largely fallen apart.

That role of both coverage of local and state politics as well as a lot of that investigative work, we feel, is being underinvested in. The commercial revenue potential of that is limited. Our view is that there is a need for philanthropy to fund some of this, the same way that philanthropy funds the opera house, museums, and parks. Having a vibrant journalistic outlet to oversee a lot of these things is necessary for a vibrant community.

You get to see so many parts of the system, probably more than almost anyone I've ever talked to. What has you the most excited about the world?

Innovation. It's very easy to get pessimistic. It's very easy to talk about the problems, to talk about the debt and deficit, and all our political dysfunction. Then you step back and look at what this country's gone through since its formation and the number of challenges it's faced, and see that it's been so robust in being able to overcome those challenges historically.

The innovation that many have provided and created over the years, and how that's improved quality of life for us—that's what we have to trust and have faith in, and I think it will continue to do so in the future.

When I do these interviews, I always ask everyone the same traditional closing question: What is the kindest thing that anyone's ever done for you?

When I was really deep into trading, it kind of created some unhealthy habits and started to change me. At one point, my brother pulled me aside and just said, “You've changed, and not for the better.” It took courage on his part to do that, and I think it caused me to step back and think about whether I agreed with that or not.

Of course, the first instinct is to say, “No, you're wrong,” and then that thought lingers, and you start trying to see yourself as a third person would. I started to think, “Maybe he's right. What do I need to change in my life?” I tried to make the other parts of my life, besides the business, high-performing as well.

I love the closing idea that saying something hard to someone you care about can be a great kindness. A cool and unique answer, John. Thank you so much for your time.

Great to be here. Thanks.