Have you ever watched a market flip so fast that last year’s “clean and flexible” fuel becomes this year’s expensive headache? That is roughly where European power sits right now. For years the working assumption was simple: burn less coal, lean on gas, fill the gaps with wind and sun. Then shipping lanes tightened, prices jumped, and the old spreadsheet stopped working. I keep coming back to one blunt question. If a single kilowatt hour is the product, which fuel still makes financial sense after carbon costs?
When Gas Stopped Being The Cheap Backup
The short version is uncomfortable. In several large European power markets, coal-fired generation has become more profitable than gas-fired generation. That is not a slogan from a coal lobby. It is what operators see when they stack fuel costs, efficiency, and carbon allowances against the price they can actually sell electricity for. After the latest round of Middle East tension and disruption around liquefied natural gas routes, the two profit paths split. Coal margins widened. Gas margins sank, and in some hours they went negative.
I find that last point easy to miss if you only follow wholesale gas headlines. A plant can run and still lose money on a spark-spread basis. When the fuel is dear and the carbon bill is not trivial, the operator has a choice: generate at a loss, or sit idle and buy power from someone else. That is how you get a winter conversation that feels like 2022 all over again, even if the political slogans have changed.
Benchmark European gas briefly pushed through 80 euros per megawatt hour this month and tagged three-year highs. That single number does a lot of work. It changes dispatch order. It changes hedging. It changes how industrial users think about night shifts and furnace schedules. And it changes the political temperature around plants that were supposed to be on the way out.
When the fuel that was meant to be the bridge becomes the bottleneck, the system starts looking for any dispatchable megawatt that still exists.
Why Coal Margins Widened While Gas Sank
Think of a power plant as a thin-margin factory. The “product” is electricity. The “inputs” are fuel, carbon permits, operations, and a bit of luck with the weather. Coal plants are dirtier, so they pay more for emissions. For a long stretch that extra cost was enough to keep them behind gas in the merit order. Then gas prices ran away from coal prices. The carbon penalty did not disappear. It just stopped being large enough to cancel the fuel gap.
That is the quiet math behind the chart everyone in trading rooms has been staring at. After regional conflict flared and LNG shipping through a critical chokepoint got messier, the lines diverged. Summer demand and storage worries stretched the gap further. Coal generation margins expanded. Gas-fired power slid deeper into the red. In my experience, markets do not stay polite when that happens. They use whatever is still connected to the grid.
Analysts now talk about European coal-fired output rising by roughly a quarter over the next six months, with gas-fired output falling in response. That is a big swing in a system that spent a decade telling itself coal was finished. It is also a swing with a ceiling, because you cannot revive plants that were demolished, converted, or legally boxed in.
The Structural Ceiling On Bringing Coal Back
Here is the part that makes the story more than a commodity spike. Europe did not just use less coal. It retired the machines. Decades of transition policy pushed plants offline. Official statistics tell a stark story. Coal once supplied more than a third of European Union electricity around 1990. By 2025 the share had fallen to a historic low near 9.2%. That is not a rounding error. That is a fleet that no longer exists in the same form.
Germany matters disproportionately. It is the largest power market on the continent and a heavy gas consumer. Forecasts for the fourth quarter already point toward coal generation bumping against the operational limit of what remains. You can run the surviving units harder. You cannot invent spare boilers overnight. Permits, staffing, fuel logistics, and public opposition all act like brakes.
So the system faces a three-way squeeze. Gas is expensive. Coal capacity is thin. Wind and solar remain intermittent, which is fine on a bright, windy afternoon and less fine on a still winter evening when factories and data halls both want firm power. I’ve found that people argue about ideology when they should argue about dispatchability. The electrons have to show up when the price signal screams for them.
- High gas prices push operators toward any cheaper thermal unit still licensed to run.
- Retired coal capacity cannot be summoned with a press release.
- Intermittent renewables do not automatically replace baseload on dark, calm days.
- Carbon costs still matter, but they no longer dominate the short-term merit order.
- Industrial demand and digital demand both want stable megawatts, not hopes.
Italy’s Nuclear Vote And France’s SMR Bet
If coal cannot scale and gas is punishing the bill, the conversation slides toward nuclear whether campaign posters like it or not. On Wednesday the Italian Senate approved a government plan to restart nuclear power. That matters because Italy lived for nearly four decades without active nuclear plants. Clearing legal hurdles is not the same as pouring concrete. It is, however, the first door that has to open.
France is taking a different route with more industrial muscle already in place. The state utility has announced plans to build 10 small modular reactors across the European Union by 2035. SMRs are sold as faster, smaller, and easier to site than giant legacy units. Maybe. Construction timelines in nuclear have a habit of stretching. Anyone promising cheap firm power next winter from a reactor that does not exist yet is selling a calendar, not a kilowatt hour.
That lag is the policy trap. Building nuclear takes years. Energy prices move in weeks. Inflation and near-term security stay exposed while the new steel is still on order. I do not say that to dismiss nuclear. I say it because the public debate often jumps from “we should have plants” to “problem solved.” Those are not the same sentence.
I think we are at the upper bound of the neutral territory. I will not exclude that we have to go in the mild restrictive territory, but as I said, it is very much dependent on how the energy prices evolve, how the price picture is evolving over the course of maybe the next month.
– Senior European central banker
That comment is doing more than rate-path theater. It is an admission that power prices still sit inside the inflation machine. If gas stays elevated, the “neutral” zone gets harder to defend. If prices ease, the committee can wait. Energy is not a side chart anymore. It is part of the monetary reaction function, whether anyone enjoys that fact or not.
Power Grids, Artificial Intelligence, And National Competitiveness
There is a newer twist that older energy crises did not have in the same way. Rapid expansion of artificial intelligence is turning electricity from a background utility into a strategic input. Training clusters and inference farms do not care about political branding. They care about uptime, interconnection queues, and a price that does not jump 40 percent because a tanker route got messy.
Grid stability now touches household bills, factory margins, and the ability to host compute. Countries that can offer firm, relatively clean, relatively predictable power will look more investable. Countries that bounce between scarce gas, constrained coal, and weather-dependent supply will look like a risk premium. That is not ideology. That is site-selection logic.
Perhaps the most interesting aspect is how quickly “energy transition” language collides with “energy security” language when prices spike. The same government that celebrated coal closures now needs those last units to hold the evening peak. The same public that wanted cheaper green power still wants the lights on during a cold snap. Adults can hold both thoughts. Markets already do.
| Source | Near-term flexibility | Fuel price risk | Build time |
| Existing coal | High if licensed | Moderate | Already built |
| Gas plants | High | Very high right now | Already built |
| Wind and solar | Weather bound | Low fuel cost | Faster than nuclear |
| Large nuclear | Firm once online | Low operating fuel risk | Many years |
| Small modular reactors | Promised firm power | Low operating fuel risk | Still unproven at scale |
What A Six-Month Coal Rebound Would Actually Look Like
A 25 percent lift in coal generation sounds dramatic until you remember the base is smaller than it used to be. Raising a diminished fleet by a quarter is not the same as returning to 1990. It is more like wringing extra hours out of aging assets while hoping maintenance outages do not cluster. Fuel logistics matter too. Coal still has to arrive, get stored, and meet environmental operating limits.
Gas plants will not vanish. They remain the fast responders when wind drops. The issue is economics, not physics. If running them destroys margin, owners minimize hours. That can tighten the evening stack and lift wholesale prices even if nameplate capacity looks fine on a spreadsheet. Capacity that is too expensive to switch on is not really capacity. It is scenery.
Households feel this with a lag. Industrial users feel it faster, especially those on contracts linked to wholesale markers. A chemical plant or a glass furnace cannot “work from home” during a price spike. Some will curtail. Some will pass costs through. Some will quietly study sites outside the region. That last group is the competitiveness story hiding under the energy story.
Carbon Pricing Did Not Vanish. It Just Got Outrun
It is tempting to say climate policy failed because coal is making money again. That is sloppy. Carbon costs are still in the stack. They still favor cleaner kilowatt hours when fuel prices are close. The current episode is what happens when one fuel’s price explodes for geopolitical reasons. The allowance system was not designed to neutralize a shipping shock through a strait.
In my view, pretending the carbon market can also be a full energy-security policy is asking one tool to do three jobs. It can tilt investment over years. It cannot conjure spare turbines next Tuesday. If governments want both lower emissions and firm winter power, they need firm low-carbon supply, storage, interconnectors, and demand response that actually shows up. Slogans do not balance the grid at 7 p.m.
There is also a fairness angle that does not get enough airtime. When wholesale prices jump, the pain is not evenly spread. Households with efficient homes and rooftop generation are insulated. Renters in older buildings are not. Energy-intensive employers in contested electoral regions notice first. That is how a commodity chart becomes a political chart.
The Uncomfortable Middle Path For Policymakers
Nobody running a finance ministry wants to stand at a podium and praise coal. Nobody running a grid wants to explain blackouts either. So you get a messy middle: keep the last coal units available, beg for LNG cargoes, accelerate nuclear paperwork, and hope a mild winter plus decent wind saves the quarter. It is not elegant. It is how systems behave when long-cycle assets meet short-cycle shocks.
- Protect remaining dispatchable plants that can legally operate through the winter peak.
- Keep gas storage and LNG access as insurance even when the politics are ugly.
- Clear interconnection and nuclear licensing bottlenecks without pretending they are overnight fixes.
- Treat large new electricity demand, including compute, as a grid-planning issue rather than a branding issue.
- Speak honestly about bills, because surprise invoices destroy trust faster than a dull speech.
Will every government do those five things well? Of course not. Some will over-promise on SMRs. Some will under-invest in interconnectors. Some will keep telling voters that weather-dependent supply is already a complete substitute for firm power. Markets will grade the homework in real time.
How Investors Should Read The Next Season
If you follow utilities, miners, gas shippers, or industrial names with fat power bills, the next six months are less about a single headline and more about a stack of conditions. Watch gas benchmarks, coal generation hours, residual nuclear availability in France, German plant utilization, and any fresh disruption to LNG routing. Then watch whether wholesale power stays elevated long enough to leak into core inflation prints.
I would not treat a coal-margin rebound as a multi-year renaissance. The fleet is smaller, the politics are still hostile, and the long-run policy direction has not magically reversed. What you have is a stress response. Stress responses can last a season or two. They can also leave scars on industrial location decisions that last much longer than the spot price spike.
Nuclear names and suppliers will get a narrative tailwind. Fair enough. Separate the narrative from cash flow. Legal clearance in Italy is a beginning. Ten SMRs by 2035 is a plan. Revenue is a later chapter. The firms that earn money in the meantime are often the unglamorous ones: fuel logistics, grid equipment, maintenance, storage, and flexible generation that can still clear the market.
Rough mental model for the winter stack: Pricey gas sets the painful cap Remaining coal fills what it legally can Renewables cut the bill on good weather days Nuclear, where it exists, is the quiet stabilizer New nuclear is a 2030s story, not a January story
Why This Episode Feels Different From The Last One
The early-2020s shock taught Europe that pipelines can become weapons and that storage is not a boring accounting item. This episode adds two wrinkles. First, more coal is already gone, so the emergency buffer is thinner. Second, electricity demand from digital infrastructure is growing in a way that looks structural rather than cyclical. You can lecture a household about shorter showers. You cannot lecture a model-training cluster into using less power and still expect the same output.
That combination makes “just wait for more wind” a thinner answer than it was. Wind and solar remain essential. They still need companions: storage, interconnectors, demand shifting, and firm low-carbon plants. Without those companions, every geopolitical flare-up turns into a power-price flare-up, and every power-price flare-up turns into an inflation argument at the central bank.
I’ve sat through enough of these cycles to notice a pattern. The first month is about prices. The second month is about who can still generate. The third month is about politics. Then someone announces a long-term plan that cannot help the current quarter. The useful analysis lives in that gap between the plan and the peak hour.
A Ground-Level View Of Reliability
Reliability is not a vibe. It is a sequence of plants, wires, fuel deliveries, and operators who can say yes at 6:42 p.m. when the forecast was wrong. Coal’s sudden profitability is a symptom of that sequence getting tight. Nuclear’s political comeback is another symptom. Gas’s price spike is the trigger. Renewables’ variability is the background condition that never left.
If you work in a business that lives or dies on stable power, the practical questions are dull and important. Can you hedge? Can you shift load? Can you site the next expansion where the grid is actually long, not just where the press release is friendly? Those questions beat another round of abstract argument about which fuel is morally prettier this week.
Europe can still get through a difficult winter without drama. Mild weather and decent hydro would help. Extra LNG cargoes would help. Plants that stay online would help. Hope is not a strategy, but physics plus logistics still decide the outcome more than commentary does. The honest stance is watchful, not theatrical.
What To Watch After The Headlines Fade
After the first wave of stories about coal beating gas on margin, attention will drift. It always does. The useful follow-through is slower. Did coal hours actually rise by anything like a quarter? Did gas generation drop, or did a warm spell paper over the math? Did Italian nuclear paperwork turn into site work? Did SMR announcements attract real supply-chain commitments or just slides?
Also watch the tone from rate-setters. If energy prices cool, the upper bound of “neutral” becomes easier to defend. If they do not, the phrase “mild restrictive” stops being a footnote. Power markets and money markets are talking to each other again. That conversation is easy to ignore until it shows up in a print that markets cannot shrug off.
I keep a simple bias. Firm power is going to be valued more honestly over the next decade than it was during the years when cheap gas made the transition look smoother than it was. That does not mean every reactor proposal is a buy. It means the system is rediscovering that electrons have to be made on purpose, in volume, on bad weather days. The current coal-versus-gas flip is just the latest reminder, written in euros per megawatt hour.
And if you only remember one thing from this stretch, remember the constraint that does not show up in a slogan. Europe can prefer a fuel. It cannot run a plant that was already torn down. The next chapter belongs to whatever can be built, licensed, connected, and kept online when the wind drops and the tanker is late. That is the unfashionable core of energy security, and it is back on the table whether anyone planned for it or not.
