There are two types of very powerful growth: when small things grow fast and when big things grow slowly. New technologies—everything from the adoption of new devices and appliances to the growth of wind and solar—are powerful when they grow fast from small beginnings. But when large systems, from energy consumption to the national debt to population, grow at a slow rate, the compounding effect can be quite powerful. If you want to catch a big thing growing slowly, therefore, you may not only need the help of small things growing quickly, but a separate strategy to get the big thing’s growth rate decisively back below 1.00%. Should global consumption of fossil fuels maintain their 1.25% growth rate over the next five years, total global consumption will reach 551 EJ in 2030. That would make for an aggregate addition of 61 EJ over the 2021 level, or a total advance of 12.4%. (For context, because EJ are hard to visualize: The entire EU ’s energy consumption last year stood at 72 EJ.) Compounding: It adds up.

Coal remains the world’s second-largest energy source, despite rapid gains by natural gas to catch up, from third place. When we look at the three individual fuels on their own, we can see pretty easily the slowdown in global coal growth as wind and solar, natural gas, and more aggressive policies have joined together in opposition. To the extent that natural gas emits at least 50% less CO₂ than coal, on a combustion basis, it’s a victory that a great deal of potential coal growth has been curbed. The alternative view is that coal has been able to continually squeak out successive new highs of consumption, despite being attacked from all sides. The persistence of coal—19th-century relic and undisputed leader of the industrial revolution—even in light of the additive model of energy transition, is frustrating if not mind-boggling.

When we break global coal growth down to its two constituent parts, there’s a clear story to be told. One of the common hopes for wind, solar, and storage is that they would be particularly competitive in less developed countries, the non-OECD, where they would fight fewer incumbents. It was theorized, just as with mobile phone adoption, that many of these young countries would essentially “leapfrog” to the latest technology, in contrast to the OECD, where incumbency had long been established. As it turns out, new demand for coal has been far more robust than anticipated. If we ask, “Which has been more powerful in wind and solar’s growth, its affordability or supportive policies?” one would have to lean toward policies. This is another way to think again about the nuanced difference between cheap and crazy cheap: Wind and solar and storage are the former, not the latter. If they were crazy cheap, they’d be unstoppable in the non-OECD.

Wind and solar and storage are of course growing quite nicely in the non-OECD, especially in China. But so far, they are playing the additive role, as coal consumption and natural gas consumption press onward. Good news, however: Data shows that China, at least for one year, brought coal growth down to zero in the power sector. And because of China’s massive weight in the index, that alone dragged down power-sector fossil fuel growth in the non-OECD also to zero last year. That will be a very hard achievement to sustain, but let’s hope it can be achieved. As you can see in the chart below (interactive on a web page—can be moused over for data—but not in mobile), wind and solar took a full 100% of total non-OECD power-sector growth last year. When you consider the relentless growth of power demand in general across the non-OECD, that’s encouraging. However, as seen in other large domains, wind and solar taking 100% of marginal growth leads to an increasingly familiar and frustrating juncture: great news to have arrived at that precise spot, but a wall of incumbency tends to block further progress. The layers of the additive model strike again.

China remains the world’s best hope to shut down economically viable fossil fuel power generation, using brute force. As a command economy China, has the freedom to pick winners and losers, to force losses on any sector they deem. China is not constrained, as we are in the West, by the sanctity of private property. One could argue they have already started down this road, given their massive buildout of wind and solar, and that through this massive buildout they are creating ever larger spare capacity in their coal fleet, therefore degrading its economics. The only hitch there is that in the past few years, when summer heat or winter cold have moved to extremes, China has been forced to turn to its mighty coal fleet to meet such spikes in demand. For the moment, therefore, we might conclude that China has taken definitive steps to undermine the future viability of its coal fleet but has stopped short of making the more decisive move.

At the current juncture, unfortunately, total coal consumption in China continues to advance, even as the country has done a good job trying to fund growth with an array of other sources. So, to its credit, coal has fallen as a share of total energy use, but alas, coal marches onward in China, reaching new highs in both 2024 and 2025.

It’s instructive to add a throwback chart here, showing the moment in the early 1960s when global oil consumption, having risen steadily since the 1930s, finally overtook coal as the world’s top energy resource. (Note: The unit of account here is the Mtoe, as this is from an older BP Statistical Review dataset. And yes, just like Btu or EJ, an Mtoe—a million metric tons of oil equivalent—is a unit of heat.) Understandably, at the time, oil’s versatility convinced people that it would snuff out coal, especially since oil was a liquid and thus far more adaptable than old, solid coal. Sixty years on, however, we’ve learned that coal can still perform many industrial tasks that can’t be undertaken as easily, either with natural gas or versatile oil.

The front end of any energy transition is convincing and powerful, as the new energy source chews its way quickly through a universe of new opportunities. The early years for coal as it overthrew wood, and oil as it overthrew coal, were positively dizzying. The past 15 years for wind and solar have been similar. From a small base, wind and solar are competing beautifully for marginal growth in power and, because they are sources of electricity, are far more adaptive to a digital world that runs on software and, yes, AI. Wind, solar, batteries, and technology fit together in ways that are modern. Coal, by comparison, is almost comically archaic. Coal is dumb, crude, inefficient, and yet packed with powerful energy that creates high heat. The heat from combustion remains crucial to global industrialism—crucial especially to creating all the steel in the ground necessary to mount new energy infrastructure, like wind and solar, for example. In renewables, the world has happened upon clean energy, and, for the most part, clean electricity. What the world lacks is a way to create heat at scale, without the associated emissions. It’s a tough and so far insurmountable reality.

—Gregor Macdonald

The just released Statistical Review of World Energy from the Energy Institute confirms that the additive model of energy transition maintains its grip. Last year, wind and solar dominated marginal growth of world energy demand, making the largest contribution to new energy supply. Yet, fossil fuels continued to grow, leaving their existing underlayer untouched. The result once again is that renewables continue to play an additive role, helping to serve growing demand but never eating into, or disrupting, the still much larger role of fossil fuels in the global economy.

The additive model of energy transition is rock solid as a descriptive heuristic, but one has to be careful about deploying it as a predictive tool. Over the long run, the existing fossil fuel base will eventually erode, and renewables (wind, solar, and storage mainly) will surely cross over from their additive role to a far more disruptive role, in which they meet not just marginal growth, but degrade the economics of incumbent energy providers, forcing more rapid retirements, and substitutions. It has been about two years now since Cold Eye Earth abandoned the outlook that this cross-over was imminent, and adopted the additive model as the best guide. And there is still no reason to think that the additive model will spoil, anytime soon. (more on this later).

Many energy observers are hostile to the additive model because, as a guide, it is far better and more accurate than their serial, year-after-year predictions of the demise of fossil fuels. In the view of Cold Eye Earth, it may not have occurred to these observers that the additive model is highly aware, and uses history to demonstrate that nearly all energy sources over the past several hundred years have reliably continued to grow, even when specific energy sources fall out of favor. Did oil overtake coal in the 1960’s to become the world’s primary energy source? Indeed it did. And did coal consumption growth slow down alot? Yes, that too. But did coal ever stop growing? No, unfortunately, it did not. And coal demand, both in the power sector and more broadly, is still growing. Global coal demand hit a new all-time high last year.

Our first view of last year’s data comes through the comprehensive chart showing the marginal contribution each energy source made to 2025’s growth in global energy demand. Here, you can easily see that renewables is taking a leadership role, on a relative basis. Cold Eye Earth is not a fan of the “renewables” category because it adds the tiny contributions of geothermal and biofuels to the far larger output of wind and solar. The alternative view of this category is that, when tallying up global energy contributions (and not just the power sector), it’s useful to bring these junior partners into the category. Here is last year’s scoreboard:

Although we must be cautious about using the additive model as a predictive guide, you can see pretty easily how those making that argument currently have the data on their side. The argument goes like this:

Just as the historian Jean-Baptiste Fressoz laid out in his book, More and More and More: An All Consuming History of Energy, every time a new and successful energy source is introduced to the world, (coal, then gas, then oil, etc), that energy source starts to dominate marginal annual growth but all the other, existing sources continue to grow. In other words, energy history is a story of layers, one stacking on top of the others, forever. We should expect the low-carbon revolution to follow the same path. Clearly renewables are dominant now in marginal growth, and while further growth of oil, natural gas, and coal will certainly slow down (and already has), actual, outright declines in those fuels will remain elusive.

As Cold Eye Earth has observed previously, the common blind-spot among climate hawks and renewables advocates is the factor that lies at the heart of the additive model’s competitiveness: economic and population growth. The global climate community is itself dominated by a view that fails to pay enough respect to economic growth, and is always and ever over-persuaded by the fact that wind and solar growth is in insane mode. In doing so, they leave out the most controlling variable to energy transition. It is absolutely wonderful that wind and solar and storage growth rates are on fire, and their insane mode growth is sure to continue. They are at least tamping down the growth of fossil fuels, and for this we are thankful. But none of that success amounts to a rebuttal of the additive model. They are instead confirming its thesis.

We should be less concerned about the ability of renewables to halt the growth of fossil fuels on a global scale, and far more concerned about their ability to compete in their own, specific lane: the power sector. No one expects renewables to pull down the growth of all carbon sources to zero just yet. We are still too early for that, with myriad sectors from steelmaking to cement and other heat-based industrial processes still bereft of any scalable alternative. So let’s not make the hurdle too high for renewables with taunts and criticisms that they “still only make up 5.9% of total global energy demand.” Listen: nuclear power only makes up 5.2% and hydropower just 2.7% of global demand, and no one is suggesting these are scams or failures, right?

Indeed, this is a good occasion to point out that the vast majority of the criticism against renewables is just nonsense and idiocy. These lines of attack come from those who don’t understand growth, and who cannot comprehend large systems and how they change over time. This impulse far outweighs the tendency of climate folks and renewables advocates to play loose with the data to tell happy stories that have yet to come true.

With that brief assessment out of the way, everyone needs to face up to this:

Despite a very successful, decade-long campaign in both Europe and the U.S. to drive out coal from their respective power sectors, global coal in power continues to grow, having hit a new all time high in 2024 at 10,750 TWh and easing last year by a scant 0.56%. Indeed, coal has absolutely crashed in the U.S. from providing half of all power supply to less than 15% of power, as natural gas alongside wind and solar have soared. Combined, the EU and the U.S. have shed nearly 1500 TWh of coal-fired power since 2010. Wow! But in that same period, global coal-fired power has advanced over 1800 TWh, not just entirely erasing the combined EU-U.S. retirement wave, but actually adding to the consumption total. In other words, in the period from 2010, which Cold Eye Earth uses as the starting line for the big global push into wind and solar, coal was able to not only survive an enormous loss of demand from the OECD, but grew so quickly in the Non-OECD that it outcompeted wind and solar consistently for a large portion of marginal power sector demand.

Peak coal theory, therefore, can now be considered a total failure. Indeed, even in the sector where wind and solar are most competitive, coal continues to press onward. It doesn’t matter one bit that coal crashed in the EU and the U.S. Those crashes say nothing about coal’s capability to keep growing in global power. You can nitpick and data hunt all day long, convincing yourself that “with all the coal retirements and now China easing back on coal that surely peak coal is finally here” but unfortunately that theory has been hanging around for more than a decade. Indeed, we all thought coal had peaked back in 2014—both globally and in the power sector. We were all wrong.

Global wind and solar continued their relentless march forward last year, nearly providing 100% of marginal growth in global power. That is both extremely impressive, while also being a perfect portrait of the additive model of energy transition. Coal inputs to global power fell by a half a percent; while natural gas inputs to global power rose by roughly the same amount, 0.6%. There is no trend, no data point, or any other reason to believe that coal and natural gas won’t either continue to grow in power, or at the very least, hold on to their current position for a while yet. This is exactly the outcome that the additive model describes, and at some point, we may want to admit the additive model does have some predictive power.

As you can see, combined wind and solar absolutely smothered all marginal growth in global power last year. Even more impressive is the leap from covering just half of marginal growth in the year prior, 2024. Well, it’s impressive until you observe that total demand growth actually fell last year compared to 2024, only reaching 855 TWh. That made the hurdle far easier for wind and solar. Nevertheless, you would be surprised how many will defiantly conclude that, based on one year’s performance, combined wind and solar are on the verge of overrunning the system—not just covering marginal growth, but eating into the fossil fuel underlayer. Advice: don’t do that.

There are a number of reasons why the additive model may continue to work as a predictive tool specifically within global power, at least for another three to five years. Let’s go through them.

• Energy transition by its very nature is all about bringing economic activity over to the electricity system. This supertrend now has a new entrant: data center growth. Together, the years like last year when total demand growth was lower than the prior year will be rare. We have entered a new domain of above average growth in electricity demand.

• The economic viability of existing natural gas power and coal power continues to stand as an immoveable wall. Retirements are not numerous enough, and wind and solar may win the bidding for all new power capacity in some domains but we don’t care about “some” domains. Clearly, coal is still competitive enough to get built, and in systems like China—where vast, untapped coal capacity is always ready to be placed into service—the upward risk of coal-fired power during hot summers, cold winters, and stronger manufacturing remains.

• Natural gas remains dirt cheap. Not a single analytical forecast about the economic viability of shipping LNG ever came true. In fact, natural gas is not just globalized through LNG but the LNG market is robust. So even in a plausible world where coal starts to (finally) die by entering sustainable decline, wind and solar would still have to share the playing field with natural gas. And North American natural gas remains plentiful, and mind-blowingly cheap.

• No large country or domain has embarked on an aggressive path to actually shutter economically viable coal and natural gas power capacity. Sure, OECD countries have deployed small versions of this effort through incentives, tax breaks, and other policy preferences. Outside of emergencies though, it is very, very hard for democracies to force losses onto the private sector. Without such aggression, therefore, we are all standing around waiting for the existing global fleet of power to age out on its own timeline. And you can already see what that outcome looks like: it looks exactly like the additive model says it will look.

• There’s a huge misunderstanding about the affordability of wind and solar and storage. Yes, they are cheaper than building new coal or new natural gas in many domains. But they are not dirt cheap. Dirt cheap is the input to truly revolutionary change. Compared to their energy content, coal, natural gas, and oil were dirt cheap in their day and except for periodic crises, they remain dirt cheap. A good definition of dirt cheap: so cheap, that the new entrant not only delivers high value to the end user, but in doing so funds its own growth. Tech revolutions are born when a new tool offers more service than the old tool, at a lower price. That is the killer combination no one can fight. Wind and solar and storage, despite wishes, dreams, and yearning, are cheap—just not so cheap that they can invade without resistance.

Cold Eye Earth continues to use a very simple model to forecast total annual growth from wind and solar power generation. And it’s almost comical how well this has worked the past few years, given the formula. We take the trailing five year growth rate and apply it to the new year. That’s it. So simple, it’s frankly simplistic. How did it perform this year? Well, the model forecasted that 774 TWh of new generation would be supplied by wind and solar in 2025 and the actual figure, already discussed, was 844 TWh. That’s a miss of about 9%— a larger miss than in years prior—but given that the model expends no effort, does no research or counting or analysis of global solar, not bad. One should also consider the rather wide error bars that the big analytical groups maintain as each year proceeds. They are doing the real work of course, the hard work, canvassing the globe for solar information—and yet that more sophisticated method has to be updated continually. Finally, revisions across global energy typically can run in the 3 % to 5% range. So a 9% error is not great, until you consider that it’s a costless forecast. What does the model say for 2026? Why, another spectacular year for wind and solar growth, of course.

Because of wind and solar’s great year in 2025, the trailing five year growth rate now lifts from a previous 16.8% to a current 17.6%. That’s great! When we apply this higher rate to last year’s level of 5525 TWh, Cold Eye Earth projects global wind and solar will advance by a huge 972 TWh this year (nearly 1000 TWh!), reaching 6497 TWh.

The project to decarbonize the world’s energy system remains a Tale of Two Cities. It is the worst of times to ponder the stubborn viability of incumbent fossil fuels and it is the best of times to marvel at the global buildout of low carbon energy. Dickens himself would surely be impressed with the clean air that now governs most world cities from London to New York. But then, someone would inevitably have to explain to him the problem of carbon build up, and the advance in global temperatures which are already affecting the world.

—Gregor Macdonald.

Subscribe now

The IEA wrote a brief obituary for the Iran war’s oil crisis, which turned out to be not much of a crisis after all. At most, the world suffered roughly eight weeks of elevated oil prices around $100 a barrel in March and April before oil began its decline. Moreover, if $100 oil for a couple of months qualifies as “the greatest energy crisis of our lifetimes,” then the world is in exceedingly good shape when it comes to energy disruption risk to the global economy. $100 oil is cheap, and not even remotely comparable to $100 oil 20 years ago, as inflation and wages have moved strongly ahead over those two decades. If your intuitions are inclined to fight this truth, go right ahead, but you will come up empty-handed. As we go to press, oil has of course responded to the Iran war’s conclusion, as the price has now drifted down below $70 a barrel. That price is actually equivalent to about $42 a barrel in 2006—but no one was paying $42 for oil in 2006. No, they were paying much more that year: The average nominal price in 2006 was $55 a barrel, which in inflation-adjusted terms is about $91 a barrel, roughly 30% higher than today. (And as we know, oil prices went far higher, for longer, during that period.) Cold Eye Earth will say it again: fossil fuel prices have not kept up with inflation, and have failed as an inflation hedge. You’d be better off having invested in other liquids, like wine or Scotch whisky.

But back to the “energy crisis” story. The IEA understandably focused on two structural factors that kept the Strait of Hormuz blockage from evolving into a global economic crisis: China’s enormous stockpiles of stored oil and the slow rate of global demand growth. (Cold Eye Earth discussed these and other factors in the 15 June issue, “Stocks and Flows.”) Here is the IEA in its summary:

The global oil market went into this crisis with significant buffers. In February, before the outbreak of hostilities, the IEA’s market balances indicated a surplus of 3.7 million barrels a day for 2026 as a whole.

Global oil supply had been running ahead of demand for 12 months, with the amount of oil in storage reaching 8.2 billion barrels. China, notably, had for many months been hoovering up huge amounts of oil imports and putting them in storage. This, combined with demand reductions both by refiners and end-users enabled China to slash crude oil imports by 40%—or 4.6 million barrels per day—between February and May, helping significantly to ease wider pressures in the global market.

When the flow of oil tankers through the Strait of Hormuz plummeted in early March, the massive global surplus that had built up over the past year gave oil markets a valuable cushion from immediate impacts.

A helpful rule of thumb to keep in mind in commodity markets is that when the rate of demand growth can be met easily by the rate of supply growth, the futures market enters a relaxed state. That’s why, outside of crises—whether small or large—the global oil market has been unconcerned about supply for many years now. No, it’s never “easy” to bring on the marginal barrel of supply, but when the rate of demand growth consistently struggles to sustain itself above 1.00% per year, it gets a lot easier.

Moreover, we should probably deflate global demand rates even further over recent years, because the crisis revealed just how much stockpiling China had been undertaking. The result may seem obvious, but it’s worth stating it clearly: Oil has been so affordable for so long that China correctly concluded that importing more oil than it needed and stockpiling it delivered a good return on investment. When the Iran war began, therefore, China’s long-term thinking paid off, essentially offering relief to the entire world for the simple reason that those stockpiles served China’s demand. China took responsibility for itself and thus, by implication, delivered a benefit to the world. The contrast between China’s adult-level decision-making and the impulsive, unplanned, authoritarian decision by the U.S. to make sudden war on Iran is disturbing. The results of that great divide could not be more clear.

One sector where the rate of demand has gone into the stratosphere, while marginal supply is essentially unable to respond, is the newly hot AI trade in semiconductor memory. Commodity analysts are typically alert to the risk that cyclical rebounds in demand can often arrive after periods of underinvestment, which invariably supercharge prices as suppliers are caught off guard. We have seen this many times in copper, uranium, and other metals, but this time analysts completely missed the fact that memory—historically a kind of simple if not dumb component in electronic component manufacturing—would become critical to the functionality and growth of AI. It’s been called one of Wall Street’s biggest whiffs that few saw this coming, but memory, it turns out, may be just as critical to AI growth as the far more complex and higher-priced chips made by companies like Nvidia.

One way to think about this is that the AI buildout has just received a lesson in Liebig’s law of the minimum (from Wikipedia):

Liebig’s law of the minimum, often simply called Liebig’s law or the law of the minimum, is a principle developed in agricultural science by Carl Sprengel (1840) and later popularized by Justus von Liebig. It states that growth is dictated not by total resources available, but by the scarcest resource (limiting factor). The law has also been applied to biological populations and ecosystem models for factors such as sunlight or mineral nutrients.

In other words, you can gather all the fancy-pants components you need to mount a project, but if you are short the normally simple stuff like soil or sand—or in this case memory chips—you are going to be highly constrained. Here, memory is placing a kind of tax on AI growth in the same way oil, when it was more scarce and demand rates were higher, often placed a tax on economic growth. (Indeed, there is a view that when oil prices rise in response to strong global growth, that oil is itself doing the work of central bankers by tightening conditions around global growth.)

Unsurprisingly, semiconductor memory has historically been thought of as a routine commodity product, oscillating between glut and scarcity. Oh, how the tables have turned! In the aftermath of Micron’s earnings report last week, many gasped that much of the capex being spent by the large hyperscaler companies (Google, Meta, Microsoft, Amazon) would now be going to the memory makers. What an outrage! But what did they expect? In the first of the recent two-part series, “Markets I—Won’t Get Fooled Again,” Cold Eye Earth suggested that many aspects of this AI hypergrowth period are reminiscent of a commodity bull market, often born on the ashes of the previous bear market:

Despite the tech-focused character of the AI mania, elements of this current boom have more in common with a commodity bull market. You will recall that the dot-com boom drove sales of personal computers, the laying of fiber-optic cable, and the early appearance of mobile phones. But mostly, dot-com was a connectivity explosion that used existing networks (telephone lines) and distributed itself rather quickly and easily using software. (The AOL disk in the mail comes to mind). The AI boom, however, has unleashed a powerful wave of demand for power sector equipment: turbines, transmission lines, and associated grid infrastructure.

One is also reminded of a rather prescient call made by T.S. Lombard back in 2021, asserting that semiconductors were the new oil and that the South China Sea (surrounding Taiwan) could potentially become the new commodity chokepoint for the world. Here we start to make contact with the market analogy that Cold Eye Earth has in mind, that the AI boom shares qualities with the unexpected and thus extremely powerful peak oil bull market in oil that ran from 2003 into 2008.

This constraint in available manufacturing capacity for memory is of course the second bottleneck to arise from the AI infrastructure buildout, the first one being the limited onramps to power supply, as many proposed data centers are not being built because they can’t plug in to the electrical grid. (Also true for myriad wind and solar projects.) For those who are against data centers and want to see them curbed, it may end up being internal supply constraints, not community protests, that sustainably curb their growth. The call, as they say, is coming from inside the house.

Lingering question: Having embarked on the pursuit of artificial intelligence, why was it not more broadly understood to the tech sector that to mimic intelligence you would also need to mimic memory?

The world has made enormous progress over the past 20 years in connecting populations to the electricity system. But there is much more to be done. As recently as 2010, the largest collection of people without energy access happened to reside in India. But now, in the wake of India’s heroic effort to extend electricity to rural regions, the focus has turned to sub-Saharan Africa, where, according to the IEA, 550 million people are still waiting to plug in to grids—grids that, in many cases, are yet to be constructed. The IEA’s just-released Energy Progress Report estimates that the total global number of people lacking electricity currently stands at 655 million.

Read more

As the Iran war winds down, a story worth telling is how, these past several months, we managed to escape an even more dramatic oil price spike. Helpfully, the story and its explanation are not that complicated. Saudi Arabia’s historic investment in bicoastal export capacity tells one part of the tale, and China’s gargantuan oil reserves, steadily built up over years, tells the other. We can probably add in a dash of real-time demand destruction, as global consumers not only reacted to the initial increase in oil prices but anticipated a lengthy period of elevated prices. True, oil did spike and prices remain elevated. But we never suffered a systemic seize-up price like the $200 and even $300 levels many feared. As Cold Eye Earth goes to press (on the same weekend Iran and the U.S. are supposedly about to sign a peace deal), both WTI and Brent oil are back under $90 a barrel.

For years now, oil has suffered from a sustained period of slow demand growth, and technology advancement that keeps reliably pulling down the cost of creating the marginal barrel. This setup has persisted for years, and it is the very best explanation for why energy equities have been such a poor investment (ex temporary price spikes and geopolitical dramas). This new topography for oil, where nothing happens for a long time, a brief interruption arrives, and then oil returns to another lengthy and listless period, says a lot about oil’s overall prospects and fortunes. It’s not great.

The fact that China was able to amass over years a 1.4 billion barrel reserve mountain without disturbing the price of oil also indicates how absolutely lousy the oil business has become, as extraction costs have risen along with inflation but the price of oil itself has been tame, and actually quite cheap. Twenty years ago, Americans were paying on average $66 for a barrel of oil, which after adjusting for inflation would be about $109 today. Stop here and consider the implication: Our “energy crisis price of oil” during the 2026 Iran war this spring took oil prices roughly to the same average recorded for all of benign, humdrum 2006. Again, Cold Eye Earth has been pointing out for about a year now that oil and gas prices today, after adjusting for inflation, are akin to walking into a grocery store and seeing that beef prices are the same as they were two decades ago. Now that oil has returned to below $90 a barrel, it’s once again dirt cheap when compared with wages.

Another perspective worth considering is that, since 2005, global oil demand has either reached or exceeded 2% growth only five times. And of those five instances, at least two of those years have to be considered outlier, rebound years, in which the global economy roared back and oil demand crawled out of a deep hole. It sure is easy to post a “growth” number when you’re starting from the depths.

A number of longtime oil bulls have been frustrated for years about oil’s sluggish performance since the go-go days of 2002–2008. But when you convert to a regime that sees global oil demand rising, at best, at about 1.0% per year, you have to expect that the futures market will see the combination of weak demand growth, steady supply growth, and the reserve-building those conditions foster. If you are familiar with the concept of stocks and flows, what has happened to the oil market since the great recession is that both stocks and flows increased. By how much? This is how much: The world mostly lost the Strait of Hormuz for 90+ days and the price of oil never reached the boiling point!

What comes next? A continuation of the past decade: investors baffled that quality names like Occidental Petroleum are dead money and can’t get off the ground, stocks return to a building phase not just in China but in the OECD, and the relentless ebb of oil demand growth as more of the world’s transportation system electrifies. Oh, and a kicker: While the Saudis were indeed quite correct over a decade ago when they added, at great expense, bicoastal export capacity, we must now acknowledge that it was China, not Saudi, that assumed the role of The Central Bank of Oil during this most recent energy disruption. Assuming they rebuild those stocks over the next few years, we are obligated to keep extending that title to China.

Breaking news on Sunday afternoon, June 14: Iran and the U.S. agreed to their peace deal.

The strategic importance of oil is fading, while that of artificial intelligence is rising. Over the weekend, the Trump administration placed an export ban on the latest AI model from Anthropic, shocking investors across the AI community. As is typical with the current government, everything about the sudden announcement felt rushed, ad hoc, and possibly petty and vengeful. David Sachs, the AI czar, tried to calm the waters over the weekend, saying both sides would almost certainly come up with a compromise. But then Secretary of Defense Pete Hegseth let the cat out of the bag with a dismissive post on X that understandably made many conclude that, yes, this was more about vengeance than any scholarly analysis of Anthropic’s model.

Read more

A surprisingly relaxed year for global power growth made it easier for wind and solar in their ongoing effort to cover marginal demand in 2025. According to the IEA, in fact, combined generation from new wind and solar, along with the help from nuclear power growth, almost met total demand. Impressive! As energy transitionists understand, that is the holy-grail crossover point that, if breached—and then breached repeatedly—would inevitably trigger emissions declines in the global power sector. Alas, we are still not there.

According to the IEA, total global power demand rose by an unexpectedly weak 850 TWh in 2025. That represented just 3% growth during a period when the IEA had confidently forecasted steady, sequential 4% annual growth for last year, this year, and the next. In the chart below, you can see the year-over-year deceleration in total demand, and how combined wind and solar growth fared by comparison. While nuclear is not shown, according to the IEA, it alone advanced by 34 TWh. Ergo, when you add that to wind and solar’s 799 TWh of growth, the total nearly covers the entire 850 TWh.

Just to remind, Cold Eye Earth remains uncomfortable with the catch-all category known as “renewables,” which is a kind of jambalaya stew composed of everything from highly unstable hydropower to dreaded biofuels (ugh). This is not to say that emerging sources like geothermal, for example, aren’t both exciting and promising, but rather that sources like wind, solar, and nuclear are superior for tracking purposes, because they adhere to manufacturing and thus learning rate dynamics in addition to being either zero or near-zero carbon sources. Moreover, as Cold Eye Earth has pointed out for years, wind and solar are the clear leaders—by leaps and bounds—beyond all the other “renewable” sources. If you’re looking for the tip of the spear in how we’ll eventually decarbonize the power sector, look always to those two (with nuclear just behind) until some radical scientific and economic breakthrough occurs that can scale. Hint: If pressed to identify where that breakthrough may arrive among geothermal, hydropower, biofuels, wave (ocean) energy, hydrogen, and some as-yet-unnamed energy source, the current choice would have to be geothermal.

Of particular note in the IEA’s assessment of last year: The agency reported that the big swing downward in power demand was driven largely by a more relaxed year for cooling demand in India and Southeast Asia. This reminds us that climate change and climate volatility are increasingly competing as major factors in our more routine efforts to track trajectories from fossil fuels, wind and solar, and overall economic growth. Indeed, it sounds like the downshift in electricity growth was a major factor in bringing total global growth for all energy down to a slower pace in 2025:

All major energy fuels and technologies grew in 2025 – but at very different rates. Overall global energy demand growth slowed to 1.3%, just below the average for the previous decade. Slower economic growth and slower growth in energy-intensive industries in some regions, lower cooling demand, and faster efficiency improvements all contributed to slower demand growth.

To put some numbers on this thorny problem, it is estimated that global air-conditioning accounted for 7% of total power demand in 2022, or about 2,100 TWh out of 29,000 TWh total. That’s just about exactly half the power consumed in the U.S. that same year. At this scale, uncertainties over how much new demand will come from, say, data centers, start to fade rather quickly because we literally find ourselves in an era where volatility in cooling demand is so substantial that it alone can turn the knob on global power and our efforts to decarbonize it. Here we run into an old nemesis, a flaw in human thinking: We are not good at understanding scale. This flaw is how we get such axioms as “One death is a tragedy, but a million deaths is a statistic.” This bias leads us rather habitually to discount the system as it currently exists—colossal, profoundly embedded, hard to dislocate—while overfocusing on local disputes around a transmission wire, an offshore wind farm, or the location of a battery installation.

Why don’t wind and solar simply race ahead and quickly bury continued growth from fossil fuels in the global power sector? Why don’t they put this contest to bed already? Why does the volume of new generation each year from wind and solar pretty reliably come up to the line of marginal growth, and then halt? Why don’t wind and solar simply overrun the system, in a spectacular viral distribution? The reason is that wind and solar, while cheap, are not nearly as cheap when compared to other historical technologies that quickly overwhelmed their markets. Essentially, wind and solar are cheap enough to be the obvious choice for replacing retiring power generation, but not so cheap as to economically destroy existing generation. Just a few examples from the past:

Read more

Financial crises tend to burn themselves into the psyche of nations. The South Sea Bubble, Tulip mania, the Crash of 1929, the German hyperinflation of 1930’s, and the Asian Financial Crisis of 1997 are just few examples of market dramas that inevitably shaped future monetary policies, market structures and each nation’s perception of risk. The traumas need not be restricted to markets. China’s multiple, historical famines—in addition to the 20th century Great Chinese Famine—have created, according to some scholars, a kind of hard, red line around food security. A study at Berkeley in fact showed that the last famine continues to have negative impacts on people’s health. The first job of any Chinese government, it’s been said, is to entirely foreclose upon the risk of starvation.

The ghost of a past trauma that clearly animates current thinking around the AI investment boom is of course the Dot Com bubble—an exuberant, deflationary growth event that caused the U.S. to understandably conclude it was ground-zero for tech futurism, and that this cornucopia of progress would press onward, well, “forever?” While a tremendous volume of crucial internet infrastructure was indeed deployed through that mania—foundational to today’s far more advanced capabilities—the bubble was ultimately humiliating, and wound up wrecking the careers of many people who both successfully, and then unsuccessfully, rode that wave. On a more humorous note, “Dot Com” finally replaced the 1929 crash as the go-to model for traders seeking a template to guide them through a historic market meltdown.

Which brings us to today: there is currently a feverish model-making community attempting to build a road-map for how and when the current tech-driven bull market will end—as they all must end—in tears. If you haven’t been following along, the AI bull market has, so far, deputized a number of sectors already from semiconductors to electrical utility and power, and more recently to optics and computer memory makers. Unsurprisingly, the various Magnificent Seven names have participated in their various ways, from the concentrated bet expressed through Nvidia, to the AI focused Google, and to other hyperscalers trying to find their footing like Microsoft to Meta, and Amazon. For the modelers of the coming bust, the imminent initial public offerings of SpaceX, Anthropic, and Open AI are classic over-the-bow warning shots that we are entering the final stages of the current AI bull run, or at least, “we are much closer to the end, than the beginning.”

Mostly though, the roadmap currently in favor among the bubble watchers overlays two time series: The number of days (or weeks, or months) which passed from the introduction of the Netscape browser in 1994/1995 to the top of the market in 1999/2000 vs. the number of days which have passed since Open AI introduced ChatGPT in 2022, through today. Some use the Nasdaq 100, or the Nasdaq composite, or even the SP500 as their market measure. In the version below, which comes from Fidelity, the chartmaker is using the Inter@ctive Week Internet Index, and finds that were the current market to match Dot Com, the bull run would top out later this year.

But here is another variation, this time from Bespoke, that anchors Netscape to late 1994 and uses the much broader Nasdaq composite index. Following this particular historical roadmap suggests the current AI driven bull market has almost another two years to run.

Wise investors will generally take note of these models, and then proceed to not rely on them. In other words, these charts function best as starting points for your own thinking, your own analyses and your risk management. Here, we must pay some due to Nassim Taleb, who insightfully pointed out that humans are so desperate for maps to guide them that they will happily take the wrong map and use it anyway, if it makes them feel better. While we know that all bubbles end in a bust, it’s not clear that the Dot Com era is the best fit for the current one. Cold Eye Earth has a slightly different map, therefore. After all, the goal of this game is to not get fooled again, right? Read on.

Read more

Cold Eye Earth is going to have to roll forward its publication schedule by a week.

Sorry about that. We’ve had a few technical issues to deal with of late (a nasty round of factory resets and clean installs on the Mac collection) and in addition your editor has a historical sleep-problem that’s usually contained to manageable levels but has flared up recently to a feverish pitch. For readers who are familiar with that particular issue, your bookshelf is no doubt already filled up with books like these.

All this said, a couple of two-parters are in the works: one on current markets, and an overdue deep dive into AI that incorporates all that’s happened anew in the past few months. The two-part series on markets will begin a week from today, May 25, 2026.

—Gregor Macdonald

Senate candidate James Talarico of Texas has voiced full support for the oil and gas industry, in a reminder that Democrats will have to broaden the tent if they want to expand their numbers in the Senate and the House. Talarico can be thought of as a kind of experiment: After numerous near misses, what combination of positions would actually give a Democratic candidate a chance to win a Senate seat in Texas? So far, Talarico seems to have hit upon the optimal mix. His background is rooted in a distinctly more progressive version of Christianity, and this particular signal—a friendly embrace of the oil and gas industry—is clearly designed to loosen calcified distinctions in the Lone Star State. The strategy appears to be working. Recent polling and prediction markets show that regardless of whether Cornyn or Paxton is the nominee on the Republican side, Talarico has converted the race into a tossup.

Talarico made his comments on the Ranch Hall podcast earlier this year, declaring that the oil and gas industry is critical to the Texas economy and to working families. He also embraced the growth of wind and solar power and correctly pointed out that Texas was a 20th-century giant in energy, and that renewables point to the state’s continued dominance in the 21st. What’s intriguing to ponder here is that Talarico will surely be criticized within the Democratic party for making such accommodating remarks, but the irony is that voicing support for the existing oil and gas industry is almost entirely without any ramification, and in the end is mostly costless. As a presidential candidate, Barack Obama took the exact same position. Talarico repeated Obama’s phrase too, an embrace of an “all of the above strategy.” And what did Obama do, practically, while in office? Instead of either promoting or interfering with the U.S. oil and gas industry, his administration got busy, creating the conditions to help wind and solar reach a takeoff point. What’s not to like?

Cold Eye Earth has laid out previously the observation that the “words” uttered by politicians, more often than not, have little to no bearing on the fortunes of the U.S. oil and gas industry. The March 6, 2023, issue, Energy and the Presidents, showed that oil production declined quite seriously under the presidency of Bush (an oilman from Texas); soared under Obama, who presided over the biggest surge in “all of the above energy” in half a century; that coal continued to crash under Trump (I) despite his promises; and that Biden oversaw new all-time highs in LNG exports and U.S. petroleum production. But at least 90% of Americans don’t know anything at all about these facts. In the suburbs of Columbus, Ohio, no one is digging through PDFs and Excel files from the EIA. From that 2023 issue of Cold Eye Earth:

Americans mistakenly believe political rhetoric governs energy policy, but the fact remains that the United States runs a free market energy system. And in a free market system, price and technology are in control. Correlations between the political perceptions of various Presidential administrations and actual, realized energy outcomes remain exceedingly low, and in many cases negative. Yes, the EPA (created by Richard Nixon!) has churned out regulations for a half-century and American presidents can certainly bend both the discourse, and the regulatory state, in their preferred direction. But it’s truly remarkable to consider how the last four Presidents, in this century, oversaw energy outcomes that blatantly defied commonplace public perceptions. And these were not marginal outcomes, but central, large scale outcomes.

There’s additional confusion because many people who do take an interest in energy are not just over-persuaded by words, or symbolic actions, but do not see or understand scale. They think that when Biden cancelled the KXL pipeline that it was a momentous event. Indeed, both the political right and the left had reactions to the announcement totally out of scale or proportion to the policy. Same too with the Biden administration’s approval of the Willow project in Alaska. These actions were almost entirely symbolic, and barely show up on a map of the U.S. energy system. The U.S. energy complex is a vast, sprawling juggernaut of enormous production, consumption, and exports. And this is what Talarico more cleverly, more astutely understands. Should he win that Texas senate seat, he will have virtually no impact on the future of U.S. energy which will roll onward, like a big wheel.

Read more

Scientific discoveries tend to cluster, arising roughly at the same time from teams or individuals working separately from each other. The phenomenon has a name— multiple discovery or simultaneous invention—and there are many examples: natural selection, calculus, the telephone. Contemporaneously, it’s been observed that more Nobel Prizes have been going to multiple persons—who, again, were working independently of each other.

Of course, the concept has great relevance to the progress of artificial intelligence. And especially to the understandable unease around AI, and even to call a stop to the development of AI. Unfortunately for those with that view, it’s simply not going to be possible to halt or even slow the forward march of this technology. Yes, an individual country or domain could undertake such action, but it would only serve to establish their isolation from the rest of the world. Indeed, in a very compelling example, historians have speculated that the Ming Dynasty successfully suppressed the growth of maritime technology and withdrew its society from engaging in any seafaring exploration. Effectively, this ceded all venturing to the rest of the world. So yes, you can stop progress for yourself, but that won’t stop others.

When we consider the intuitions of those who habitually want to halt technological progress, we can further speculate that some portion of society has probably always taken such a position: against seafaring, against railroads, against cars, against telephones, against stem-cell research… and now this same layer of society is back again in full swing, against AI.

A good proxy for these inclinations can currently be found in the anti-data center movement, which unfortunately is arising most strongly on the left side of the political spectrum. No surprise there. Led by plodding, linear, zero-sum thinkers like Bernie Sanders, groups that have successfully stopped transmission projects and even lots of wind and solar projects now look to do the same with data centers. Mr. Sanders—who apparently has never seen a complex-contingent system, and doesn’t care to know how they work—represents a puritanical New England tradition that is eager to jump into the fray and take the side of those who abhor building and construction and development of any kind. Sure, ratepayers are justifiably concerned about the impact of future electricity prices as power demand grows, but news flash: Global power-sector growth is already on a strong upward march, because energy transition itself makes that inevitable. If you are against data center growth because it will place upward pressure on grid demand, then you should also be against electric cars and buses and trucks, electric heating, and the entire effort to swing as many industrial processes as possible from the realm of fossil fuels over to the power sector.

Now, if the federal government were not currently in the hands of a kakistocracy, we would likely be engaging in constructive approaches to these challenges in everything from grid management and transmission access to AI safety policy more generally. We’d have professionals like Jigar Shah (former director of the DOE Loan Programs Office) back at the helm working on approaches that welcome the spectacular power sector-growth necessary for transition with a solution set that actually addresses those needs. Shah currently has a post up at LinkedIn, The Data Center Reckoning: Why Communities Are Saying No — And What Has to Change, that’s right on point.

In the current government, however, we have a far more serious and widespread case of derangement, where the impulses to control everything are fully authoritarian. This has given way to some of the most awful outcomes of all, in which natural gas turbine growth, to take just one example, is now flourishing “behind the meter,” thus adding further lock-in to what is already a very problematic path dependency forming on natural gas. Worse, the clowns and sociopaths sprinkled across this administration demonstrated in the Anthropic dustup last month that they would, in the face of what they perceived as intransigence on the part of the company, blacklist them—effectively imposing a corporate death penalty on the AI startup. And we actually witnessed “Department of War” officials going full-blast on social media, accusing the Anthropic CEO of being a liar.

Political reality suggests, however, that the real threat of AI—job losses, possibly on a large if not very large scale—is going to animate the AI discourse once those career-ending effects start to come into play. You may have noticed that Anthropic’s CEO, Dario Amodei frequently steps to the microphone these days to warn that AI could smash both existing white collar jobs and their future growth. While not everyone agrees, we already have signals from the market that the prospects for computer science majors—probably the hottest undergraduate major of the past decade or so—have suddenly turned downward.

In recent decades the U.S. has developed into a country with a special talent for stagnation, and an inability to solve problems or to get anything done. We can’t build high-speed rail, we’re way behind on EV adoption, we can’t repair or replace our tunnels and bridges, and we have infrastructure projects, especially in the power sector, that have been bottlenecked for years. Governance, meanwhile, has lapsed into such an inefficient state that the new mayor of New York is getting rave reviews for fixing a series of small things that have long since been resolved in the rest of the developed world.

As people age, they become more risk-averse and increasingly opposed to change of any kind. The U.S. currently has an aging population, with the replacement rate falling to very low levels through a combination of lower fertility rates and lower immigration. This is a recipe not for a correction in our ways but for a deepening of our current course, in which there’s strong bipartisan sentiment to build nothing, change nothing. A favorite example: the long-delayed L.A. bike path project which, according to reporting last year, should have been completed by now. Your faithful correspondent actually did some reporting on this effort seven years ago, in 2019, and it should be said that the portions of this project that have been completed are amazing. However, the hard part—connecting the southern portion to the northern portion near downtown—remains fully stuck in the mud. Notably, in an example of how America’s neglected problems can stack on top of one another, one of the concerns about the formation of the center-city portion of the bike path was that it would negatively impact the homeless population. But this was just one of myriad hurdles that transformed the project into what the linked SFGate article called a “bureaucratic hell.”

One public figure who’s currently sending the message that the task of the next administration cannot be restricted to “simply putting back things to the way they were” is Pete Buttigieg, currently on a speaking tour in states like Kansas and Oklahoma. While oratory and public persuasion are not the only deciding factors in making a presidential run, Mayor Pete has shown again and again that he can marshal words against our current Empire of B.S. with consistent success—and to the distress of his interlocutors!

More specifically, as the former DOT secretary, Buttigieg got a full dose of the regulatory miasma that has built up over a half century in this country, and described it this way: “We got so good at stopping all the bad things that we stopped all the good things from happening too.” Buttigieg hasn’t declared yet for 2028, but he seems to have a “theory of the case,” and so his decision to run seems inevitable.

Read more

Americans aren’t buying what President Trump is selling. In a televised speech to the nation last week, the president once again laid out a series of incoherent arguments in his ongoing attempt to explain the war. The attack on Iran, apparently, is simultaneously a war of choice and something that was necessary. The closing of the Strait of Hormuz, meanwhile, occupies not two but several conflicting framings: an action taken by Iran that has nothing to do with the war; a condition that will dissipate naturally all on its own; and a problem the rest of the world needs to solve, because the U.S. is on the verge of washing its hands of any responsibility for its future protection. The president followed up his speech several days later with a bizarre Truth Social post on Easter Sunday, in which he praised Allah, dropped an f-bomb, and threatened Iran with total annihilation if they do not open the Strait. Betting markets responded to the whiff of executive-branch desperation by pushing the Impeachment contract at Kalshi to new highs, raising the odds to nearly 26% that Trump is both impeached (in the House) and convicted (in the Senate). Just to add, that is functionally a parlay bet, relying on not one but two outcomes to pay out.

A practice most professional bullshitters generally respect when baffling audiences is the disciplined avoidance of any details. Here the president failed to adhere to that tactic when he asserted, several times and forcefully, that the passage of ships through the Gulf has no effect on the U.S. or its economy. His reason? We produce our own oil and don’t need any Gulf oil. Observers have often remarked that Donald Trump genuinely doesn’t understand how complex, contingent systems work—such as the global economy and all its interlocking parts—and here we have a tidy case study in the president’s deficiency.

But the president also committed a cardinal sin of politics: You cannot assert to the public specific claims that are easily falsifiable. Gasoline prices are available for all to see, and on a national average are back above $4.00 per gallon—last reached in 2022, at the outset of the Russia-Ukraine war. His assertion that the U.S. is insulated from the effects of high oil prices only has relevance to the country’s trade balance—not to the effect on American consumers. Unlike some other petrostates in history, the U.S. does not subsidize prices at the pump. We pay the global price. There is no local price. Clearly there’s a common intuition that big producer states of commodities confer cheap local prices to the domestic populace, but that’s possible only if you are not running a free market in those commodities at home. It’s easy to imagine someone making this same mistake, say, upon arriving in New Zealand assuming that, because that country is a huge exporter of dairy products, local milk will be cheap. Nope. The U.S. could double its production of oil and New Zealand could double its output of powdered milk and Americans and New Zealanders would still be paying the global price for oil and milk respectively.

Here, we can assert a maxim: There is no way to reduce the hit to consumption from a spike in oil prices by producing more oil. If that were true, the U.S. would be basking in cheap oil prices right now, having nearly tripled its output of petroleum (not just crude) since the low-output years, around 2005. On the contrary, the U.S. has done basically nothing to reduce its structural dependency on gasoline—though you could assert it’s a form of accomplishment to have held consumption steady as the population grew over the past two decades.

Read more

Forever is a long time. And yet the events now set in motion by the U.S. attack on Iran are likely to give way to a new energy security regime of very lengthy duration. While energy history is anchored by the great, singular transitions, it’s also the case that wars, embargoes, and shortages also play a reshaping role, cutting deep grooves in the psyches of nations and kicking off new policy regimes. In the aftermath of Russia’s attack on Ukraine, for example, the EU’s natural gas consumption stepped down about 20% from its 2021 highs; and that demand destruction has roughly been maintained as the EU has raced forward to adopt heat pumps and other gas-saving measures. Going back a bit further, the 2004—2008 global bull market in oil, which resolved into a fearsome spike, put a halt to decades-long OECD demand growth. And that growth never recovered (see chart below). Going even further back, oil enjoyed a dominant 50% share of all energy consumed in the world coming into the oil crises of the 1970s. Today? Oil has fallen to just a 33% share of total world energy consumption. Pain, as always, is a stern teacher.

When George H.W. Bush attacked Iraqi forces back in 1991, driving them out of Kuwait, he was criticized for not following up Desert Storm with a march on Baghdad. American hawks wanted Bush to “finish the job.” But Bush, among several considerations, understood that despite the serial tensions that defined the Middle East then (as now), a kind of equilibrium had long since formed around flows of oil from the Persian Gulf, and disturbing that equilibrium was far more likely to open up a box of furies, and other nasty surprises. Here we might recall the Odyssean dilemma, in which our hero has to navigate a strait and choose whether to sail closer to a six-headed monster, on one side of the passage, or a whirlpool on the other. Figuring the whirlpool was certain death, Odysseus sailed nearer to the monster, losing a few men in the process, but survived. March on Baghdad? No way, concluded Bush.

By attacking the country that owns the most beachfront real estate on the Strait of Hormuz, you might say that President Trump, by contrast, has sailed closer to the whirlpool. Everyone knows that Iran is a brutal theocracy and has been meddling and causing trouble since forever through its proxies around the region. Attacking Iran, however, with the intent to drive them into submission has pierced the stability of a system that is far larger than either the U.S. or Iran: the global economy, and its continued dependence on petroleum. Proper war planning for an attack on Iran ideally would have recognized that this larger system requires at least as much, if not more, attention and care than the direct attacks on the Iranian state.

What are some of the painful adjustments that therefore are likely to evolve as the risk premium around oil remains elevated for some time to come? Let’s go through some of the most obvious.

• In a repeat of 2022, global coal consumption is likely to soar. The pain point in 2022 was natural gas via the Russia-Ukraine war and Putin’s gas policies, but this time around it’s the reduction in Gulf LNG capacity that’s going to deliver a lot of pain to Asia. Global coal consumption reached a new all-time high in 2022—and then hit two more highs in 2023 and 2024—entirely undoing the moderate progress made in coal demand-reduction over the previous decade. A country like Japan, which has an enormous industrial base but must import energy to run it, is at great risk in this scenario. If Japan can’t get enough LNG, it will partly turn to coal, and you can apply that to other nations across Asia that manufacture but are dependent to some extent on imported energy. As someone quipped on the internet this week, maybe it’s time for Japan to do its part to clear the Strait of Hormuz:

• U.S. exports of both coal and natural gas in the form of LNG will quickly rise to meet current export capacity; and if demand for those two fuels doesn’t let up, the U.S. energy economy will start exploring ways to expand that capacity. We must also remember that the U.S. has been expanding exports of both these fuels steadily over the past decade. U.S. coal exports have risen over this time partly due to the reduction in domestic demand. LNG capacity has risen in part because North America has economically recoverable natural gas reserves that are nothing less than gargantuan. So export momentum is already in the mix. Finally, U.S. petroleum product exports (not crude oil) may be constrained from increasing much as the combination of lower lower global supply but also lower global demand makes may offset each other. However, there’s no question that prices for those products will rise.

• Leaving aside the associated snarling of global supply chains that may constrain all production of any kind, demand for renewables and planning for more renewables will soar, along with demand for battery energy storage systems (BESS). The Iran war is shaping up to become a second major blow in less than five years to fossil fuel dependency in the power sector. The world will become weary. The Asian countries, led by India, are the domains where we are likely to see enormous jumps in renewables growth. China already builds renewables at a frantic pace, so the rate of change possible in China is far less. India, meanwhile, has a lot of catching up to do, and this probably triggers that long-overdue need to build new solar capacity.

Let’s pause here to remember a golden rule on which you can rely: Fossil fuel price shocks are a wonderful gift to producers of those fuels in the short term but are terrible for those producers in the long term. ‘Twas always thus.

Read more

Crafting a working description of our current energy transition continues to be a challenge. We remain stuck in a version of the classic Dickens novel, A Tale of Two Cities, in which it’s seemingly the best of times for the continued buildout of wind, solar, and batteries while we are forced to endure the worst of times when it comes to emissions, which keep on growing. A recent editorial piece in the New York Times by David Wallace-Welles, Don’t Look Now, but the Green Transition Is Still Happening, illustrates the challenge perfectly. And the problems start right away, with the title, which depends greatly on how one defines transition.

Is our current energy transition one in which old forms of fossil energy are being replaced by low carbon sources of energy? Not really. A more accurate description would be that energy transition has gone far enough to see low carbon energy dominate marginal demand growth—thus halting new growth of fossil energy—while leaving all the existing fossil energy demand, unfortunately, in place. The structure of Wallace-Wells’ essay is a familiar one, in which you kick off the argument with an extensive list of stellar low carbon achievements: This January, Norway sold only 7 vehicles with an internal combustion engine; Pakistan now sometimes produces more electricity from rooftop solar than from its main power grid; 90% of new U.S. power capacity is composed of wind and solar. And so on.

Regular readers of climate journalism will recognize this touting of achievements and the declarations are well deserved. Cold Eye Earth feels obligated to cite these wins also, and just recently took the view that global road fuel demand has almost certainly peaked, because EV adoption in China has finally reached critical mass, closing the door on global road fuel demand going much higher. So again, there is no shortage of good news items, but the bigger picture is not great. And presumably, no one reading this is confused about the fact that the bigger picture is the only one that matters. Thus, starting about half way through Wallace-Well’s essay, we get the other half of the story:

New power changes the existing system only on the margins, and the old system is pretty dirty. To this point, new green energy has mostly supplemented rather than displaced fossil fuels. Globally, emissions are still climbing, if slowly, and temperatures are not only rising but rising at an accelerating rate, raising the uncomfortable possibility that the world’s climate system may be more sensitive to emissions than almost anyone had bargained for.

So, in the title, we have the claim that the transition is still happening, but in the body we get the admission that low carbon energy has mostly supplemented, not replaced, fossil fuels. Doesn’t sound like an actual transition, does it? Effectively, the headline claim is undermined by the facts and you do have to wonder how long this storytelling architecture can sustain itself before reaching full burnout. At its worst, this format preys upon common human intuitions that if lots of good news is being generated that it invariably leads to a great outcome.

Wind and solar once again covered all the growth in U.S. power generation last year. But power sector emissions rose anyway. The culprit? A huge spike in coal-fired power generation. Here we have a truly unfortunate example of how renewables can meet all marginal demand growth, but a change in composition of fossil fuel utilization (coal took back some share from natural gas) allows emissions growth to find a way upward. Let’s take this counterpoint to energy transition in two parts.

First, the U.S. is a major world domain for wind, solar, and battery growth. Last year, wind and solar reached a nearly 19% share of total U.S. power generation. 2025 was also the sixth year in a row that wind and solar have basically covered all marginal growth in power sector demand growth. You can see that marginal coverage quite easily, just by glancing at the following chart:

This chart is a portrait of additionality, the much criticized term that makes the simple and direct observation (just as Wallace-Wells acknowledged) that energy transition so far is merely supplemental to the world of fossil fuels. Is it a victory to have at least halted the growth of fossil fuels—as demonstrated here, in the U.S. power sector? Indeed it is a victory. However, the reason why renewables are competitive for new build power and less so against incumbent power—both in the U.S. and around the world—is due to simple economics: renewables are cheap, but they are not dirt cheap. The former gets you wins in the marketplace for new capacity. But the latter doesn’t get you radical disruption to existing capacity.

Second, power sector emissions actually rose last year. Coal stormed back into the power sector, causing emissions to rise by 4.2%. There is no question that the transition from coal to a combination of natural gas and renewables has been successful at driving a huge volume of coal out of the system. But the U.S. like many other regions in the world still has extensive coal-fired capacity that can be brought on when needed.

When spare capacity in global oil producers either expands or sits at comforting levels, oil prices have a very difficult time sustaining high levels. This is the regime we’ve inhabited for the past few years in the aftermath of the 2022 Russian invasion of Ukraine, and the associated sanctions which greatly disrupted global oil and natural gas markets.

Read more

The U.S. Supreme Court struck down the Trump administration’s tariff scheme, holding that it was unconstitutional in a 6-3 decision. While there were differences of opinion within the majority about the relevance of the major questions doctrine, the message from justices Roberts, Barrett, Gorsuch, Kagan, Sotomayor, and Jackson was clear: Executive branch actions are reviewable by courts, and the president does not possess unlimited powers. The decision appears to have shattered Trump emotionally, and the president had what most observed as a temper tantrum before the TV cameras, even going so far as to accuse the justices of being under the influence of foreign interests.

Last November, Cold Eye Earth reported on the oral arguments presented to the court, and that’s when we all learned, based on two hours of questioning, that the plaintiffs could probably count on at least six justices to ultimately find against the administration as the three liberals on the court were likely to be joined by Gorsuch, Barrett, and Roberts who would ultimately write the opinion for the majority. The Cold Eye Earth summation at the time works well as a re-cap of the case, now decided:

When President Trump imposed across-the-board tariffs on the entire world this past April, he arrogated powers to himself and the executive branch that no president before him has attempted to claim. Justice Barrett pressed the solicitor general fairly hard on the fact that the powers enumerated in the IEEPA legislation not only do not include the word tariff, but “the actual verbs” cited in IEEPA—investigate, block, nullify, void, prevent, prohibit, regulate, direct, and compel—have not been used elsewhere by Congress as stand-ins or substitute concepts for tariffs. Chief Justice Roberts appeared quite resolute in his questioning, saying several times that tariffs are indeed taxes—a theme also sounded out by Justice Sotomayor. Gorsuch, meanwhile, appeared to be rather agitated by the top-down implications of the case, as he repeatedly laid out a scenario whereby either Congress and/or the Supreme Court hands too much power to the president. The country would discover over time, in his view, that such power, once granted can never be clawed back. By the time the plaintiff’s lawyer, Neal Katyal, rose to make their case, it was already clear the government was facing a steep climb: IEEPA does not grant tariff powers in the text, and even if it did, granting revenue-raising powers to the president is something Congress simply can’t be allowed to do.

The standout event of the majority’s opinion surely has to be the very lengthy, 46-page concurrence by Gorsuch, in which he takes to task three of his colleagues on the left and three more on the right. Each set of three, in his view, has long been playing hide-and-seek with the major questions doctrine, and he suggested pretty clearly that the three conservatives (Kavanaugh, Alito, and Thomas) were denying executive powers to Democratic presidents, and expanding them on behalf of Republican presidents while suggesting the three liberals (Kagan, Sotomayor, and Jackson) were doing the same in reverse, favoring power expansion to Democratic presidents and leaning against Republican presidents (as they did in this case). The man has a point! For Gorsuch, the major questions doctrine and the related non-delegation principle were central to this case. After chastising the conservatives who dissented from the 6-3 opinion for pretending these questions were not applicable, he then pointed to the liberals who had concluded that “the statutory language of IEEPA” was sufficient to deny the president its use as a global tariff vehicle. For Gorsuch, they too were avoiding the major questions doctrine.

Gorsuch’s view may very well have ramifications as Trump attempts to stitch his tariff regime back together using other legal facilities (supposedly) available to him. By foregrounding the robust theme that the executive must generally go through Congress to either enact or sustain tariff actions, he is transmitting to the federal judiciary (in part) a guiding principle they may use as future cases come before them.

Suggested Reading:

Quick Thoughts on the Tariff Decision and the President’s Angry Reaction, by Jack Goldsmith

A breakdown of the court’s tariff decision, by Adam Feldman

Supreme Court, on a 6-3 vote, blocks Trump’s IEEPA tariffs, by Chris Geidner

Unsurprisingly, the Trump team returned to the tariff well the very same day, announcing their backup plan to achieve tariff continuity. Let’s pause here and recall the old proverb “As a dog returns to its vomit, so a fool returns to his folly.” The court’s decision opened an escape hatch for Trump—whose public approval craters to new lows each week—to actually step down from this failed scheme. But, as the saying goes, you’ll never guess what happened next: Trump erected, again by executive fiat, a new slate of global tariffs at a 10% rate, then adjusted that rate upward to 15% in a subsequent social media post. If tariffs were alcohol, we might say the president’s on a bender.

Read more

Coal-fired power generation popped higher in the U.S. last year. Coal consumed to make electricity soared by a fairly distressing 13.8%. But the spike represents no change to the overall trend of long-term decline. Since 2010, coal’s misfortune—a decline of roughly 1,000 TWh—has been exploited by natural gas and combined wind and solar. If we use a baseline of roughly 4,000 TWh to represent total U.S. generation over that time, coal in power has been more than cut in half. The misfortune will continue.

Meanwhile, natural gas and combined wind and solar have each advanced by about 800 TWh since 2010, not only covering coal’s losses but advancing further to cover total demand, which finally budged from that 4,000 TWh level and has now grown to 4,500 TWh per year.

Because natural gas has lower emissions than coal, there’s a common perception that dumping one for the other is an unqualified win. But that’s not true. Natural gas still produces strong emissions, and natural gas can also hitch its wagon to path dependency, eventually curbing (some of) the expected growth from combined wind and solar. Cold Eye Earth offers the following delineation to untangle the confusion:

• Operationally, combined wind and solar benefit greatly from having enough natural gas on the grid to support their intermittency. The engineering now available to modern, highly-calibrated natural gas power capacity is so advanced that grids can titrate the amount of gas power on the grid from hour to hour—stepping aside when wind and solar are going strong, and stepping back in when they weaken. So far, therefore, this fine-tuned ramping capability has worked perfectly with the rise of wind and solar. Up to some level of penetration therefore, we must allow that this flexible natural capacity has been necessary to help wind and solar spread their wings.

• Alas, once you go beyond that point then new natural gas capacity simply becomes another emissions problem to be solved, neutralizing a great portion of the improvement renewables secured in lowering emissions. Sure, you want to love and celebrate an extra 800 TWh of combined wind and solar, but doing so in the face of a fresh 800 TWh from natural gas is troubling. And unfortunately, this is the path that the U.S. has now taken. And no, wind and solar will not displace young, newly built natural gas capacity.

Facts change over time, morphing into new facts. It’s likely that many still believe the U.S. has an intractable coal problem, when in fact we now have a new problem in natural gas. This is the old “fighting the last war” phenomenon. No, coal hasn’t been zeroed out yet (as it has in the United Kingdom, for example), but directionally we are headed to the same destination. The U.S. has more than enough natural gas capacity in power, and we should be fighting further growth now with the same intensity and planning that was brought to bear on coal starting 20 years ago.

Long-term U.S. tariffs on Chinese solar panels raised prices and triggered a decline in solar industry employment. Those are the findings of a new 2026 study, “Beyond tariffs: A better approach to green industrial policy,” which examined these effects from the starting point of 2012. Unsurprisingly, the study’s broadest conclusion, that “overall human welfare declined,” derives from the simple fact that human health improves when emissions decline and worsens when emissions either rise, or hang around, not falling fast enough.

What’s delicious here is that it’s a perfect demonstration of the insights offered by the British economist David Ricardo more than 200 years ago. Even today, many people have trouble understanding what Ricardo saw: we live in a dynamic economy, not a zero-sum, static one, and tariffs affect far more than just import prices as they eventually shrink overall flows in targeted sectors, reducing earnings, total market size, and purchasing power of workers and consumers.

This is why Cold Eye Earth devoted significant time to Ricardo last year, when Trump returned to office with his caveman economic views. Here we used the example of bike frames and bike components, and the recognition that Japan makes the best components at the most competitive cost, and the U.S. tends to have an edge when making bike frames:

Read more

Global power generation from wind and solar has reached a level that’s equal to total generation from all sources in the United States. And this is 2024 data — 2025 data will arrive in late spring—so global wind and solar have almost certainly risen above this level. Recorded at 4,623 TWh of generation in 2024, global wind and solar likely added another 800 TWh in 2025 and now roughly generate about 5,400 TWh annually. That is very strong growth.

Indeed, the growth of wind and solar globally has been so strong that they equalled the total generation from all sources in the EU as recently as 2021, right around the 2,900 TWh level. Only three years were required to grow global wind and solar generation another 1,700 TWh to the 2024 level of 4,623 TWh. While the effort to kick off actual declines in global generation from fossil fuels remains a difficult problem to solve, there is clearly no problem at all with either the rate or the volume of wind and solar growth globally.

The most important metric for power sector decarbonization is of course total global demand growth for power, and 2025 has turned out to be notably weaker than expected. When we measure how much of this new demand globally is matched by the growth of wind and solar generation, weaker growth in total demand naturally lowers the success threshold for wind and solar, in their ongoing effort to entirely kill the growth of fossil fuel generation. Most estimates of last year’s total demand growth from all sources point to a 3.3% annual increase—far below the 4.3% increase in 2024 and also below the IEA’s previous headline forecast for three straight years of 4.0% (or even higher) growth.

Setting aside the percentages for a moment, it’s not hard or complicated to understand the basic proposition here: If total wind and solar generation added a fresh 800 TWh, we need total global power demand growth to either match that volume or, better still, fall below it. Alas, even with last year’s slower growth in total demand, wind and solar once again did not (quite) cover that gain. This also means that unless hydropower or nuclear power came to the rescue last year (again, we are still working with estimates until official 2025 data is released), then global power sector emissions either increased or stayed flat.

So how has 2025 shaped up so far, as we put it all together? Wind and solar growth are a tad stronger, and total demand is a tad weaker, but combined wind and solar have once again missed out on covering all the growth. Based on current estimates, 2025 likely saw 1,031 TWh of total demand growth, against wind and solar’s 800 TWh. Furthermore, using an updated IEA estimate of 3.7% for 2026 total power demand growth, the projection suggests once again that wind and solar growth don’t quite make it across the line.

Cold Eye Earth would like to remind readers of two concepts that are truly helpful in understanding the interplay between the growth of clean generation, and legacy, incumbent generation:

1. A small thing growing fast is a powerful phenomenon that will shock you with its ability to overtake a large thing.

2. A large thing growing slowly can be devilishly hard to constrain, and will resist its overthrow durably for a long period of time.

We see the same phenomenon across a number of domains, including the intellectual realm, where new and popular ideas can grow and spread quickly, even as older, incumbent ideas that “should have died already!” press onward.

One way these kinds of accretive trajectories can play out is through a sort of collapse, in which incumbency or the status quo is eroded at its base for a long while and finally topples over after somehow managing to stay upright for much longer than most would have expected. It’s harder to see such an outcome in the world of global power, where hardware is dominant and decay is slow. Subsidy schemes, however, can help in producing collapse-like scenarios because they produce life extension in the sectors they seek to protect. So, for example, we have seen collapse unfold in regimes where a currency can no longer be supported, or sclerotic businesses fall apart when a tariff regime, meant to protect them, withers.

We might want to ask ourselves, for example, where might there be a large inventory of coal-based power generation that enjoys if not depends on strong state support? China. Not making predictions, but this is the kind of analysis that is worth pursuing. Wind and solar (and storage) will eventually cut into incumbent generation growth on a global basis, but we will still need to see the kind of collapse (again, as an example) that we saw in coal across the U.S. and EU starting around 2010. We have seen the collapse of China’s real estate sector in recent years—it’s not unthinkable that the same could happen to their bloated, under-utilized coal sector.

As a U.S. based publication Cold Eye Earth feels an obligation to briefly address the extraordinary events of the past week. Rather than passion however, here is a brief historical view that may shed some light: The modern constitutions of most western nations are quite clearly animated by the brutality, the tyranny, and the chronic dysfunction and injustice of the authoritarian political economies which dominated human existence for millennia. Free speech and assembly, trials composed of evidence and juries, and protection against garrisoned troops in cities, home invasion and assaults on the body are repeated themes across western legal frameworks for the very obvious reason that these were all long-endured flashpoints of suffering—seared into the collective memories of humanity.

The murders at the hand of the State which occurred in the Twin Cities of Minnesota this month hit the bullseye of these rights and liberties, and that is why the deaths of two—just two—innocent people are in fact a national crisis, which is now drawing the attention of citizens across the entirety of the political spectrum. The State then followed up these murders with a series of pronouncements and accusations that were so transparently false, that it offended not only everyone already shocked by the lethality of the State, but then brought along a new wave of Americans rightly offended that the State would lie to them so blatantly, so obviously.

Just to add: This all unfolded in the same timeframe which saw a U.S. President literally threaten an ally with an invasion of its territory, thus completely blowing out of the water another important democratic principle—the respect for sovereignty.

Cold Eye Earth is allergic to making specific political predictions, but given that the publication has a large international readership, your faithful correspondent can offer the following view: If you are wondering whether these events are considered to be major here in the U.S., and whether they have penetrated the soul and the psyche of the nation, raising the spectre of an existential threat to democracy itself, and an urgent need to take action, the answer is yes.

—Gregor Macdonald

Editor’s note: today’s issue is shorter than usual because writing and thinking became difficult this weekend, and while I am usually unforgiving about such vagaries of emotion, this time it seemed entirely appropriate to honor the difficulty.

An emerging bright spot in our current energy transition is that global road fuel demand appears to be peaking. Here at Cold Eye Earth we maintain a healthy skepticism towards any claim of a peak in fossil fuel demand or emissions globally, and equally assert that these declarations can neither be reliably “called” in real time. Peak declarations have scurried about for nearly twenty years now, nearly all of them wrong. But at least in the past, the “peak game” in energy demand forecasting paid respect to a time-window, in a nod to uncertainty. Today’s peak callers rush to the stage after quarterly data to yell “peak” only to throw themselves against this wall repeatedly as the years pass by.

The prospect that global road fuel is currently in the midst of peaking is composed of a number of data points, but owes most of its viability to long-run trends that began a decade or more ago. Put another way, the very same trajectories that got people excited about a peak in oil demand were overly embraced when they first appeared. Those trends needed time—and lots of it— to broaden and deepen. Let’s take a look at them.

• The arrival of the first EVs, and the array of incentives to buy them, 2005-2015. We can think of this period as the early investment phase, the start of a long journey that would eventually curb sales of internal combustion engine (ICE) vehicles. Europe and the U.S. were the first participants in this stage which had no effect at all on oil demand but prepared the ground for an eventual shift. A great example of how this lengthy return on investment played out in a specific domain is California, which remains the leading edge market for EV sales in the U.S. The seeds were sown. And eventually, they bloomed.

• Disincentives to ICE vehicle ownership in the form of fees, higher petrol taxation; the rise of the walkable city and the flowering of emissions zoning and congestion pricing; and the exploding popularity of EVs, 2010-2020. Although adoption of EV in the U.S. during this period was sluggish, (and remains so) globally the car itself became an increasingly powerful symbol of the problem of climate change. Europe, for example, had long been a leader in ICE disincentives—mostly through stiff petrol taxes—but then cities from Paris to London to Barcelona began a second initiative to attack ICE vehicles through travel fees, and car zoning. Just as in California, those initiatives accumulated over time and then broke through.

• The juggernaut of China’s EV-forward transportation policy, and the peaking of global ICE vehicle sales, 2015-2025. In the year or two after 2018 we began to safely conclude that global ICE sales had peaked, as they were unlikely to turn around or recover again given global sales growth of EV, especially in China. What we did not know is how rapid and aggressive (see the chart below) the ensuing 5-7 years would become for EV sales in China, as they stormed both passenger and commercial markets. The best estimate for EV sales last year (based on 11 months of data) is that they got close to reaching 50% of China’s total market. This means it’s an easy layup to project that EVs will push past 50% of the market this year. Put another way, the ICE vehicle, in China at least, just became the rotary phone.

Now, to take the next steps to a thesis of peaking global road fuel demand we need to understand that China has largely driven global oil demand higher for years, despite ongoing declines of oil consumption in the OECD. It may seem obvious to say so, but without China’s demand growth global oil demand would have peaked by now—which is to say that when China’s oil demand peaks, it makes it very hard (nearly impossible) for the rest of the world to take oil to a new higher level. Why should we speculate now, however, that this inflection point has been reached? Let’s consider the following:

1. Data from the IEA now shows that road fuel demand in China is flattening out. The IEA tracks both gasoline and gas/diesel oil, and incredibly both have failed to grow for a couple years now, and are projected to (not) grow again this year. Furthermore, it seems quite notable statistically that 2026 gasoline demand is no greater than it was in 2019, and the same is true for gas/diesel oil. The chart below is from the December 2025 Oil Market Report:

   

2. China’s oil demand actually fell in 2024 by 1.7%. As always, one cannot make a trend from a single data point. But we have not seen China’s oil demand fall even once this century, until last year. Not during the global financial crisis. Not during COVID, either. Yes, demand slowed during those periods, but still grew. Another indication: From 2021-2022 China’s oil demand barely grew at all.

Further growth in global road fuel demand has a new opponent and that is affordability. You will have noticed in each of the charts just presented that there’s also a kind of dropoff that’s seen across each data series after the pandemic year, 2020. Here, it’s not 2020 itself that interests us but the way in which gasoline sales in California, oil consumption in Europe, and EV sales in China lurch notably after the pandemic. We need to consider the strong possibility that, through a combination of disrupted purchasing power from inflation, high sticker prices for new cars, and the cost of insurance and maintenance, purchasing a car has become a far more daunting prospect. Now add to this the common knowledge that young people across the world feel that housing and living costs are already burdensome, and that leaves little to nothing left over to handle the purchase of a car. In the U.S., the average price of a car hit another new all time high last year, and this trend looks unstoppable. Indeed, one wonders that the fate of personal vehicle ownership is to become somewhat of a luxury item, supported by upper income consumers. Should this happen, the conclusion is not hard: the inventory of vehicles on the road will shift upward in price, and thus the total stock of vehicles will stagnate, and then slowly decline.

It’s riveting, to say the least, to witness so many constraints raining down on oil demand when oil itself is at twenty-year, inflation adjusted lows. Clearly, the low cost of petrol cannot outrun the high cost of vehicle ownership. Indeed, all fossil fuels are incredibly cheap these days and have been cheap for a while. The price of natural gas or coal almost seem trivial now compared to the incredible increases in construction costs, maintenance, and operational expenses of building new power plants. While applying Baumol’s Cost Disease (or Effect) doesn’t seem to fit easily into this equation, you get the idea: deflation in one part of the economy can free up purchasing power that lands in some other part of the economy, pushing prices upward. Natural gas in the U.S. has to be one of cheapest fossil fuels in the world on a price-to-BTU basis, for example, and it drives the bulk of U.S. power generation. Yet, household electricity bills never stop rising because the infrastructure and labor costs to maintain the grid totally overwhelm the energy-cost to produce that electricity, rendering the price of natural gas a trivial afterthought. To make this point plain: imagine if every energy source was priced at zero—what do you think would happen to all the costs associated with delivery of that energy, to you?

It’s highly likely that the affordability problem for vehicles is of course now expressing itself in the U.S. As you can see, U.S. gasoline demand has been oscillating for two decades, but in conformity with most of the other data series, it couldn’t recover after the pandemic year. And now, in 2025, it’s starting to decline.

Was the U.S. economy weak last year? No, it was not. But gasoline consumption fell from that lower plateau anyway. While it’s certainly not the case that U.S. gasoline consumption is about to enter a steep decline, there’s nothing the U.S. can do to help out now with global road fuel demand. And yes, car purchases among young people age 18-34 is falling as you might expect.

Oil consumption is not likely to add much to global emissions growth in the years ahead. That feels like a momentous statement to make, but unless there’s a massive and sustained technological setback or reshaping of the world order that renews a big call on oil, it’s no longer clear where the demand will come from to get us to successive new highs in consumption. Yes, oil has other applications (as do natural gas liquids) but the downward force right now is clearly brewing in the transportation sector, which accounts for roughly 60% of total global oil demand. Petrochemicals, plastics, and other applications for oil are simply not capable of counteracting the trend in transportation. And the current price of oil, as it’s been sustained at these low levels for some time now, is likely reflecting all these factors.

—Gregor Macdonald

Despite a very encouraging 15-year decarbonization effort, the project of global energy transition continues to face stubborn hurdles. In order to judge the state of progress, it’s important to anchor oneself to a starting line, and Cold Eye Earth continues to prefer the year 2010 as the best overall reference point for our current transition. That’s the year that utility-scale wind and solar really started to land with some force, when EV adoption as a sustainable trend became visible, and when national policy strategies and corporate efforts started to emerge strongly. A different and perhaps even better argument is that 2010 was the year that previously noisy data that reflected the early, nascent stirrings of decarbonization kind of snapped to attention and began to press forward with clarity.

If we look back at that period, we’ll see that most experts were already anticipating that hard-to-abate sectors like steel, cement, air travel, and heat for both industrial and commercial/residential needs were not going to be dislocated easily. That is still the case. Fifteen years have gone by and, at best, we are still at the experimental, proof-of-concept stage in trying to dislodge coal in steelmaking, jet fuel in aviation, and heat in the built environment. These science experiments have been encouraging. Green steel is possible. Hydrogen fuel cells in trucking and even in aviation have potential. But nothing is scaling, and we know why: price. So in these areas, we continue to nibble around the edges of the problem in creative ways, but without any breakthroughs. One example: After all the hopes for hydrogen and its promise of versatility, the accelerated adoption of heat pumps has taken a far bigger stride (notably lowering natural gas consumption in Europe). And what’s poignant in that comparison is that heat pumps are established tech, while hydrogen applications remain largely on the drawing board. Solar panels, wind turbines, batteries, and EVs are in the same category—established, and no longer experimental.

The hurdle to global energy transition that was less anticipated was economic growth itself. A double-edged sword, modern growth provides the momentum and liquidity necessary to get new infrastructure built, while at the same time making the decarbonization hurdle rate far more challenging. Consequently, new energy sources have tried and so far struggled to cover marginal demand, while also displacing existing legacy sources. To be fair, this is not an easy mandate to fulfill. Indeed, we are still waiting for the moment when wind and solar will sustainably cover 100% of global power demand growth. (Reports indicate this may have happened in 2025. We await data confirmation, of course. But a stern lesson of energy history is to always be wary of extrapolating long-term trends from short-term data.)

All this said, we are obligated to regard the rise of wind and solar as one of the main, established supertrends that will—given enough time—eventually cause other sources of legacy power to enter long-term decline. Indeed, if we were not under time pressure to deliver reductions in global CO_2, we would see the current pace of energy transition in far more positive terms. The progression you see below is frankly quite impressive. Since 2010, the world has moved combined wind and solar from 1.76% to 17% of global power generation. That’s an enormous accomplishment in a short time span, and one that has obviously avoided a very large tranche of emissions growth. In the 2025 estimate, growth of power generation ex-wind and solar is held steady at roughly 26,630 TWh, based on projections by Ember. Whether the rather slow rate of total demand growth, estimated at just 2.7%, endures the finalization of this year’s statistics remains to be seen.

Cold Eye Earth has written extensively about the fleet turnover challenge in our efforts to transform global power. There is no natural law that prevents combined wind and solar from cutting into all legacy sources, forcing them into decline. That wind and solar have not been able to do so can be found in a much simpler, more conventional explanation: New wind and solar are cheap enough to compete for new generation and to dislodge old, very uneconomic generation. That’s the story, right there in the chart above. What wind and solar can’t do is dislodge economically viable generation. That this line of demarcation is not intellectually challenging to understand suggests that those who keep resisting it do so because they understandably don’t like the message.

But turn the orb slightly and the prismatic light that starts to shine through is still pretty positive: Legacy generation gets older every day. Rust, as they say, never sleeps. And despite the fact that economic growth makes this all the more challenging, let’s marvel at how adept wind and solar have been in pushing through these myriad challenges regardless. Economic growth is like the proverbial steamroller, in front of which one tries to pick up nickels and dimes before the barrel squashes all that lies before it. Wind and solar have neatly survived that experience for more than a decade, and they will indeed prevail. Eventually.

The other global supertrend with real promise is the adoption of EVs, though this trend might escape your notice if you live in the United States. (Sad face.) Global sales of internal combustion engine (ICE) vehicles peaked in 2017, but it’s only now that we are seeing its first effect: Global consumption of gasoline is starting to flatline. This is due to China’s almost singular effort to remove ICE vehicles from the market. The policy mandate has had a profound effect not just within China but now outside of China too. If you blinked, you may have missed that EVs are now nearly 50% of China’s vehicle market—passenger cars and commercial vehicles combined.

Why was China able to use EVs to drive out the incumbent technology, while global wind and solar have still not crossed that threshold in the global power domain? Multiple reasons. For many Chinese consumers, the EV was their first car. The country was only at the beginning of the S-curve of personal vehicle ownership when EV technology began to roll out. So, unlike mature vehicle markets in the OECD, there were far fewer vehicles on the road with residual utility value. Readers may recall the following example from the March 31, 2025, issue:

A family keeps a 2010 Toyota in the garage so their two college students can drive it when they are home for summer, working jobs. No one in the family loves the Toyota, and they are all math literate and understand that running costs for a new EV would be far lower. Ah, but the EV represents a new capital cost to the family, whereas the 2010 Toyota is already paid for. So the value proposition of the 2010 Toyota is poor if you were buying it today off a used car lot. But not if you already own it. The value proposition of keeping the Toyota, the value to be captured in its ongoing utility, therefore, is excellent.

The OECD has a far greater fleet turnover problem when it comes to EV adoption, not only because these regions began adopting cars widely more than 100 years ago, but because cars now last longer too, with lifespans lengthening notably in the past 20 years. The OECD was also early to electrification, so power plant penetration also began here over 100 years ago.

Unsurprisingly, road fuel demand is no longer growing (much) in China, and as a result, it’s no longer growing (much) worldwide. According to the IEA, for example, global gasoline usage barely grew in 2025, and it’s expected to not grow at all in 2026—even though total oil demand is forecasted to rise by 0.8%. The global oil market is on the verge therefore of losing road fuel as a pillar of growth.

The profound effect that the Chinese EV initiative is having on transportation fuel demand is now spreading outward, reaching markets across the non-OECD where legacy automakers like Ford and Volkswagen are now seeing much stiffer competition for sales. Here’s a representative data point that underscores the tectonic shift underway: Through the first 10 months of this year, China’s exports of EVs rose 90% compared to 2024, reaching more than 2 million units sold. How dominant has China become in world car markets? Well, Chinese EVs now have a whopping 75% share of EV sales in emerging markets (non-OECD nations). Ponder that data point for a moment. In a global vehicle market that runs at around 85 million units per year, 20 million of those units are now EVs. And China is providing 10% of that market now just through exports!

This quiet war on oil that China kicked off last decade is entirely rational. Leadership correctly judged that were a nation of 1.4 billion people to fully adopt ICE vehicles, the air pollution alone would have become politically unmanageable. Last decade, the populace indicated to leadership that coal-driven air pollution was becoming a crisis, and leadership listened. Already the world’s largest importer of oil, the prospect of an emerging ICE fleet in non stop expansion would have only added another fearsome layer to the country’s carbon output from not just coal, but industrial uses of oil.

There’s a familiar lesson in all this to Western nations, who habitually want to avoid any policy-pain when it comes to cars, car ownership, and voters—who invariably are also car owners. If you want to transform large systems rapidly, you will need to employ both the carrot, and the stick.

There’s also a surprise. In most forecasting over the past decade analysts have generally regarded oil as a far more challenging fuel to overthrow than, say, coal. But it’s coal instead that has shown nine-lives-capability—faking us out with a ghostly peak in 2014 before going on to reach new, successive all-time highs—while it’s oil that looks like its losing its grip on one growth driver after another.

Tipping points are achievable when scalable technologies are given enough time to gain momentum, triggering network effects as they distribute their magic. We are now in a position to categorize the scalable solutions that are already in place, or have potential to start scaling soon.

• Technologies that have proven their ability to scale, and which are already heavy movers in their respective sectors: wind and solar in global power, and electric vehicles in global transport.

• Technologies that have proven their ability to scale, but which are not yet heavy movers in their respective sectors: nuclear in global power; heat pumps in buildings; grid-scale batteries; high-voltage direct current (HVDC) technology in power transmission.

• Technologies that have not yet proven their ability to scale, but which have the potential to become heavy movers in their respective sectors: geothermal in global power, and robotics in global manufacturing, transport, mining, transportation (autonomous), and agriculture.

• Technologies that either remain on the drawing board or have working prototypes whose future ability to contribute meaningfully to energy transition is uncertain: the entire promise of hydrogen, from heat generation in industrial settings to battery cells in air and ground transportation; materials recycling in everything from batteries to spent nuclear fuel; myriad conservation and efficiency efforts, from building materials to networked electrical flows; and carbon capture technologies.

The last two categories receive considerable attention in the press, because humans are wired to care alot about breakthroughs. But for our purposes, we should concentrate on the second category first, while keeping an open mind to all the rest. It’s that second category of nuclear, heat pumps, grid-scale batteries and better transmission technology that’s in a position right now to join wind, solar, and EV as new forces for energy transition.

While it’s true that global emissions are likely to hit a new all time high in 2025, at least the estimated 1.1% increase is within the normal range for annual additions. As readers know, Cold Eye Earth has been critical (and correct) of nearly all peak emissions analysis up to this point in time. And yet, now that we have arrived at mid-decade, the next five years look brighter. Wind and solar will likely do a “pretty good job” of suppressing emissions growth in global power. Even more enticing is the prospect that global EV adoption has already reached the suppression point for road fuel. It’s not impossible to imagine, for the first time, that oil consumption will no longer add to annual emissions increases. That’s encouraging.

If you are feeling slightly more optimistic about the next five years, you need to see the following if you want to retain such optimism: wind and solar and EVs press onward in their disruption of their respective sectors, and then nuclear, heat pumps, and grid-scale batteries will need to demonstrate they are not just ready to join the fight, but are ready to move aggressively enough to change the nature of the fight itself.

—Gregor Macdonald

To better manage the holiday period, Cold Eye Earth is shifting its publication schedule ahead by one week and will release its next issue on Monday, 29 December 2025. We wish everyone a warm and fulfilling holiday season. While you’re here, a preview of the next issue: EV sales are about to reach 50% share in China, the world’s top vehicle market. That’s a very big deal, with broad ramifications for emissions and the overall project of energy transition. See you soon.

—Gregor Macdonald

In the current energy transition, the most important metric to follow is how much of yearly global power demand growth can be covered by low-carbon resources. Cold Eye Earth, along with its preceding publications (gregor.us and terrajoule.us), has tracked this metric for over a decade, hoping that this threshold would soon be crossed. But with the exception of a couple of years when total wind and solar grew moderately toward that target, we have continued to wait patiently for a better outcome. (Total global power demand actually fell in the pandemic year, 2020, but the that year remains an unhelpful outlier in all datasets.)

Through the first three quarters of this year, however, reporting indicates that total global demand growth for power is coming in much lower than forecast, which considerably lowers the bar for combined wind and solar to meet all new demand. This is surprising. Not on the wind and solar side of growth, but rather that total growth, which is supposed to be running at 4.00% or so, is coming in lower, reportedly at just 2.7%. Given all the drama and excitement over accelerated AI-driven demand for power, one has to wonder why the high rate forecast, coming from the IEA, for example, is not coming true.

London-based think tank Ember, which tracks this progress, has just reported its latest year-to-date estimate:

In the first three quarters of 2025, solar generation rose by 498 TWh (+31%) and already surpassed the total solar output in all of 2024. Wind generation grew by 137 TWh (+7.6%). Together, they added 635 TWh, outpacing the rise in global electricity demand of 603 TWh (+2.7%).

As previously discussed in the November 10 issue, wind and solar generation growth this year continue to track with the simple (if not simplistic) growth model used by Cold Eye Earth, which projects the two technologies will advance by 775 total TWh this year. But if they have already grown by 635 TWh through Q3, then they may complete the year at a level higher than that projection, above 800 TWh. That’s great news for wind and solar, but also for the Cold Eye Earth model, which may have correctly estimated this year’s growth within 5%.

The big miss here, again, is total demand growth, which is also on pace to come in around 800 TWh—leaving a big gap against the ongoing IEA forecast. In the chart below, Cold Eye Earth previously modeled its usual year-to-year wind and solar growth forecast, and also a three-year run of total demand growth of 4.0% per year. (Note that there are small discrepancies in the comparison of 2024 to 2025’s total demand growth because 2024 was a leap year.)

If total global power demand comes in at just 64% of the forecast this year (800 TWh vs. 1250 TWh), not only will it be a nice illustration of how dominant this factor is to the direction of energy transition, but it will also suggest that much of the expectation of higher growth from AI-related workloads has not yet landed. Other factors that might explain the miss: lower global economic growth as a result of disruptive U.S. trade policies, and there may be weather-related effects too. As we know, extremely hot years can cause global air-conditioning to spike. And any easing of that demand, from an anomalous cool summer, would have the opposite effect. It’s also good to remember that real-time data collection has higher error-risk. After all, even data that’s gathered retrospectively often has to be revised also. (Agencies like the U.S. EIA revise data continually.) So, Ember’s estimate of total demand will likely be revised. But, by how much is uncertain. Finally, there is the forward march of energy efficiency, but that is a small and steady factor and certainly could not account for the huge difference between 2.7% and 4.0% growth of global power.

Progress in AI is pressing onward as a new round of models are released. Google’s Gemini 3 was let loose about three weeks ago and has wowed the tech community with its stellar capabilities. Claude Opus 4.5, from Anthropic, emerged a few days later. And rumor has it that OpenAI—which reportedly sounded the alarm over this mounting competition with a company-wide memo, calling a code red from Sam Altman—has decided to hurry up and release its latest version of ChatGPT this coming week.

Your faithful correspondent thought it might be useful, therefore, to report how these models actually perform on a real-world task. The usual province of Cold Eye Earth is to report, of course, on all the energy demands that might emerge from AI. But there has been so much public focus on that theme that it’s left a deficit of reporting on how AI might work for the average person. In this case, we’ll take a look at how these models handle taxes and financial planning.

Background: I have a fairly complex tax picture that begins with having been a longtime independent worker whose earned income is exclusively reported on Schedule C, for small businesses. This means I have a number of deductions that come through that channel, from annual healthcare premiums to the qualified business income deduction. I also have a Roth IRA, a regular IRA, inherited IRAs, and a personal (Solo) 401(k). Looking further ahead, over the next ten years there will come the question of Social Security (and when to start taking it). And finally, all these factors weave into Oregon State tax laws, as I live in Portland. For example, Oregon’s tax return starts with the federal AGI (adjusted gross income) and Social Security is not taxed in Oregon. However, Oregon has a very progressive income tax table (partly owing to the fact that we have no sales tax on anything—not even meals or alcohol) and therefore smoothing income, when possible, to avoid annual spikes is of some benefit in this state. One area of tax simplification that has occurred since 2020 is the introduction of a new, super-high standard deduction at the federal level, which has wiped out itemized deductions in the housing area for many—though certainly not all.

Traditionally, I’ve handed all this work over to “a tax guy.” Here in the U.S., having a tax guy is your best choice to navigate the ridiculously complicated U.S. tax code. I inherited one of these wizards from my parents, back in Massachusetts. Because I’m a “Schedule C guy,” I found that my tax guy routinely pushed back on some of my deduction efforts, while uncovering other deductions hiding under rocks. That’s exactly what you want: all valid deductions taken, while steering clear of any claims that would violate the rules.

Let’s see how the various AI models handled all these factors. I asked them to estimate my 2025 tax burden in a single scenario, my 2026 tax burden in several scenarios, and then I asked for 10-year projections that started to incorporate either IRA withdrawal or Social Security income or both. To be honest, I had no idea these models could come even close to handling all this information, so here’s an early preview of my conclusions: As the models produce answers and even downloadable spreadsheets, you can almost see in your mind’s eye how so many accounting and financial planning positions will eventually disappear from the landscape.

Claude: Dario Amodei, the CEO of Anthropic, has been clear that the angle his company is pursuing in AI is specialization; and they are largely focused on business customers. They want to develop various speciality channels that would focus on pharmaceuticals, chemicals, biology, math, and science more generally. Claude at the moment has decisively won the computer coding channel, for example. When you hear tech CEOs say that most of their coding workforce now uses AI to write the code, thus leaving editorial work to them, this is why. Unsurprisingly, Claude is also extremely good with numbers and data.

I paid for a subscription to Claude some time ago for exactly this reason. I wanted to analyze some exceedingly large data spreadsheets from the EIA. One of my first queries was asking Claude to estimate the average age of natural gas power plants in the U.S. Longtime Excel wizards might wonder why this query couldn’t simply be accomplished using Excel’s native capabilities. Probably could! But Claude performed double sorting in this task: It had to pick out all the natural gas from every other listed plant type (coal, nuclear, solar, etc.) and then moved on to its second sort. This all happened in a flash, by the way. And yes, it felt like a shazam kind of moment.

So how did Claude do on taxes? Its user interface (UI) is particularly good for these tasks. It unfailingly produces an estimate that is formatted like a professional report, and it offers a spreadsheet version to boot. Claude looks gorgeous. But it got very hung up on the qualified business income (QBI) deduction. To be fair, the IRS’s instructions for Form 8995 are slightly loose, and less clear. But that’s not the only area where Claude had trouble. Its main weakness was being unclear where, exactly, in the order of operations to apply the QBI deduction. Indeed, it was not able to correct this mistake until some time later, when I pitted a different model against it, forcing Claude to admit it had never quite gotten the QBI correct.

Claude also had trouble analyzing the optimal point in the future to start taking Social Security, and interestingly the model had such high confidence in its ability to perform such longer-term projections that it offered these queries to me—only to prove less able to solve them. Indeed, Claude was super eager to spit out all sorts of ancillary analysis: tax efficiency of the Solo 401(k), tax drags from taking Social Security too early, and how best to handle required minimum distributions from an inherited IRA and future withdrawals from regular IRAs. It was as though Claude was super-informed about all these measures and scenarios from a top down view, but then got hung up on the details, and had to correct itself through further prompting. Still, I was very impressed. The AI model eventually delivered to me eventually high-quality estimates. This echoes other reporting on the current ability level of these interactive AI models. They work best when the user has enough knowledge to challenge some of their answers.

ChatGPT: Lightning fast, and exceedingly good, I was surprised that ChatGPT needed a corrective prompt. The glaring error it made was easily avoidable: it completely missed the basic fact that Oregon’s return starts with the federal AGI, and as a result it got all its Oregon estimates wrong. But the fact that ChatGPT got everything else right paints a portrait of a genius who can spit out differential calculus answers while driving a car, but then shockingly blows through a stop sign. “How lazy do you have to be,” I asked, “to get the most basic fact wrong about Oregon’s return?” Chat GPT apologized (all the models apologize a lot) and quickly fixed its error.

When I reported Claude’s answer to the QBI question, however, ChatGPT suddenly went on the offense. “Claude AI made a big error by incorrectly figuring the basis amount, on which to apply QBI!” I acknowledged that ChatGPT was quite right, and it seemed quite pleased with itself and proceeded to produce a bibliography of tax essays and other material supporting its (correct) take. Bottom line: ChatGPT correctly understood that higher contributions to my Solo 401(k) would actually reduce the basis on which to take the 20% QBI deduction, thus making the overall QBI deduction smaller. Again, note the strange asymmetry: ChatGPT made the most elementary of errors on the Oregon return but completely aced a far murkier area of tax law in QBI.

Gemini 3. Gemini 3 got everything right the first time through. That I did not expect. It felt like magic. Actually, it made a small error in calculating my standard deduction, but claimed that the way I had formed and formatted that question had led to its error. Really? Oops. Gemini was right. Which shows that in AI the quality of the query heavily controls the quality of the outcome.

Conclusions: AI is very much all around us and taking place quite actively in the present, even if the narrative focuses on a more distant big-bang future where AI supposedly becomes super-intelligent. The demand is already here, and the productivity effects on labor (as well as the dislocating effects on labor) are already taking place—but in myriad small ways, rather than the revolutionary outcome many are still waiting for. While it’s true that the return-on-investment question looms large for those megacap corporations loading up on chips and data centers, it is no longer possible to imagine how consumers, businesses, and institutions won’t soon become hopelessly addicted to these capabilities. Given that AI is ultimately energy-driven computing power, upward pressure on future power demand is not in question.

U.S. power demand growth in 2025 is also coming in below expectations, and this may go some way to explain why global power growth is falling far short of the EIA’s 4.0% forecast. To be sure, U.S. power demand growth is currently stronger when compared to the past two decades. But the hue and cry over runaway power demand here in the U.S. from rapidly rising AI power demand is simply not coming true. Not yet, at least.

In the department of good news, however, the projected increase from 2024’s total demand of 4393 TWh to this year’s estimate of 4500 TWh has once again been handily covered by growth of combined wind and solar. And given that wind growth this year is awful, barely above 1.0%, indicates that the heavy lifting is being done by solar. The all important ratio is looking pretty good, therefore, here in the U.S. This leaves only the ongoing question of when, if anytime soon, fossil fuel power generation will actually fall.

—Gregor Macdonald

In its latest World Energy Outlook the IEA reopened the door to continued growth of global oil demand. This is a change from their outlook of the past few years, to be sure, where they, like others, forecast little to no demand growth after 2025. In the opinion of Cold Eye Earth, however, we shouldn’t be too alarmed by the IEA’s shift in posture. After all, it’s just one of their scenarios (current policy scenario, or CPS), and in this latest CPS, oil demand continues rising—by an aggregate 13% in the 26 years from 2024 to 2050. Looked at more closely, that’s a scant 0.5% of growth per year until midcentury. And it’s probably not a coincidence that this tracks with the sub-1% growth that oil is experiencing lately.

Readers will recall it was British Petroleum that first made the big call on “scant to nil” post-2025 demand growth, way back in 2019. And since that time, a number of analysts have gathered around this year as a pivot point. Why is the IEA suddenly so alarmed? It’s not clear. But here is one possibility: The effort to stop oil demand growth requires intense effort, and in recent years China has led that effort valiantly, followed by a solid and ongoing effort in Europe. But North American EV adoption just isn’t that great, and the current administration is of course doing everything possible to return oil to a stronger position not just here in the U.S. but in this entire hemisphere . If we examine a related IEA chart, "Global new car sales by drivetrain,” one can see how although ICE vehicles globally peaked years ago, around 2017, the rate at which we stop buying ICE altogether and increase purchases of EV still matters.

Here, meanwhile, is the fossil fuel chart from the IEA, “Oil and gas demand do not peak,”revealing their shift in the CPS.

You will also note that the IEA sees no peak for natural gas, either. That particular aspect of their forecast looks more solid. Gas has two things going for it. (1) As a partner to variable renewable sources, gas greatly helps the balancing-needs of global power grids—and that will be true even as battery energy storage systems (BESS) continue to grow. (2) In its competitor role, gas has sustained a strong, ongoing capability against coal, and much of the global natural gas fleet is young.

Indeed, the favorable growth prospects for natural gas offer additional help in understanding why growth prospects for oil are so poor: In global transportation, there remains oil dependency but not oil growth. And outside of transportation, whether it be chemicals or an array of industrial applications, there’s simply not enough organic growth to offset oil’s poor prospects.

While we want to be cautious about using the signal from oil prices to strengthen the argument against (much, if any) oil growth in the future, it has become difficult to ignore that the global oil complex has been building inventories, growing spare capacity, and suffering the low prices that inevitably follow those emerging conditions. For at least a year or two now, it’s also been possible to declare that oil prices, on an inflation-adjusted, basis are dirt cheap. Oil was at one time considered an inflation hedge, and for short periods oil may still retain some dwindling qualities under those circumstances. But it’s pretty shocking to see oil at $60 a barrel in the year 2025. That’s after 20% aggregate inflation since 2019-2020 or so, and doesn’t even include the inflation from the previous 20 years. Oil prices first reached $60 a barrel in 2005!—which shocked people and rightly so. But $60 oil in 2025 is little more than a pale, thin shadow of those go-go years.

To be blunt: Today’s oil prices do in fact indicate that demand absolutely sucks, and the futures market (which is not infallible) doesn’t see that changing. It’s also fair to say that the IEA often makes new shifts and declarations right around the time the trends they are seeing reach their apex. To be sure, the Crazy Ivan move in the U.S. toward fossil fuels has been jarring, and the importance of the U.S. in these equations no doubt adds to the overall dark vibes about energy transition in general.

Presently, according to the IEA, global oil demand is on course to advance 0.7% in 2025 and by a similar percentage next year. But overall, oil demand has grown only 3.2% in the aggregate in the five-year period from 2019 through 2024. From a climate perspective, the central problem that oil presents, therefore, is not that it’s going to keep advancing, but rather that it’s just not on course to decline.

The IEA has of course updated their other main scenario, the Stated Policies Scenario, or STEPS, and that has led them to muse and speculate over various peaks. Cold Eye Earth likes this scenario much better, and overall it seems more in tune with current trends. (Indeed, the CPS increasingly looks like a rearguard move by the IEA that gives them credibility in case “all this energy transition stuff” doesn’t work out very well.) For example, in the STEPS, oil demand stops growing around 2030; natural gas reaches its maximum level ten years from now, around 2035; nuclear generation finally grows again on a sustainable basis; and coal keeps going for a while but really starts to fall off hard by 2030. You can see all those projections in the chart below, also from the IEA’s 2025 Outlook.

It’s no surprise that nuclear fares so well in the STEPS. Nuclear is the perfect example of a technology that truly lives and dies on the back of policy changes. And yes, that remains true even if you posit that the new wave of SMRs comes to fruition. Everyone is aware that many countries around the world are coalescing to promote a nuclear revival, but to achieve the IEA outcome shown here will require countries to actually follow through on their plans.

Coal, meanwhile, and the decline the IEA forecasts, is therefore the shadow of these nuclear plans. The titanic ascent of wind, solar, and storage—along with natural gas—has certainly helped to blunt coal growth. But to trigger actual declines in coal will require all the current effort plus something extra—like nuclear—to force coal consumption to a final tipping point.

A final thought: Cold Eye Earth remains allergic to the common assumption that peaks convert quickly to declines. They don’t. However, we should keep an open mind for the possibility that some fossil fuels, like coal, could actually not just decline but fall off a cliff as the encirclement of gas, renewables, and nuclear reaches critical mass.

Wealth and income inequality are back on the front page as public figures race to converge on the issue of affordability. The tip of the spear in current discourse appears to be centering around the election of Zohran Mamdani for New York City mayor after he ran a campaign absolutely laser-focused on affordability, while at the same time President Trump has run into deep trouble in voter popularity as the public has decided his administration not only messed up but is now gaslighting everyone about its record on inflation. A fresh CBS poll released on November 23 showed that the public has now essentially transferred its ire—once trained on Biden and the last administration—to Trump. And so Trump is now the punching bag for the fact that prices are either still going up or are already unaffordable.

Cold Eye Earth, however, likes to check in with the top-down view, and that means taking a look at the latest Gini coefficient reading for the U.S. Is the U.S. more unequal than most of its OECD counterparts? Indeed it is. But this comes as no surprise. The U.S. has only partly been running a political economy that adheres to the “free-market welfare state” that’s more common in Europe. Yes, the U.S. has had a safety net since FDR established Social Security. But the U.S. diverges in two important ways from the OECD: Income tax burdens on higher-income earners are much lower in the U.S., and even now in the 21st century the U.S. still doesn’t offer universal healthcare. Those two factors alone—if you run them long enough—could deepen wealth and income inequality. Whereas the U.S. has a Gini coefficient of 59 (the higher the number, the worse inequality), the E.U. comes in at 48. The full data page, with the interactive version of the chart below, can be found at Our World in Data.

The outlook for the U.S. is not particularly bright, however. While the American public doesn’t like inequality much and is increasingly bitter about affordability, when given the opportunity to gravitate toward the EU model, most U.S. voters decline to vote for such changes. Meanwhile, in the background of such ongoing policy preferences, the U.S. and the rest of the OECD continues to roll through a series of technological changes that for decades have increasingly replaced workers with automation. If you believe that AI poses the risk that this automation goes into hyperdrive, then the outlook for inequality is due to worsen rapidly.

Using AI has likely become routine for most people, and this fact seems underappreciated. This “light AI” which comes in the form of advanced search, queries about daily life from finances to taxes to home repair (no doubt already helping personal productivity), is surely playing a role in the quickly rising demand that makes purveyors of AI services so excited. Cold Eye Earth uses AI semi-regularly to analyze datasets, for example. Each model has its own strengths.

Here’s a fun one: my recent travels I was gifted this nice chunk of rock with a couple of fossils, and so I took the photo above, and asked Google’s Gemini to make its estimates. Here is what Gemini returned:

That is a beautiful piece! Based on the appearance, preservation, and the rock matrix, these fossils are almost certainly fish from the Green River Formation in Wyoming, USA.

This geologic formation is famous worldwide for producing incredibly detailed, Eocene-age (about 50 million years old) fish fossils like the ones you have.

🐟 Likely Identification

The two most common genera of fish found in this formation, and the ones that match the general body shapes in your image, are:

1. Lower Fish: This appears to be a good example of Knightia eocaena.

   • Features: It has a smaller, more streamlined, herring-like body, which are the most abundant fossils found in the formation. They rarely grow over 10 inches, and this one looks to be in that typical size range.

2. Upper Fish: This one might be another Knightia, or possibly a juvenile Diplomystus dentatus.

   • Features of Diplomystus: Diplomystus is the second most common fish and is known for having a deeper, “hatchet-shaped” body and an upturned snout (though the snout isn’t perfectly clear in the photo). They were predators that often fed on the smaller Knightia.

• News briefs • Concentration risk is growing elsewhere too, and not just in measures of inequality. Investors are quite aware that nearly all the performance in the U.S. stock market increasingly comes from a handful of mega-cap corporations, leaving small companies to tread water. • Mike Green of Simplify Asset Management thinks that the way the U.S. measures poverty has a deep flaw, and as a result we are likely undercounting poverty and underestimating the actual household income needed to support a family, in the year 2025. • New England, which has been reluctant to build gas pipelines, transmission pipelines, or enough renewables, has increasingly found itself struggling with very high winter energy prices. Now comes a plan that takes a page out of the EU’s response to the Russian gas crunch: build heat pumps, and lots of them. • Amazon will work with Energy Northwest and X-energy to build an SMR in eastern Washington, on the old Hanford site, home of the existing Columbia Generating Station. • Synchronous condensors, a technology originally developed nearly a century ago, are expanding quickly again as they solve stability challenges on grids that are increasingly composed of variable renewable energy (VRE). •

—Gregor Macdonald