AI, Policy and the Future of US Power Trading

[Music]

Lee Price: Hi there and welcome to J.P. Morgan's Making Sense. I'm Lee Price from the FICC Market Structure and Liquidity Strategy team. In this episode, we're going to focus on power, a key component in the ongoing evolution of U.S. energy markets. With significant energy demand growth driven by AI data centers, power capacity constraints and rising prices are emerging as key considerations for market participants. Against that backdrop, policy developments, geopolitical factors, and new market entrants are reshaping the energy trading landscape with opportunities and risks to consider. To shine a light on these market structure themes, I'm joined by my J.P. Morgan colleague, Chris Wu, head of U.S. power trading. Chris, thanks for being on today.

Chris Wu: My pleasure. Thanks for having me.

Lee Price: So it's an understatement to say that the energy market is in focus. The development of AI has sparked global competition and seemingly insatiable demand for power. AI data centers consume massive amounts of power and the significant growth expectations have altered the forecast for energy markets. To get a bit more specific, the U.S. Energy Information Administration is estimated that data centers could consume as much as 12% of U.S. electricity by 2028, which is up from 2% in 2020 and 4% in 2024. There's also a recent estimate calling for data center grid power demand to nearly triple by 2030. So AI is clearly the most significant element here, but it's not the only driver of electronification that we're seeing. You think about things like cloud computing, crypto markets, industrial electronification are also top of mind. For power trading, there are several developments at play here that could shake up the market structure, and particularly given the environment that we're coming from where load growth has not experienced meaningful change in almost two decades. So Chris, I'd like to hear some of your observations on the current market environment. Do you see much impact so far for trading the U.S. power grid?

Chris Wu: Thank you, Lee. The short answer to that question is a resounding yes. The U.S. power environment has seen considerable changes over the past 12 to 18-month period. First, the prices in the two most liquid trading hubs in the U.S., PJM and ERCOT have seen prices rise to all-time highs across the curve. The rise in prices began in Q2 of 2024 when the data center story first gripped the U.S. markets. Another major development has been the massive generation consolidation that kicked off with last year's Constellation/Calpine merger. The independent power producers of the world are putting their money where their mouth is as they have been touting the load growth story back to the beginning of 2024. The biggest players in the U.S. power markets are signaling that this load growth is coming and they are getting deals done which reflect that outcome.

Lee Price: Alright. So let's try to put that in perspective a little bit. Is there any precedent for what we're seeing in terms of the demand forecast for power?

Chris Wu: So I've been in the market since 2007, and in my career, I've never seen such a focus on load from a U.S. and a global perspective. The last time we had one particular focal point was when EV started becoming mainstream, but that story never gained the traction that data centers are having today. And I think the main reason is that there is a finite amount of EV vehicles that we can own, right? But I think if you think about the nature of data centers and hyperscalers, the growth there is infinite, because as an example, when you move from a ChatGPT-3 to a ChatGPT-4, the amount of compute needed doesn't double, it goes up exponentially. So I think this time around, the product that we're talking about has a very, very different demand profile than EVs did. I have recently started to see chatter and some articles that kind of parallel what's going on today with tech company valuations and sort of this power consumption story where people are trying to compare it to pre dot-com bubble, so '95 to '98, trying to sort of connect the large thermal build out that happened between 2000 and 2010 directly to the technology sector in that pre dot-com era. I think that is kind of ignoring the more macro, that occurred during that '95 to 2000 period, which is really the deregulation of the U.S. power markets. Before that time, there was no real IPPs in the market, right? Everything was regulated. The utilities owned everything from generation to transmission to load. I think when the markets deregulated and IPPs were, then allowed to market that power across, all different kinds of entities, I think that's really what spurred that build out. And also we had natural retail load growth. We still had a fairly robust manufacturing sector at that time before globalization really took a lot of that a lot of that away. So, I don't think that that parallel is warranted.

Lee Price: Okay, so a few really interesting comparisons there, but it feels like we're getting into pretty unprecedented territory here. And given the surge in power demand, capacity constraints and grid reliability might come into play. So we start with power generation. As of 2024, natural gas supplied over 40% electricity for U.S. data centers according to the International Energy Agency. Renewables, so think about wind and solar, supplied about 24% with nuclear around 20, and coal around 15%. So perhaps we can take a look at renewables to start. Over the past several years, national energy investment is focused on renewables with large scale solar, wind and battery projects under construction or in development. However, that trend may be set to ship. Chris, what are your expectations for renewable power generation going forward? And are there certain factors you're looking out for here?

Chris Wu: Absolutely. There's been quite a bit of changes in the renewable environment, but renewable power is here to stay. Despite the recent shift, away from federal subsidies, renewable generation will continue to be built in the U.S. for a couple of different reasons. One, you have a lot of commitments, in terms of reducing our carbon footprint. Those are led by large corporates, certain states, and I don't see that landscape changing, right? The other one is also, if you look at the pairing of renewables with conventional generation, it has served, to bring additional capacity online over the past 10 years. I think the big change, obviously, is cost. When you take subsidies away, future projects are going to cost 25 to 50% more, and that's just with the subsidies taken away alone. We're not factoring into the equation yet, the cost of supply chains, natural inflation. So I think the pace over the next two to three years, you're going to see it slow down considerably as the market reworks its economics, goes back to the drawing board a little bit. So we won't build at the speeds that we saw in the past decade. But as we exit 2028, I do think that you'll kind of start to see the development cycle start to grow again.

Lee Price: Got it. So cost and reliability are some of the considerations we'll be paying attention to. And as you mentioned, there are notable recent estimates that are calling for a decline in the near term, for renewable build before returning to modest growth a little bit farther down the road. So shifting gears to non-renewable, natural gas is the largest power generation source in the U.S. and also serves an important role as kind of the primary balancing source to meet seasonal variation in power demand. So with power demand surging globally, due to AI and other factors we've spoken about, many anticipate natural gas becoming increasingly important. Nuclear is another one that could play a big role over a longer time horizon. Chris, what are your observations here?

Chris Wu: I think gas and nuclear generation are a must, if the U.S. wants to execute on its infrastructure growth. Natural gas is the most obvious choice given the U.S. advantage in domestic gas production. It also is much more cost-effective and timely than nuclear is at the moment. Small nuclear reactors are still five years away from an initial rollout phase. And then the large-scale greenfield nuclear has seen an extremely long lead time, and is very costly. Our latest example is the Vogtle plant in Georgia, which ended up taking 15 to 16 years from the initial licensing, to going operational in 2023 and 2024, with a price tag of roughly $35 billion, which was two and a half times the original proposed cost. The problem with gas generation at the moment is just a congested pipeline on large turbines. Currently, there's a backlog of turbines that stretch all the way out to Cal 32, and the cost of building that gas generation has gone up from 750 to $1,000 per kilowatt to 2,000 to 2,500, just given the boom in demand. The difficulties in the development environment and U.S. power is undergoing significant policy change, and that takes time, right? I would expect the markets to be volatile and tight over the course of the next two to three years until we can see the light at the end of the tunnel. The third generation source that I would love to mention here are batteries. Batteries have provided stability across the day by adding demand during periods of overgen, especially in the solar hours, and shifting that generation to the high demand periods of the day. It has proven to work quite well in California markets, where solar has pushed the midday demand below zero and evening peaks had become somewhat unmanageable. The batteries have stabilized that market very well across all seasons. The caveat for batteries under the current monetization methods, they're only financially viable in markets with either high ancillary services values or wide intraday energy spreads caused by severe solar penetration. So currently, they're mainly built in California and Texas. In other major markets such as PJM, the build out pace is far lower. But the technology has proven to work, and it provides a great way to move solar generation across the rest of the day. So I think storage is going to become an increasingly crucial role, to keeping up with power demand growth. So I've briefly touched on a few factors contributing to the mix of energy production going forward. I've obviously mentioned policy change, but I know you and the FICC Market Structure team have been tracking some of those developments.

Lee Price: Yeah, I think what's happening in energy in a lot of ways is something we've been tracking internationally where really since the start of the year, we've seen domestic policies that prioritize growth, innovation, and national competitiveness. So it's a theme that's taking place across major economies. There have been several policy actions that outline the energy priorities of this U.S. administration, including if you think back to Inauguration Day, there was an executive order titled Unleashing American Energy. So, you mentioned the shift away from renewables and EVs in terms of subsidies, but much of what that order was trying to do was instructing federal agencies to establish the U.S. as a leading energy producer and to encourage new energy exploration. And later in February this year, a separate order established, the National Energy Dominance Council, which is really an executive branch entity that coordinates with the federal government agencies. So thinking about the Department of Energy, the Department of Agriculture, to carry out the president's energy objectives. So, you could broadly break those down into a dual mandate of ensuring affordable and reliable energy for Americans, and then also generating enough electricity to power these AI ambitions we've been speaking to. So, there'll be certainly more to come here, and it'll be interesting to see how these developments encourage energy exploration and investment going forward. But the political focus highlights that when it comes to supply, power generation is just one part of the picture. So you have transmission and also distribution, which is achieved by the power grid, and in some ways, that's an even more challenging part of the picture given the aging infrastructure. It seems that the AI story has brought increased attention to the limitations of our current U.S. power grid, and the time-consuming process that is required to make improvements or bring new projects online. Chris, when you think about how transmission investment, can speed up or slow down this AI growth story, what comes to mind for you here?

Chris Wu: Transmission really is the forgotten piece of the puzzle a lot of times, right? And I think a part of that is due to the fact that from the moment that the energy markets deregulated, at the end of the '90s, transmission was actually not a piece of that deregulation. So, transmission sat with the vertically-integrated utilities, and most of that policy has not changed since that time. I think what that has led to is a very difficult environment for private companies to build and compete in those transmission markets. But transmission development can unlock multiple pockets of generation, which are geographically locked due to its weather dependency. Think wind and solar in West Texas, solar in the southwest, wind in the Great Plains, wind in the Northwest. There's a lot of stranded generation, which cannot then lead to supplying parts that have stranded load. Cost and many other physical factors are an issue with transmission development, but it has long taken a backseat to the generation development story. And I think that we need more strategic planning around this topic and perhaps also regulatory and policy changes, because transmission is a key to the efficiency of the total system, and it could actually offset the need to build, more generation in certain areas when really transport is what's needed.

Lee Price: Right. So we've got long project lead times as well as pipeline congestion, but we are starting to see more tangible policy proposals to address some of those issues. So as an example, last month, the Department of Energy proposed a rule to expand the authority of the Federal Energy Regulatory Commission, which would give it jurisdiction over the interconnections of loads larger than 20 megawatts. So the goal here being to accelerate the ability to add connections to the grid. And in the meantime, we've seen efforts from certain entities to produce onsite power for AI data centers that would bypass the power grid or use a mix of grid and onsite power, so kind of a hybrid setup there. What do you think of this trend, and what would arise in offsite generation mean for trading the power grid?

Chris Wu: So, data centers are obviously well-aware of their development backlogs and grid issues, which is why they are already starting to look at behind-the-meter generation. A lot of that would be powered by small, less efficient generators. Think about the caterpillars of the world that produce these smaller 30, 35, 40 megawatt generators. I kind of make the example of a generator for your home, basically, that turns on in emergency situations, but imagine 200 of them, 300 of them, right, piled, on a single site. And that's really to get their product to market in the fastest time possible. I think the ultimate plan for the tech companies and the data center companies is not to be generation owners, but they need to bridge the gap between how fast it takes them to get to market and ultimately one day to reconnect back to the grid. I think given the potential revenue for data centers, really all they care about right now is that speed, right? I think they're willing to say, look, we understand five years from now, maybe seven years from now, we will ultimately integrate back into the overall power grid, but we know that we can't do that today, and so we need to figure out what that means for us and what the fastest strategy is. For example, a lot of the Texas data center developments are already accumulating these small generators. Now, I think in terms of what that means from a trading perspective is actually not that much, right? Because for the next five years, if you get a majority of development within certain markets that are behind the meter, that is not going to change power prices, for those particular markets. Now, however, these are all gas fire generation, so that could change the gas trading environment, in those markets because you do have to go to market to procure that gas supply. But I think where it leads to a change in power trading is eventually that five years down the road, that seven years down the road, are we going to be able to get a significant amount of supply online for that moment when those data centers do actually come to the power grid?

Lee Price: Right. No, that's really interesting. I think this kind of bring your own energy trend is going to be an area to watch. And Chris, just given the topic we're discussing today, it would be remiss of me not to mention the recent launch of J.P. Morgan's Security and Resiliency Initiative to boost critical industries, which aims to support the U.S. in modernizing infrastructure, fortifying supply chains, and implementing policies that promote growth. And there's four key areas of the initiative, and one of those will focus on energy independence and resilience, including battery storage, grid resilience, and distributed energy. So, continued private and public sector investment in this space will be something to look out for with both power generation and transmission. We're seeing pressure on existing infrastructure and the desire to build, thinking about data centers, storage, transmission, liquefaction, terminals for gas, pipeline and shipping. Thinking about some of the current challenges associated with U.S. energy and with volatility being such a feature of this market, have you seen more sophisticated or financial participants entering the market recently?

Chris Wu: Absolutely. Over the past three years, we have seen a growing number of macro funds, CTAs, and quantitative funds, enter the financial markets. Power is fast becoming a macro theme as the financial world realizes its prominent place in energy commodities. I see this as a long-term macro shift, and just like oil or gas, it will be a material part of any commodity business for the foreseeable future.

Lee Price: Really interesting. Thanks, Chris. And kind of bringing us full circle, the supply demand imbalances being a key factor that contributes to the price volatility in power markets that may be bringing some of these new participants in. Being able to navigate these forces as well as new policy frameworks and competitive dynamics will certainly be keeping yourself and market participants busy. So we're almost up on time, but we've covered a lot of ground today with AI-driven energy demand, power generation, transmission challenges, and the evolving trading landscape. There's really a lot to look out for in this market. And as we head into the winter months, we may see these trends continue to heat up. And Chris, I may have some follow up questions for you next time I get my electricity bill. (laughs) I can't wait to see what's next. And Chris, thanks so much for joining today.

Chris Wu: Thank you, Lee.

[Music]

Voiceover: Thanks for listening to ‘Market Matters.’ If you’ve enjoyed this conversation, we hope you’ll review, rate, and subscribe to J.P. Morgan’s Making Sense to stay on top of the latest industry news and trends, available on Apple Podcasts, Spotify, and YouTube.

The views expressed in this podcast may not necessarily reflect the views of JPMorgan Chase & Co, and its affiliates, together J.P. Morgan, and do not constitute research or recommendation advice or an offer or a solicitation to buy or sell any security or financial instrument. They are not issued by Research but are a solicitation under CFTC Rule 1.71. Referenced products and services in this podcast may not be suitable for you, and may not be available in all jurisdictions. J.P. Morgan may make markets and trade as principal in securities and other asset classes and financial products that may have been discussed. The FICC market structure publications, or to one, newsletters, mentioned in this podcast are available for J.P. Morgan clients. Please contact your J.P. Morgan sales representative should you wish to receive these. For additional disclaimers and regulatory disclosures, please visit www.jpmorgan.com/disclosures

© 2025 JPMorgan Chase & Company. All rights reserved.

[End of episode]