Fighting Words: The Energy Transition in 2026

(DESCRIPTION)

Title card: JP Morgan, Eye on the market. JP Morgan. Presentation. Title: 16th Annual Energy Paper, Fighting Words. Image: A man in a long coat and cowboy hat stands in a desert between two hills. He sprays liquid from a hose connected to a rusty oil drum, and he holds a propane tank that shoots fire from a nozzle. A man on the left wears a cowboy hat, vest, and jeans, and he wields a solar panel and a small wind turbine. A man on the right wears a cowboy hat and brown clothing, and he wields two glowing green rods. There is a radiation symbol marked on his back. All the men point at each other. A video box on the right. Michael Cembalest has short hair, glasses, and wears a dark zip-up jacket with light blue trim. He sits in front of a virtual modern shelf that holds vases, gold decorative items, books, and geometric objects.

(SPEECH)

[WHOOSH]

Welcome to the Eye on the Market podcast for March 2026. This one is our annual energy paper podcast, accompanied by the actual Eye on the Market itself, which has 200 charts in it. This one is called "Fighting Words" the energy transition in 2026.

And let me show you the cover art for a minute. So

(DESCRIPTION)

Slide: Fighting Words: debates, arguments and battles in the energy world, 2026.

(SPEECH)

the cover art is inspired by this movie called The Good, the Bad and the Ugly, which was Clint Eastwood and Lee Van Cleef and Eli Wallach. And at the end, there's a three way shootout and they're all pointing at each other in this shootout. That's what I find when I talk to a lot of people in the energy world. It's a three way shootout. You've got one guy with a wind turbine and a solar panel, who thinks renewables are the only way to go. And then you've got a guy in the back who's holding a natural gas canister and a diesel gasoline pump, who's committed to fossil fuels. And then you have the guy on the right who's holding two light green colored nuclear rods, that's the way they at least they appear on The Simpsons, and he's the nuclear guy. And so they're all shooting at each other because they think there's only one way to address both energy security and cost and sustainability.

And so-- and reminds me of a speech I gave a couple of years ago where I was talking about the energy transition and the role of natural gas. And there was a client that was so upset about the comments I had made that this person chased me out to the beach. I was in Miami. And I had my work shoes on, so I couldn't run, so that-- at least very far. And then when I stopped, I tried to explain and put some context around my comments and it didn't do any good. By simply chronicling the pace of the transition and talking about the role of natural gas, people ascribe all sorts of assumptions and motivations to you, even when there's no basis for doing so anyway. So those are the kinds of fighting words that I'm referring to.

And so this year in the energy paper, we're going to go through all the debates and arguments and battles that are taking place. So,

(DESCRIPTION)

Slide: Fighting Words: debates, arguments and battles in the energy world, 2026. A bullet point list.

(SPEECH)

for example, are data centers responsible for rising power prices? What's going on with the backlash right now in terms of requiring data centers to pick up a greater share of the tab of new generation. Can more of them be added to the grid if they agree to be curtailed at moments of peak demand?

As China continues to flood the world with cheap energy transition equipment, whether it's wind, solar, batteries, or EVs, how is that going to work over the long run if all their companies have deeply negative operating margins? What is the primary energy fallacy and why does it matter? One of the biggest debates in the energy world that has some real term implications for energy policy.

What's the real cost of solar plus storage when used for baseload power? There was a report that came out last year that claimed that it was in the same ballpark as natural gas and we do a full endoscopy on that report this year. Do renewables raise or lower power prices?

Here's a big one. Almost every client has asked me at some point in the last year are SMR, Small Modular Reactors, are they feasible? How much will they really cost? How much will they address data center needs before the year 2030? We're going to go into detail on that.

What should we make about all the hype over geothermal, geologic, hydrogen, Bloom solid oxide fuel cells, virtual power plants? We get into that this year. Why is it so hard to make money selling EVs. Very few companies can.

Why are battery arbitrage revenues collapsing in ERCOT? What are the real emissions footprints of natural gas and LNG supply chains? And how on Earth did XAI get permits to install natural gas turbines on tractor trailers and then become exempt from nitrous oxide emissions rules?

And then another topic why are sustainable aviation, motor, and shipping fuels still stuck in neutral after all the tax credits and people trying to do it? And then lastly, why are green hydrogen and carbon capture the poster children for unrealistic visions in our energy paper every single year? And obviously nothing has changed about that.

So these are the kind of debates, arguments, and battles that we get into in detail. We have a separate chapter on each one of these topics. I'm just going to go through a little bit of the content on this podcast and hopefully you're watching this instead of just listening to it because this is a very chart-based experience.

So where are we?

(DESCRIPTION)

Slide: The pace of the energy transition in 2026. Why is it so important to understand this? A line chart titled Decarbonization has been a mostly linear industrial transition since 2010 tracks renewable share of useful final energy from 1990 to 2025 across Europe, LatAm and Brz, China, Japan, US, Asia ex Japan and China, Africa, Russia and C.I.S., and Middle East. Most regions show gradual increases after 2010, with Europe and LatAm and Brz rising the most. China and the US trend upward steadily, while Japan and Asia ex Japan and China grow moderately. Africa and Middle East remain relatively flat with slight gains. Text: Source: Energy Institute, EIA, JPMAM, 2025

(SPEECH)

Well, we always lead with this chart, which is what's the pace of the energy transition? And, to me, that means looking at renewables as a share of all useful final energy. That latter phrase is very important. We discuss it in the paper.

And so far since 2010, this has been mostly a linear transition, very different from the S shaped technological adoption curves that you see in the technology sector. These are linear transitions. These are industrial transitions. And they take time and lots of capital.

And so now the pace of linearity is a little different here across countries, but these are linear transitions. And this has been really important to understand. Because if you read a few years ago all the things that were being written, you probably would have thought, A, there would be a much faster pace of these renewable share uptake and, B, there would be absolutely no economic benefit to investing in anything related to natural gas production or turbines or equipment and nothing can be further from the truth as we all know.

So

(DESCRIPTION)

Slide: The pace of the energy transition in 2026. Why is it so important to understand this? A line chart titled Projections of a US power supply-demand gap tracks gigawatts from 2024 to 2033, comparing Peak supply needed and Peak supply anticipated. Peak supply needed rises steadily each year, while Peak supply anticipated increases until around 2028 then declines gradually, creating a widening gap. Text: Source: NERC, Schneider Electric, 2025.

(SPEECH)

now the United States is all of a sudden looking at a fairly substantial power supply demand gap by the end of the decade. How big it is? Nobody knows. But it's pretty clear to me that there's going to be a power supply demand gap, a lot of which is going to be filled or attempted to be filled with combined cycle turbines that people use on the grid and by behind the meter generation with different kinds of natural gas combustion equipment.

(DESCRIPTION)

Slide: Turbines. A bar chart titled Global gas turbine orders measures gigawatts from 2001 to 2028. Blue bars show volatile historical orders that spike in the early 2000s, peak again around 2007 and 2011, then decline and fluctuate below the dashed line labeled Current production limit in the late 2010s before rising after 2022. Gold bars labeled Projected jump sharply above the production limit through the mid to late 2020s, then ease slightly by 2028. Text: Source: Bloomberg News, October 2, 2025.

(SPEECH)

Mitsubishi, Siemens, and GE Vernova are the other big dogs in the combined cycle turbine space. They've all announced some modest expansion to their production. Whether it's going to be enough to meet this projected demand is not clear yet.

(DESCRIPTION)

Slide: Why is it so important for investors to understand the dynamics of the transition? A table titled Gas turbine and solid oxide fuel cell manufacturer cumulative total return since January 2022 lists companies, products, returns, benchmarks, and benchmark returns. Bloom Energy leads solid oxide fuel cells with 697 percent versus Russell 2000 at 26 percent, while Mitsubishi Heavy Industries posts 1818 percent versus Nikkei 225 at 120 percent. Other firms including GE Vernova, Siemens Energy, Caterpillar, Doosan, Rolls Royce, and Wartsila show strong gains compared with their respective benchmarks. Text: Source: Bloomberg, JPMAM, February 25, 2026. GE Vernova and S&P 500 returns show since GE Vernova’s listing on March 27, 2024.

(SPEECH)

And look at what's happened to the stocks of companies that make gas turbines and solid oxide fuel cells, which can consume natural gas and create power. The returns on these stocks are astronomical and we compare them on the right to their respective home benchmarks.

And, again, the reason why it's so important to understand the dynamics of the transition is it allows you to understand everything else around it, and, in particular, the role of natural gas equipment. And so let me go to the next one here, having a challenging Zoom day. That's for sure.

And

(DESCRIPTION)

Slide: Data Center Load Projections. A stacked bar chart titled ERCOT large load interconnection requests measures gigawatts from 2022 to 2030 across categories including No studies submitted, Under review, Planning study approved, Approved but not operational, and Observed and energized. Total requests rise sharply after 2026, driven mainly by large increases in No studies submitted and Under review, while approved and energized portions remain relatively small. Text: Source: ERCOT, October 2025.

(SPEECH)

when we think about data centers-- and the first section in the piece and the longest one is on this whole question of data centers, what's their impact on power prices, and a bunch of other things. But the numbers are just staggering when you look at ERCOT as one example and you look at large load interconnection requests. That's fancy language for the magnitude of data center requests, which are large loads to join the ERCOT grid.

And a couple of years ago, this was 20 gigawatts. And by 2030, they're expecting it's 200. So this is growing by a factor of 10. Similar

(DESCRIPTION)

Slide: A line chart titled Large load adjustment forecasts for PJM measures megawatts from 2022 to 2026 and compares 2022, 2023, 2024, and 2025 forecasts. Each successive forecast projects higher adjustments, with the 2025 forecast rising most sharply and reaching the highest level by 2026, while earlier forecasts show more gradual increases. Text: Source: Monitoring Analytics, JPMAM, June 2025.

(SPEECH)

story in PJM, which is the mid-Atlantic grid, where data center ally. Again from 2000-- 2022 to 2025, you've had an explosion in large load forecasts, which are the large data center users that are trying to get access to the grid.

And

(DESCRIPTION)

Slide: Data Centers. A line chart titled PJM capacity prices are soaring tracks Base Residual Auction rate in US dollars per MW per day by delivery year from 2007, 2008, to 2027, 2028 for Mid Atlantic Area Council MAAC and Regional Transmission Organization RTO. Prices fluctuate over time with periodic spikes, then surge sharply in 2026, 2027, and 2027, 2028 to above 300, marking the highest levels on the chart. Text: Source: NRG Energy, July 23, 2025. (a) uncapped would have been $530/MW.

(SPEECH)

then you have this extra thing in PJM, which a lot of you may have read about, where capacity prices are soaring. And think of capacity prices as the way that particular grid pays insurance premiums to people that have generation capacity to commit to be available on the system on a given day. And these costs are being passed through to retail consumers for electricity. The White House had a meeting with the 13 governors of the PJM states. They're trying to encourage PJM to force data centers to engage in an auction where they fully pay and amortize down their new generation equipment. So you can really sense that there's a little bit of a data center backlash going on. Everybody's trying to figure out how this is all going to get paid for.

And

(DESCRIPTION)

Slide: Data Center Rate Surcharges. A horizontal bar chart titled Specialized data center power rates still trail cost of new generation compares US dollars per MWh across utilities and service types. Blue bars represent Base electricity cost and tan bars represent Specialized data center surcharge, with most combined rates clustering well below markers for Combined cycle gas turbine and Solar plus storage on the right. Utilities listed include Idaho Power, Duke Energy Progress, Kentucky Power, Arizona Public Service, Xcel, Entergy, and Dominion. Text: Source: Wood Mackenzie, JPMAM, 2025.

(SPEECH)

one of the reasons this is complicated is some utilities have begun charging data centers, special surcharges. They're still not enough. So there's a chart in here-- I think it's very important to understand that when you take the base electricity cost and you add the special data center surcharge, in many cases, it's still not enough to pay the cost of new generation, whether you're talking about a new combined cycle turbine or a more expensive solution, which is a solar storage combination, that's being used as base load power.

So-- and

(DESCRIPTION)

Slide: Data centers: B.Y.O.G. (generation). A table titled BEHIND THE METER ALTERNATIVES compares generation type, size per unit MW, all in capex $ per kW, lead time months, ramp rate minutes, land use acres, and efficiency. It lists aeroderivative gas turbines, industrial gas turbines, small combined cycle gas turbines, medium speed reciprocating engines, high speed reciprocating engines, fuel cells, and H Class combined cycle gas turbines, with efficiencies ranging from 35 to 60 percent. Text: Source: SemiAnalysis How AI Labs Are Solving the Power Crisis The Onsite Gas Deep Dive December 30, 2025 JPMAM.

(SPEECH)

because of that now you have data centers thinking about building their own behind the meter generation. Now, currently, to be clear, this is almost never happening. Less than 1% of data center power demand in 2025 was based on-site generation. But this is rapidly in the pipeline. So by 2030, we could have somewhere between 30% and 35% of all data center demand being sourced with behind the meter generation. Now, behind the meter you're not building a large 500-megawatt combined cycle turbine, you're stringing together smaller industrial gas turbines and fuel cells, reciprocating engines, and even some aero derivative gas turbines, which have been essentially modified plane engines.

So-- now, all these things take time, whether it's 12 months, 24 months, 36 months, sits-- five to seven years in the case of the big combined cycle turbines. But this is the big question is-- and this is the limiting factor on all of this data center stuff is access to power.

(DESCRIPTION)

Slide: Renewables as baseload power for data centers? A line chart titled 1 GW data center behind the meter buildout, $ per MWh compares Unsubsidized and Assuming 40% ITC costs across increasing shares of solar and battery capacity. Both lines rise as solar percentage and battery GWh increase, with the unsubsidized line climbing more steeply and remaining higher throughout. Text: Source: NREL, JPMAM, EIA, December 2025.

(SPEECH)

Now, sometimes people ask me, why don't renewables get used by data centers for behind the meter generation? They can. And, remember, if you look over the lifetime of a data center, only 15% or so of the total cost of the whole thing, capital plus operating and maintenance and fuel, is power. So if there's a power solution that's a lot more expensive, some of the hyperscalers might just go and pay it because either because it allows them to get hooked up sooner or virtue signaling or whatever their reasons happen to be.

But based on the analysis that we've done and other people have done separately, there's a common finding here, which is compared to a natural gas turbine a solar plus battery plus natural gas backup solution is a lot more expensive. And you can see in this chart, it's either 2 or 2 and 1/2 times the cost. But, again, if it's only 15% of the cost of a data center, maybe Google and some of the other companies pay it.

(DESCRIPTION)

Slide: Can the US keep pace with power demand? A line chart titled Generation and capacity additions since 1950 tracks US electricity generation in TWh and the percent share of new capacity additions from coal, nuclear, and gas. US electricity generation rises steadily from 1950 to the mid 2000s, then levels off. The share of new capacity additions fluctuates at high levels through the late 20th century, then declines sharply after the mid 2000s to low levels by 2025. Text: Source: EIA, JPMAM, 2025.

(SPEECH)

And so that-- the data center question is really all about whether or not the US can keep pace with power demand. Now, from 1950 to 2000, the US steadily added power generation capacity every single year, but 90% of all of the power that was added was very large coal, gas, and nuclear plants. The share of new capacity additions coming from those categories has dropped almost to zero because we're now in a world where most of what's being added is renewables and storage. And there's plenty of capacity being added. The question is, what does it really mean for consumers and for the grid?

And

(DESCRIPTION)

Slide: Can the US keep pace with power demand? A line chart titled US electricity generation and storage capacity additions: nameplate and ELCC weighted, Gigawatts, annual tracks additions from 1950 to 2025. Nameplate capacity fluctuates over decades with major spikes around the early 2000s and mid 2020s, reaching new highs. ELCC weighted capacity follows a similar pattern but remains consistently lower, with smaller peaks and a more gradual rise in recent years. Text: Source: EIA, PJM, MISO, ERCOT, CAISO, Thundersaid Energy, 2025.

(SPEECH)

one of my favorite charts in this piece is this one because, yes, the US added 65 gigawatts of new nameplate capacity in terms of storage and generation last year. But if we adjust that capacity for its intermittency and its reliability, the same way that NERC and FERC and the ISOs do by using these things called effective load carrying capability factors, all of a sudden that 65gw becomes 25 gigawatts real fast, which is a fancy, jargony way of saying every megawatt of generation capacity that's added to the grid is not the same. And adding a megawatt of coal or a megawatt of gas or megawatt of nuclear means something very different than adding a megawatt of solar. So that's the big question on data centers.

On-- we

(DESCRIPTION)

Slide: China. A bar chart titled China manufacturing capacity vs production, 2024 compares production and capacity across sectors. Solar PV modules show 587 GW production versus 1,045 GW capacity, wind power nacelles show 96 GW versus 204 GW, battery cells show 890 GWh versus 2,830 GWh, and electric cars show 12 million units versus 22 million units, with capacity far exceeding production in each category. Text: Source: IEA, 2025.

(SPEECH)

have a large-- the second big deep dive is on China. I'm not going to go through all the details here. The most important point about China is they are producing so much in terms of solar, wind, batteries, and EVs, and have so much spare capacity that the prospects of them continuing to flood the world with this stuff is still out there, despite the deeply negative profitability of a lot of the Chinese companies that are doing this.

Now, if your perspective is one of you want the fastest energy transition possible because your focus is on decarbonization and climate issues, this is all great news. Because we have a page in here that shows that the emerging world and Europe are aggressively hoovering up all of the solar panels, battery exports, and EVs that they can from China. US not so much for the reasons that we all know. If

(DESCRIPTION)

Slide: China, exporting the transition. Three line charts titled China solar PV exports by destination, China battery exports by destination, and China EV exports by destination track US dollars in billions from 2015 to 2025 across Emerging markets, Europe, US, and Other. Solar exports rise strongly to Emerging markets and Europe before dipping slightly, battery exports surge sharply to Europe and the US after 2021, and EV exports climb rapidly after 2022 led by Emerging markets and Europe. Text: Source: EMBER, September 2025.

(SPEECH)

you're a perspective [INAUDIBLE], you're an investor, or you run a company that makes this stuff in other countries, this is terrible news because the Chinese are continuing to compress margins and value across the board.

(DESCRIPTION)

Slide: China: Exporting the Transition. A line chart titled China nuclear will be a fraction of wind and solar even after buildout, Terawatt-hours compares China wind and solar with China nuclear from 2005 to 2030. Wind and solar generation rises sharply after 2015 and accelerates to around 1,800 TWh by the mid 2020s, while nuclear increases gradually to about 450 TWh and remains far lower even assuming 37 GW under construction are completed by 2030. Text: Source: Energy Institute, JPMAM, 2025.

(SPEECH)

Now, there's a lot of discussion about nuclear and China. And certainly China is now the world's leader in terms of nuclear fission, in terms of cost and speed. But even if they complete all 37 gigawatts they have under construction by 2030, their future generation, in terms of terawatt hours from nuclear, will still be way below even today's wind and solar and will be even further below the wind and solar generation that's taking place in China in 2030. So sometimes I think the focus on nuclear fission is a little bit too much in China. The paper also gets into Chinese research on cutting edge hydro, a lot of fusion research they're doing, sodium ion batteries, and things like that.

But, again, here's the weird dynamic.

(DESCRIPTION)

Slide: China and Solar Power. A horizontal bar chart titled Net profit margin of select listed solar companies shows percent trailing 12 months. First Solar, DMEGC, CSI Solar, and Canadian Solar post positive margins up to around 25 percent, while many others including Jinko, Longi, Trina, JA Solar, Risen Energy, TCL Zhonghuan, Jinyuntong, and Daqo report negative margins, with some falling below negative 40 percent. Text: Source: Bloomberg, JPMAM, Q3 2025

(SPEECH)

Of all of the solar companies in the world, the Chinese ones generally have deeply negative profit margins. And in any other country this wouldn't survive, but in China, for policy reasons, it can. And so we could be we could be facing another couple of years of lots of deflation and learning curve price declines taking place in a lot of renewable transition stuff.

Now,

(DESCRIPTION)

Slide: China and renewable supply chains. A bar chart titled China share of manufacturing capacity by product shows percent shares across Polysilicon, PV modules, Batteries, PV cells, Permanent magnets, and Wafers. All categories display very high shares near or above 75 percent, with Wafers and Permanent magnets the highest and PV modules slightly lower than the others. Text: Source: Global Critical Minerals Outlook 2025, IEA, JPMAM, May 2025.

(SPEECH)

as everybody knows, China also, and for that reason, dominates manufacturing capacity in solar cells and batteries and polysilicon and modules and magnets and wafers and everything else. And this gets to this question about why the US is doing what it's doing? The administration is very nervous about the supply chain dominance that China currently has, how hard it would be for Western nations to catch up.

We wrote a couple of years ago about how almost half of all Chinese critical mineral activities are unregulated. They're unregulated companies. So, I mean, a regulated company in China probably doesn't even adhere to that many rules. Can you imagine what an unregulated company in China would be doing? So it's very hard for the West to compete. And the United States, the current administration, is very reluctant to put the pedal to the metal on a renewable transition that deepens US reliance on China for all the reasons you can imagine.

And

(DESCRIPTION)

Slide: China and critical minerals. A bar chart titled Top producer share of refined strategic mineral output shows percent shares by country. China holds dominant shares near or above 70 to 100 percent in minerals including Gallium, Graphite, Rare earths, Manganese, Silicon, and Cobalt, while Indonesia appears prominently in Nickel. Shares generally decline across minerals from left to right, with Zirconium among the lowest. Text: Source: Global Critical Minerals Outlook 2025, IEA, CRM Alliance, May 2025.

(SPEECH)

this-- we have a chart in here on strategic minerals specifically that's kind of amazing. So this chart looks at 20-- the 25 important strategic minerals and it plots for each one what country is the top producer. For every single one, except one, the top producer is China. Now, sometimes the top producer has a 90% market share and sometimes the top producer can have a 40% market share, but China is the top producer in every example except nickel, for which the answer is Indonesia.

So-- but, again, it's important to understand why there's so much focus in the US on energy independence as opposed to climate. That's what the current administration is focused on.

(DESCRIPTION)

Slide: The new lens: energy independence. A line chart titled Energy dependence and independence tracks net imports of oil, natural gas, and coal in million tonnes of oil equivalent from 1980 to 2025 for Europe, China, US, and Russia. Europe remains a large net importer with fluctuations, China rises steadily to become a major net importer, the US shifts from net importer to net exporter after 2015, and Russia stays a consistent net exporter. Text: Source: Energy Institute, JPMAM, 2025

(SPEECH)

And after 45 years of trying, a couple of years ago, the US finally became a net exporter of fossil fuels. And that's a combined figure that looks at oil and gas and coal.

At the same time, Europe and China are continuing to slug it out. Each of them is importing roughly the same net amount of those kind of fossil fuels. So this administration is definitely focused a lot more on energy independence and has downgraded the importance of some of the other things that we were used to in prior administrations.

So

(DESCRIPTION)

Slide: Small modular reactors: investors are bullish. A line chart titled US small modular reactor equity indices total returns tracks index performance from January 2024 to early 2026 with Index 0 = Jan 2024. The iSelect US Nuclear Power SMR Index and the KEDII US Nuclear Power SMR Index both rise sharply through 2025, peak above 350 to 400, then pull back but remain well above starting levels. Text: Source: Bloomberg, JPMAM, February 11, 2026.

(SPEECH)

let's talk about SMRs, small modular reactors. The administration likes those. They're putting a lot of support behind them. And the investors like them. The-- last year there's two different SMR indices that did very well. The sweet spot for investing in a lot of these kinds of projects is before there's a proof statement because the vibe and the things people say is what drives investor sentiment.

When we get closer to finding out how much these things are actually going to cost and who's going to be willing to pay for them, I think is going to be a much choppier market. So I'm not sure that SMR investors are ever going to reap the gains they invested-- that they got last year when these things just took off.

So let's just talk for a minute. What

(DESCRIPTION)

Slide: What is an SMR anyway? A bullet point list.

(SPEECH)

is an SMR anyway? Well, it's a reactor less than 300 megawatts. It's allegedly produced with some kind of modularity and serial production. And there's four main types being pursued. You can build a small version of a light-water reactor using the same technology that's existed for decades and represents around 90% of the existing nuclear fleet or you can build some newfangled thing, a fast neutron reactor, a graphite-moderated high-temperature reactor, a molten salt reactor.

The important thing to about these is they're still in the process of having their first of a kind plants approved. The NRC, the Nuclear Regulatory Commission in the US, just approved the TerraPower one, which is sodium cooled fast reactor combined with molten storage-- molten salt energy storage. And then the high-temperature gas-cooled reactor, the NRC is in the process of reviewing the final license.

So that's what SMRs are. They're small. They can be based on old technology or they can be based on some kind of new technology that still has to be licensed and approved.

The

(DESCRIPTION)

Slide: Small modular reactors: reverse learning curve. A line chart titled The original learning curve led to larger nuclear plant sizes tracks average capacity of global nuclear plants in gross MW by construction start year from 1950 to 2025. Capacity rises sharply from small units in the 1950s to around 1,000 MW by the 1970s, fluctuates widely through the 1980s and 1990s, and stabilizes near a post 1980 average around 1,000 MW despite periodic spikes above 1,400 MW. Text: Source: Power Reactor System Database, JPMAM, September 2025.

(SPEECH)

issue that I have is energy investors and energy policy people are infatuated with learning curves. And it is amazing to see the learning curve of solar, wind, and battery prices with the benefit of increased production. You really have had some incredibly steep learning curves and now those learning curves are even extending to offshore wind.

If you really believe learning curves, I think you need to pay attention to the nuclear learning curve that took place in the 1950s, when people started out building 50 to 100 megawatt plants and then found that the fixed costs were so enormous that they moved from 50 to 100 megawatts t0 1,000 megawatts or gigawatt. And SMRs, to me, is trying to crawl back down the learning curve from what used to be learned. So I think you're swimming upstream here to try to modularize something that has such embedded high fixed costs. And so I'm a little bit of a skeptic.

And so--

(DESCRIPTION)

Slide: Small modular reactors. A bullet point list.

(SPEECH)

but let's define what skepticism means and let's define what success would be. So the levelized cost of a combined cycle gas plant under a whole bunch of different assumptions that we make, is somewhere between $55 and $85 a megawatt-hour, different utilization rates, different natural gas prices, whatever. So $55 to $85.

I would say that if a small modular reactor ever came in today's dollars at $125 to $130, that's a win because now you're talking about baseload power, lower carbon intensity, et cetera. The problem is I can't find anybody, except for the companies building them, that thinks they're going to be able to do that.

The former chair of the NRC is estimating first of a kind SMR costs of $200 to $400 megawatt-hour. The Tennessee Valley Authority did their own analysis last year, came out to about 200 mega-- dollars a megawatt hour. Studies prepared for some of our corporate clients, for them specifically, for behind the meter stuff-- SMRs, 2 and 1/2 to 5 times the cost of grid-based electricity. And even in China, which has modularized almost everything, conceded a couple of years ago that the SMR costs are going to be roughly double the large plants that it's already building. So I'm dubious about this. I think the proof statement is still out there.

And then I have another issue as well. When you think about how SMRs-- what's the premise of an SMR? Sometimes it's got new technology, sometimes it doesn't, but the premise of an SMR is that you can build it more cheaply. Think about your own life experiences of capital projects. How do you make a capital project cheap? Now, if you can mass produce them like mobile homes or can or-- you can bring your unit cost down. That's not what we're talking about here.

How do you actually reduce the unit cost of something when you're going to build 10 of them instead of three of them? I don't think that learning curve is an automatic. I think you have to bring down the costs of the core functions of the plant itself to make it cheaper. And that's what makes me nervous. And because-- there have been some changes in terms of the oversight of which government agencies are overseeing SMRs and the Trump administration has now moved more of that responsibility under the OMB and away from independent scientific agencies.

And

(DESCRIPTION)

Slide: Is the Trump Administration well positioned to oversee SMR development? A line chart titled Change in agency staff levels from previous year tracks percent changes from 2016 to 2025 for agencies including Centers for Disease Control and Prevention, National Institute of Standards and Technology, National Oceanic and Atmospheric Administration, National Institutes of Health, National Science Foundation, Department of Energy, Food and Drug Administration, NASA, and Environmental Protection Agency. Most agencies show modest fluctuations around zero through 2024 with occasional spikes, then all drop sharply into negative territory in 2025. Text: Source: Nature, January 22, 2026.

(SPEECH)

the broader question I have is if you think about issues related to public health and safety, energy, and science more broadly, this administration has been gutting those functions.

(DESCRIPTION)

Slide: Is the Trump Administration well positioned to oversee SMR development? A pie chart titled Disruptions and terminations of independent federal advisory committees at science agencies, January 2025 to June 2025 shows three segments. Active accounts for 70, Terminated accounts for 29, and Disrupted accounts for 9, with Active representing the largest share. Text: Agencies include: NSF, EPA, HHS and USDA. Source: GSA FAC Database, Union of Concerned Scientists, June 30, 2025

(SPEECH)

And when you look at specifically at the independent federal advisory committees at science agencies, they've eliminated-- they eliminated it a little more than a quarter of them just in the first six months of last year. So my question is an administration that's gutting agencies that are responsible for public health, safety, science, and energy, the right administration to be overseeing and responsible for, in an oversight manner, the kind of choices that are made for SMR developments, given the consequences if something goes wrong. So that's my question. And I don't that there's an answer to it, but that's my question.

(DESCRIPTION)

Slide: The hype on geothermal. A bar chart titled Shares of global useful final energy shows percent contributions from Hydropower, Nuclear, Wind, Solar, and Geothermal, divided into Electricity and Heat. Hydropower has the largest share near 5 percent, followed by Nuclear and Wind, while Solar is smaller and Geothermal is the lowest with a small heat component. Text: Source: Energy Institute, EIA, IEA, JPMAM, 2025.

(SPEECH)

Another-- in addition to SMRs, the administration is also very optimistic about geothermal. Geothermal is a weird thing. It's been around forever. It's been around for a long time. 80% of geothermal that exists is used for heat, only 20% of it's used for power generation. You have to go deeper for the kind of temperature gradient differentials to create power instead of heat. And most energy databases that you look at don't even list out geothermal separately. It gets lumped in with kite energy, wave energy, hamster wheels, and the other kinds of things like that.

And

(DESCRIPTION)

Slide: Geothermal in context. A line chart titled Global wind, solar and geothermal electricity generation tracks TWh from 1990 to 2025 for Wind, Solar, and Geothermal. Wind rises steadily and accelerates after 2010 to around 2,500 TWh, solar surges sharply after 2015 to above 2,000 TWh, while geothermal remains relatively flat with only slight growth. Text: Source: Energy Institute, Geothermal Energy Journal, JPMAM, 2025

(SPEECH)

if you have a chart in front of you, like the one I'm showing now, that looks at geothermal power compared to wind and solar, you can't even see the geothermal. It's so small.

Now, there's a little bit of a renaissance in geothermal.

(DESCRIPTION)

Slide: Geothermal costs appear to be coming down. A bar chart titled Levelized cost of energy for geothermal power shows US dollars per MWh for 2021 through 2024 across Hydrothermal flash, Hydrothermal binary, Near field EGS binary, and Deep EGS binary. Hydrothermal technologies remain below 120 with modest changes, while Near field and Deep EGS start near 450 to 500 in earlier years and decline significantly by 2024, though they remain higher than hydrothermal. Text: Source: National Laboratory of the Rockies, January 2026.

(SPEECH)

The National Laboratory of the Rockies-- you may not be familiar with it, it's the-- we used to call it NREL, the National Renewable Energy Laboratory-- these kinds of things were rebranded along with the Gulf of America or whatever the hell it is. According to this NREL, levelized cost for geothermal, particularly enhanced geothermal, are coming down. Enhanced geothermal is referring to when you're using fracking techniques to do geothermal the same way that you would for natural gas extraction. We'll see.

The

(DESCRIPTION)

Slide: Geothermal; a drilling revolution based on hydraulic fracturing. A line chart titled Utah FORGE well drilling rates measures depth in feet versus hours and compares several drilling periods with standard oil and gas rates. The Sep 2017 line declines gradually over many hours, later periods such as Jan 2021 and Jul 2023 reach deeper depths more quickly, and the shaded Standard oil and gas rates area shows faster drilling to about 13,000 feet in fewer hours. Text: Source: National Laboratory of the Rockies, January 2026.

(SPEECH)

good news is it looks like drilling costs are going down and drilling speeds are going up. Just from 2017 to 2023, the industry has shortened by a factor of 3 the time it takes to drill a well of, let's say, 10,000 feet and is approaching the standards used in the oil and gas business. So there are some benefits there. But geothermal is still a pretty complicated thing and ultimately power is a very competitive market.

And

(DESCRIPTION)

Slide: Power is the ultimate competitive market. A bullet point list.

(SPEECH)

geothermal will get its chance to compete, there's been a handful of PPAs, which are price agreements, signed in 2024 and '25. They were $20 to $25 a megawatt hour higher than wind and solar PPAs, so they're more expensive. But they're going to get their chance to compete. I'm not sure how scalable it is, but we'll see. And we have-- the piece that we-- the section that we have on geothermal walks through all the details of everything you'd want to about temperature gradients and standard binary flash geothermal and enhanced geothermal and all those kinds of things.

I am going to wrap up soon. Oh, I thought this was kind of amazing. Why

(DESCRIPTION)

Slide: Why is it so hard to make money in the EV business? A table titled EV operating margins lists brands, country, public or private status, EV sold in 2024, 2023, and 2022, EV operating margins, latest EV margin type, and trailing 2 year EV price return. Chinese brands such as Xiaomi and BYD show positive margins highlighted in green, while many US and European automakers including Ford, GM, Rivian, and Lucid report negative margins highlighted in red. Text: Source: Company disclosures, JPMAM estimates, 2026. A line chart titled Auto and EV equity indices total returns tracks index performance from 2020 to early 2026 with Index 100 = January 1, 2021. The S&P 500 Auto Industry Index and MSCI World Auto and Components Index fluctuate sharply with peaks in 2021 and 2025, while the Bloomberg EV Index drops significantly in 2022 and recovers only partially by 2026. Text: Source: Bloomberg, JPMAM, February 11, 2026.

(SPEECH)

is it so hard to make money in the EV business? There's a bunch of different companies that make EVs, some of them are pure play companies and some of them are diversified companies, like Toyota and GM and have EV segments.

So we track down the operating margin of the pure play companies and the individual EV segments inside diversified companies. And four-- just four EV companies, Chinese, and Tesla are profitable. And the rest of them all over the world don't make money, same goes for investors. If you look at the chart on the right, investors investing in an EV index have done much worse than just investing in the regular auto industry.

Will this change? I don't know. Old-- as the days go by, all I read about is more write-offs from Stellantis and GM and Ford and some of those companies are restating their core principles, which is we're moving back to customer preference, which is a euphemism for we're going to build what people want to buy rather than what the government mandates imply that we should build. Anyway, that's what's going on in the space.

(DESCRIPTION)

Slide: Essential Chart.

(SPEECH)

There's an essential chart section in the back, about 25 pages, and it's just charts and charts and charts and charts on all these different topics about EVs and renewables and the grid and energy storage and fossil fuels and LNG and energy independence, nuclear, carbon capture storage, hydrogen, sustainable fuels, everything. So I just want to show you a couple of things in there.

(DESCRIPTION)

Slide: Essential charts: US electricity grid. Four charts compare US transmission line growth and capacity buffers. Bar charts show transmission line miles added by ISO from 2013 to 2024, with growth peaking mid period and slowing recently. A line chart by voltage shows 345 kV additions highest and volatile, while 500 kV remains lower. Two line charts show summer and winter generation capacity buffers declining steadily across regions through 2037, with several regions approaching near zero margins. Text: Source: S&P Global, JPMAM, 2025 and NERC, JPMAM, January 2026.

(SPEECH)

One of the things we do is we track the grid. And the administration has also, in the US, talked a lot about wanting to break the logjam on infrastructure, and specifically that would be the electricity grid, liquids pipelines, and gas pipelines. And if they can get anything done on either one of those three, I'd be impressed because it's a very complex maze of local, state, regional, and federal rules that make it very difficult to build in the US.

As things stood-- at the middle of the end of last year, there hasn't been that much progress on transmission line growth, particularly the high voltage stuff, over 345 kilovolts. So we'll see if that picks up. What's notable here that we show on the bottom of the page, [INAUDIBLE], which oversees the grid, computes these things called anticipated reserve margins. And it's their way of saying how much safety buffer is built into the grids in different parts of the country based on the kind of capacity they're building and the kind of capacity that they're retiring.

And you can see that there's a little bit of a five alarm fire in both PJM, which is mid-Atlantic, and MISO, which is the Midwest grids, because of the balance between what's-- on in terms of the reliability of what's being added versus what's being retired. And so a lot of the regional ISOs are struggling, delaying certain decommissioning and things like that. But you haven't heard the last on this one.

(DESCRIPTION)

Slide: Essential charts: pipelines. Four charts show US liquid and natural gas infrastructure trends. Annual US liquid pipeline capacity added rises to a peak around 2018 then declines sharply after 2020, while US natural gas pipeline capacity added fluctuates with several large additions and smaller projected under construction volumes after 2025. Annual US miles of pipeline added peaks around 2014 then trends lower with a modest rebound by 2023. A line chart titled Waha Hub Pecos County Texas gas prices shows volatile prices in 2024 and 2025 with several sharp negative spikes. Text: Sources: EIA, JPMAM, 2024 and NGI Daily Gas Price Index, December 30, 2025

(SPEECH)

And the bright spot in terms of pipelines, if you view more infrastructure as a better thing, is it looks like there's some progress coming on new natural gas pipeline capacity. A lot of it is linked to the LNG exports to get gas out of some of the basins and towards the export area in the Southeast. But that's about it. What did I want to wrap up with?

(DESCRIPTION)

Slide: Essential charts: fossil fuels - US. A stacked bar chart titled Estimates of US natural gas supply measures trillion cubic feet from 1988 to 2024 and compares Proved reserves with Total potential resources. Total potential resources rise steadily from under 1,000 to about 4,500 trillion cubic feet, while proved reserves increase more gradually and remain a smaller portion of the total. Text: Source: Potential Gas Committee, 2025.

(SPEECH)

Let me just wrap up with a couple of things here, one, the S-- there's something called the Potential Gas Committee. And every few years publish what are proved reserves that are considered commercially viable and under existing technology but what are potential resources. And as you can see here, these things tend to rise over time. The US natural gas supply is booming and may continue to boom. It's

(DESCRIPTION)

Slide: Essential charts: LNG capacity utilization. A line chart titled Global LNG capacity utilization tracks LNG export capacity and LNG imports in billion cubic meters alongside utilization percent from 2005 to 2035. Export capacity and imports rise steadily through the 2020s, while utilization fluctuates near the mid 90 percent range before dropping sharply around 2030 and then stabilizing at a lower level. Text: Source: Oxford Institute for Energy Studies, October 2025

(SPEECH)

doing the same in other countries and so, therefore, for investors, have to watch out for this because for the first time really since the LNG market became an important thing, we're facing a period of declining utilization rates, which is a way of saying LNG imports around the world are going to go up but the capacity for people to export to them is going to go up even faster. So by the end of the decade, there's a glut predicted that will drive some of the utilization rates down to levels that we haven't seen before.

And

(DESCRIPTION)

Slide: Case study: renewable shipping fuels. A scatter chart titled 2026 levelized cost of shipping fuels shows US dollars per metric ton of very low sulphur fuel oil equivalent across benchmark fuels and alternatives under Low, Base, and High cases. Conventional fuels such as Fuel oil VLSFO and LNG cluster below 1,000, while blue ammonia, e ammonia, e methanol, and DAC CO2 combinations span much higher ranges from about 1,500 to above 5,000 depending on technology assumptions. Text: Source: Covalence LCOF Model v1.0, JPMAM, January 2026.

(SPEECH)

then last topic and then I'll let you go is for the last 10 years or so every year Greentech Media, Rocky Mountain Institute, certain investor conferences, very optimistic sell side reports, particularly from Goldman's Carbonomics series, are often very optimistic on carbon capture, green hydrogen, sustainable shipping fuels, sustainable aviation fuels, sustainable motor fuels. All I can tell you is that in the history of this effort, this Eye on the Market energy paper, we focus on the thermodynamics and the chemistry and the geology first and the market second. And that's helped us in so many ways avoid allocating capital to things that don't make sense.

And let me just show you this chart on shipping fuels. So there are some alternative to shipping fuels. One is just continuing to produce them the way that you do and then bolting carbon capture on it. But the other solutions are things called e-ammonia and e-methanol that involve getting a hold of some CO2 and combining it with gene hydrogen-- green hydrogen through electrolysis and-- et cetera, et cetera. The costs are astronomically multiples of magnitude higher than traditional fuel costs.

And it's so much higher that it's hard to envision the kind of carbon tax that would be necessary to equilibrate and make a purchaser indifferent between those two. And look at the chart in the piece that shows the original projections for sustainable aviation fuel and what they've actually turned out to be, one of the biggest gaps you'd ever see.

And then we always talk about another one of my favorite charts is the highest ratio in the history of science is the ratio of academic papers published on carbon capture divided by the actual amount of megatons of carbon capture taking place. And here's another-- one last example of how you have to pay attention to the details.

There

(DESCRIPTION)

Slide: Case study: battery electric shipping. A bar chart titled Effective volumetric energy density compares MJ per liter for a Diesel engine propulsion system and a Battery electric propulsion system. Diesel shows 3.68 MJ/L, while battery electric shows 0.28 MJ/L, with an arrow noting diesel provides about 13.14 times higher energy density. Text: Source: Int. J. Nav. Archit. Ocean Eng., Zhang et al, August 7, 2025.

(SPEECH)

was a splash recently over the last couple of months about China introducing an a battery electric container ship. Yeah, OK-- they did. It carries 740 TEUs, which are 20 those TEUs-- equivalent units. And that's a very, very, very small container ship. As a matter of fact, below 3,000 TEUs is the smallest category of container ship that even gets tracked. The most of them that are used are much, much bigger than that. And the issue is volumetric density. How much weight and how much volume do batteries take?

And the last chart I want to show you looks at the volumetric energy density. How much density can you fit in terms of megajoules of energy per liter of volumetric space? And for a diesel engine propulsion system, you're talking about something that's 13 to 14 times higher than the battery electric propulsion system, even when accounting for the fact that electric batteries are much more efficient, A, let's call it 80% compared to the 20% efficiency of an internal combustion engine. Even with that, you're looking at energy density that's 13 to 14 times different.

So, again, another example of where focusing on the details is important to understand the feasibility of things and where might there be investment opportunities and where they might not be. All

(DESCRIPTION)

Title: 16th Annual Energy Paper, Fights Words. Image: A man in a long coat and cowboy hat stands in a desert between two hills. He sprays liquid from a hose connected to a rusty oil drum, and he holds a propane tank that shoots fire from a nozzle. A man on the left wears a cowboy hat, vest, and jeans, and he wields a solar panel and a small wind turbine. A man on the right wears a cowboy hat and brown clothing, and he wields two glowing green rods. There is a radiation symbol is marked on his back. All the men point their weapons at each other.

(SPEECH)

right, thank you all very much for listening, If you're still here. And the energy paper is out. We have hard copies for some of you. Just reach out to your coverage teams and you can get one. And thanks for listening and I will be back to talk to you soon, maybe to have some fun and walk through all of the different rebuttals to the Citrini AI is going to be a black hole that destroys the world email that went viral last week on Substack. Thanks for listening. Bye.

[WHOOSH]

(DESCRIPTION)

Logo: J.P. Morgan.