Hydrogen has attracted more narrative capital than actual capital deployment. Strip out the policy announcements and the conference keynotes, and what remains is a narrower, more honest picture: an industrial gas with a century of commercial use, a genuinely difficult storage and transport problem, and an electrolyzer cost curve that has not yet bent the way solar did in the 2010s.

Where It Actually Works Today

Roughly 95 million metric tons of hydrogen are produced and consumed globally each year, almost entirely as grey hydrogen derived from steam methane reforming. The buyers are not utilities or transport fleets. They are ammonia producers, petrochemical refiners, and steel mills. These are captive, industrial use cases where hydrogen is a feedstock, not a fuel, and the economics have been settled for decades.

Within that industrial base, a subset of operators is observing legitimate near-term substitution. Ammonia synthesis via green hydrogen carries a cost premium today, but the ammonia export corridor between Australia and Japan has structural policy backing on both sides, and several projects have moved past feasibility into front-end engineering. The feedstock substitution play, not the transport fuel play, is where commercial activity is measurably concentrating.

Where It Is Still Pilot-Stage

Hydrogen mobility is the category most exposed to marketing inflation. Fuel cell heavy trucks have demonstrated operational viability in controlled corridors, notably Hyundai’s XCIENT deployments in Switzerland and early California fleet trials, but the refueling infrastructure constraint is structural, not solvable by announcement. Green hydrogen at the pump currently costs between four and eight dollars per kilogram in most developed markets, against a cost parity threshold closer to two dollars for serious fleet economics.

• Long-duration grid storage via hydrogen remains sub-commercial. Round-trip efficiency losses of 60 to 70 percent are a physics constraint, not an engineering iteration.
• Residential heating via hydrogen blending into gas networks is proceeding in the UK under regulated trials, but blending above 20 percent by volume requires appliance replacement at scale.
• Steel decarbonization via direct reduced iron is the most structurally credible emerging use case. SSAB’s HYBRIT project in Sweden produced fossil-free steel commercially in 2021, and Thyssenkrupp has committed capacity in Germany, though both remain dependent on green hydrogen supply that does not yet exist at required volumes.

The Economics Question

Green hydrogen production costs are approximately four to six dollars per kilogram in most geographies today. The threshold at which green hydrogen begins to displace grey in industrial settings is generally modeled at two dollars or below, which requires electrolyzer capital costs near 300 dollars per kilowatt and low-cost renewable electricity input consistently under two cents per kilowatt-hour. Neither condition exists at scale today.

The Inflation Reduction Act’s Production Tax Credit of three dollars per kilogram for qualifying clean hydrogen shifts the domestic U.S. calculus materially, and several electrolysis projects that would otherwise have been unfinanceable are now in development. The structural question is whether that subsidy bridges to cost-competitive production or simply creates a compliance-dependent industry. Observers closer to project finance than to policy tend toward skepticism on the latter scenario.

The Operator Read

The industrial feedstock layer is the durable commercial signal. Operators watching ammonia, refining, and green steel supply chains are closer to actual hydrogen demand than anyone watching mobility announcements. The electrolyzer manufacturing buildout, concentrated currently in China and a handful of European firms including Nel ASA and ITM Power, is the upstream leverage point if cost curves do eventually move. The honest framing is that hydrogen is a real industrial commodity with a conditional future as an energy carrier, and that conditional clause still carries significant weight.

Natural gas moves freely in most of the country. In the Northeast, it does not. The Appalachian Basin sits on some of the most productive shale formations in North America, yet constrained takeaway capacity means that gas produced in Pennsylvania and West Virginia routinely clears at a significant discount to Henry Hub, while consumers in New England pay among the highest electricity prices in the continental United States. The gap between those two facts is structural, not cyclical.

The Infrastructure Ceiling

The core problem is pipeline throughput into and across the Northeast corridor. Projects like the Constitution Pipeline and the Northeast Supply Enhancement were either abandoned or blocked through a combination of state-level permitting denials, environmental litigation, and Federal Energy Regulatory Commission procedural friction. The Atlantic Bridge expansion added marginal capacity in 2017, but aggregate firm capacity into New England has not grown meaningfully in nearly a decade.

The result is basis differential volatility that operators and power generators cannot hedge efficiently. During cold snaps, Algonquin Citygate spot prices have historically spiked to multiples of Henry Hub, sometimes exceeding $30/MMBtu when the national benchmark sits below $3.00. That spread is not a trading anomaly; it reflects a hard physical ceiling on deliverable supply during peak demand windows.

How This Reprices Power

New England’s grid operator, ISO-NE, runs a capacity market that already prices in the region’s supply risk premium. But the more observable dynamic is on the energy side: gas-fired generators operating under firm transportation contracts have a structural cost advantage over those relying on interruptible service, and that advantage widens materially in winter. The constraint also keeps older oil-fired peakers economically relevant in a region that would otherwise have retired them. Constrained pipeline access effectively subsidizes fuel oil’s grid role through scarcity pricing.

• Winter firm transport contracts on Algonquin and Tennessee Gas Pipeline trade at persistent premiums over interruptible equivalents.
• LNG import terminals at Everett, Massachusetts remain operationally relevant specifically because pipeline alternatives are insufficient during demand peaks.
• Electricity consumers in Massachusetts and Connecticut carry a winter capacity cost embedded in their rates that most of the country does not.

The Longer Structural Read

Several dynamics are converging that complicate any near-term resolution. The political environment in Massachusetts and New York continues to resist new fossil fuel infrastructure at the state permitting level, regardless of federal authorization. Meanwhile, electrification mandates are increasing winter peak electricity demand even as the region’s dispatchable gas capacity faces fuel supply uncertainty during the same peak windows. That combination tightens the margin for error in grid operations.

Offshore wind build-out is progressing, but intermittency and the 2023 to 2024 contract renegotiations that stalled several major projects suggest the transition timeline is less linear than policy documents imply. The gap between current dispatchable capacity and the theoretical buildout trajectory is where physical risk concentrates.

The Operator Read

Capital allocators evaluating Northeast power assets are looking at a market where structural scarcity, not demand growth, is the primary pricing driver. Firm transportation rights, peaking assets with fuel flexibility, and generation positioned in constrained load pockets carry observable optionality that is difficult to replicate through financial instruments alone. The constraint is not new, but the convergence of electrification pressure with stalled infrastructure permitting is deepening it. Observers with exposure to capacity markets or physical generation in this region are watching basis spreads and ISO-NE forward capacity auction results as more revealing signals than headline gas prices.

Across PJM, MISO, and several Southeast balancing authorities, coal plants that were penciled in for retirement are running. Not because the politics shifted back, but because the load math did. Data center buildout, onshoring of energy-intensive manufacturing, and the slower-than-projected ramp of firmed renewables have created a capacity gap that coal, already depreciated and still connected, is filling by default.

Where the Lights Are Still On

The operating picture is not uniform. Roughly 170 GW of coal capacity remained on the U.S. grid entering 2024, down from a peak above 300 GW, but retirements have decelerated noticeably since 2022. PJM has granted retirement deferrals to plants that failed to secure deactivation approval, citing reliability concerns. In MISO, several utilities have filed to extend operating licenses on units originally slated to close in 2025 and 2026. Internationally, Germany restarted idled hard coal capacity during the 2022 energy crisis and kept units available through 2023. Poland has legally committed to coal through 2049, though that date is under revision.

India and Indonesia present a structurally different picture. Both are commissioning new coal capacity, not extending old. India added approximately 15 GW of coal-fired generation between 2022 and 2024. The IEA’s observation that global coal power generation hit record highs in 2023 is the detail that European energy policy tends to bracket out of the conversation.

The Policy Logic, Such As It Is

Regulators operating inside organized markets face a structural bind. Capacity markets were designed to price reliability, but they were not designed for a transition scenario where thermal retirements accelerate faster than storage and transmission can absorb the gap. The result is what FERC has described, cautiously, as a reliability gap risk in portions of the Eastern Interconnection through the late 2020s.

The policy response is essentially a managed delay. Utilities receive cost-of-service recovery in some jurisdictions to keep units available even when they are uneconomic on an energy basis. This is not a reversal of decarbonization targets on paper. In practice, it extends the operational runway of assets that were supposed to be off the grid.

The emissions accounting on extended coal operations versus the counterfactual is genuinely contested. Running an old coal unit at low capacity factor to provide grid insurance is a different calculus than running it as baseload. The structural tension between reliability mandates and emissions trajectories has not been resolved; it has been deferred.

What the Asset Picture Reflects

Thermal generation assets trading below replacement cost, which most coal plants do, carry a specific kind of optionality. The optionality is not that coal makes a structural comeback. It is that the transition timeline has a range of outcomes, and fully depreciated assets with existing grid interconnection and fuel supply agreements sit inside that range with low carrying cost.

Several infrastructure-focused capital allocators have noted, without publicity, that distressed coal plant acquisitions in PJM corridors have drawn interest from buyers who are not coal operators at all. The thesis is site control, interconnection queue position, and existing transmission access, assets that take years to replicate from greenfield.

The Operator Read

The structural observation for operators and allocators is not that coal is resilient as a fuel story. It is that the grid transition has dependencies that are running behind schedule, and that creates an environment where assets positioned as reliability infrastructure rather than generation assets carry different logic than the headline narrative suggests. The capacity market reform conversations happening inside FERC and at the state commission level will determine how long that logic holds. Operators watching those dockets are watching the real timeline.