Electrolyzer Economics Beyond 45V: Where Efficiency Meets Bankability

Most electrolyzer investment cases are really 45V cases wearing a technology costume. They turn on a production credit that, under OBBBA, expires for any plant that has not begun construction by 1 January 2028. Strip the credit out, and a harder set of questions appears: what does the machine cost to run on real, intermittent power, and what is it worth to anyone when the subsidy is gone? The answers are not the ones the capex headlines suggest.

The Capex Race Is the Wrong Race

The loudest story in electrolysis is the collapse in stack prices. Chinese alkaline systems now land near $750 a kilowatt, against $1,300 for Western alkaline, $1,600 for PEM and roughly $2,500 for solid oxide. That looks decisive until you build up the cost of the hydrogen itself, because electricity, not the stack, is 60 to 75 percent of the levelised cost. A $10 per MWh move in power swings the hydrogen cost by about $0.50 a kilogram, which is a far larger lever than the entire spread between Chinese and Western stacks.

Read the chart by colour, not by bar. The electricity block is identical across every technology, because it is set by the power price and the system efficiency, not by who made the stack. The technologies differ only in the capex block, the smaller one. Cheaper Chinese alkaline takes PEM-equivalent hydrogen from roughly $4.55 to $3.19 a kilogram, a real saving, but it does not change the fact that the dominant cost is power. An operator who wins the capex race and loses the power-and-utilisation race has optimised the wrong variable.

The Utilization Paradox

Here is the part that catches even experienced investors. The instinct is to chase the cheapest electricity, which means dedicated wind or solar. But the cheapest renewable power is intermittent, intermittent power runs the electrolyzer at a low capacity factor, and a low capacity factor spreads the fixed capex over fewer kilograms, driving the per-kilogram cost up. The cheapest power can produce the dearest hydrogen.

The numbers are stark. Cheap intermittent power at $18 per MWh, run at a 20 percent capacity factor, yields hydrogen at about $7.64 a kilogram. Firmed renewable power at $32 per MWh, run at a 50 percent capacity factor, yields $4.55. The more expensive electricity makes the cheaper hydrogen, because utilisation, not the headline power price, is the master variable. The entire curve says the same thing: moving left, toward lower utilisation, costs more than almost any power-price saving can recover.

This is where policy turns against the operator. From 2030, 45V requires hourly matching, every megawatt-hour consumed must be matched to clean generation in the same hour. Without overbuilding generation or adding storage, hourly matching caps the hours an electrolyzer can run on qualifying power, which pushes the plant toward the low-capacity-factor left side of the curve, exactly the wrong direction for cost. The credit you are chasing actively fights the utilisation you need to be economic. Pricing 45V without pricing its matching constraint is how a model flatters a plant into a financing it cannot sustain.

The 2028 Cliff Changes the Question

OBBBA cut the 45V runway to the bone. The credit, worth up to $3 a kilogram at the lowest carbon-intensity tier, now terminates for any facility that has not begun construction before 1 January 2028, more than seven years earlier than the original schedule. For an asset class whose models lean on that credit to bridge a $4-plus production cost down to a sub-$2 number, an expiry inside the development window of most projects is not a detail. It means the underwriting question is no longer what the plant earns with 45V. It is what the plant earns without it.

What the Cancellations Already Taught

That question has been answered in the market, painfully. Through 2025 developers cancelled roughly 60 major clean-hydrogen projects, and only somewhere between 6 and 9 percent of announced capacity has reached a final investment decision. Fortescue walked away from sanctioned projects in Arizona and Gladstone; Air Products took a write-down of up to $3.1 billion and exited US plants, including one that lost its 45V eligibility on a hydropower technicality; BP cancelled its 1.5GW Duqm project. The common thread is not technology. It is the absence of a buyer. A merchant electrolyzer producing hydrogen at $4 to $7 against a grey-hydrogen price near $2, with a credit that cliffs in 2028 and no mandate pulling the molecule into a market, is an orphan. The lesson is to underwrite the offtake, not the credit.

Two failures are specific to the equipment, not the project, and both belong in diligence. First, the manufacturers themselves are surviving on selective marquee orders rather than volume: Plug Power has run a hard cost-cutting programme through a slower-than-expected electrolyzer market while still posting negative operating cash flow, and Nel and ITM have announced single large orders rather than a broad order book. A supplier living order-to-order is a counterparty risk, not just a vendor. Second, nameplate efficiency is not delivered efficiency. Stacks degrade, and the rate under real variable load, with start-stop cycling, is poorly characterised even for SOEC, which shows a tidy 0.5 percent per thousand hours only under controlled conditions and must still prove a 60,000-hour life. Underwrite the kilograms a stack delivers over its life with degradation, not the day-one number on the datasheet.

Which Electrolyzer to Watch: AWE, PEM, SOEC, AEM

If power and utilisation decide today’s economics, technology decides where the next cost step-change comes from. Four architectures compete, and they are not equivalent bets.

Technology	Edge	Constraint	Disruption potential
Alkaline (AWE)	Cheapest (~$750/kW in China), mature, durable	Slow ramp; poor fit for intermittent power; oversupplied	Low: the cost floor, not the disruptor
PEM	Fast ramp, compact, flexible for variable renewables	Requires iridium, a scarce platinum-group metal that caps terawatt-scale growth	Medium: flexibility is valuable, but iridium limits it
SOEC	Highest efficiency (uses heat); can co-electrolyse CO2 and water straight to syngas; reversible to a fuel cell	Degradation under variable load; needs a proven 60,000-hour life; wants steady high-temp heat	High: efficiency plus a direct path to e-fuel feedstock
AEM	Earth-abundant nickel/iron catalysts, no iridium; aims to pair alkaline cost with PEM-like flexibility	Catalyst-layer and membrane durability still unproven at scale; early commercial (Enapter, Ionomr)	High: if durability is solved, it breaks the iridium constraint

The investor read is that alkaline is the floor, not the future, and the genuine disruption sits in the two less-mature rows. SOEC matters because its efficiency attacks the dominant cost, electricity, and because co-electrolysis gives it a direct line into e-fuel syngas, the embedded route that survives the 45V cliff. AEM matters because iridium is a real, under-discussed ceiling on PEM: a technology that delivers PEM-like flexibility on nickel and iron, rather than a platinum-group metal, would change the supply-chain maths for the whole sector. Both bets turn on the same unglamorous variable, stack durability under real variable load, which is precisely the data point to demand and the one least often volunteered. Watch durability curves, not nameplate efficiency.

One frontier sits beyond even AEM and belongs on the radar rather than in a financing case: membrane-free electrolysis. By removing the membrane, and with it the platinum-group-metal catalysts, decoupled and supercapacitive architectures attack the cost ceiling and the supply-chain ceiling at once, and developers such as H2Pro are running early pilots at roughly the half-megawatt scale. The honest caveat is that the efficiency case is unproven at commercial scale and the long-run durability data does not yet exist, so this is a technology to track, not yet to underwrite. It earns its place by the same rule as the rest of the watchlist: judge it on durability and scale-up evidence, not on pitch-deck efficiency.

The Beyond-45V Revenue Stack

None of this makes electrolysis uninvestable. It relocates the value away from the credit and toward what survives it.

Only one line in the revenue stack actually dies at the cliff, and it is the one most projects are built around. The streams that survive are the ones a credit-led model tends to ignore. Flexibility is the largest: an electrolyzer is a fast, controllable load, and selling that controllability into reserve and ancillary-service markets can lift revenue by around 16 percent and cut the hydrogen break-even by five to six percent, while also relieving the utilisation problem by giving the plant something to earn when it throttles down. The oxygen co-product, about eight kilograms for every kilogram of hydrogen, is a real revenue line wherever a medical or industrial buyer sits nearby. Solid-oxide waste heat lifts effective efficiency when the host site needs heat. And the most durable option of all is to stop selling hydrogen merchant and embed it as a captive input to an e-fuel, where it inherits the RIN, LCFS and 45Z stack of the finished liquid and a demand that is mandated rather than hoped for.

Does the Crude Spike Help? Mostly Not, and That Matters

It is tempting to assume the 2026 run in crude lifts every clean-energy boat. For electrolytic hydrogen, it mostly does not, and seeing why sharpens the whole thesis. Green hydrogen does not compete against oil. It competes against grey hydrogen, which is made from natural gas by steam-methane reforming at roughly $1 to $2 a kilogram in the gas-rich United States and $5 to $6 in Europe and Asia. The benchmark that green hydrogen has to beat is set by the gas price, not the oil price, so a crude spike does little directly to close the gap.

There are two second-order effects, and they pull in opposite directions for different operators. In markets where gas is oil-indexed, Europe and much of Asia, a sustained crude move can drag gas, and therefore grey hydrogen, higher, modestly improving green’s relative position; in the United States, where Henry Hub gas is largely decoupled from crude, even that help is faint. The effect that actually matters is the one this analysis keeps returning to: a high oil price lifts the value of refined products and the e-fuels that displace them, which rewards hydrogen sold as a captive input to an e-fuel rather than as a merchant molecule. In other words, the crude spike helps the embedded electrolyzer and barely touches the merchant one. The macro reinforces the structure: take the hydrogen value inside the hydrocarbon, not beside it.

How We’d Underwrite It

Underwrite the plant on its 45V-off economics and treat the credit as upside, not the thesis, because the credit has an expiry inside the build window. Solve for utilisation before power price; a firmed, higher-capacity-factor configuration usually beats a cheaper-but-intermittent one, and the hourly-matching rule makes that truer over time. Insist on a non-merchant revenue case: a flexibility contract, a local oxygen buyer, or, best, a captive offtake into an e-fuel that carries its own demand and credit stack. Be sceptical of any model whose returns depend on the cheapest possible electricity at a capacity factor that power cannot actually sustain. And treat the Chinese capex story as a welcome tailwind, not a thesis, since it moves the smaller of the two big levers. In a market where the subsidy is leaving and the molecule has no natural buyer, the bankable electrolyzer is the one that earns its keep as a flexible, integrated asset rather than a merchant hydrogen machine.

What Clears an Investment Committee

1. Model the 45V-off case first. The credit cliffs at begin-construction in January 2028. If the plant only clears with $3/kg of credit, it does not clear.

2. Solve utilisation before power price. A low capacity factor spreads capex over fewer kilograms; cheap intermittent power can produce dearer hydrogen than firmed power at higher utilisation.

3. Price the hourly-matching constraint. From 2030, matching caps qualifying run-hours unless you overbuild or store. Build that into the capacity factor, not a footnote.

4. Demand a non-merchant revenue line. Ancillary services, an oxygen offtake, useful heat, or a captive e-fuel buyer. A merchant molecule with no demand-pull is the orphan that fills the graveyard.

5. Underwrite delivered kilograms, not nameplate. Apply a real degradation curve over the stack life and test the supplier’s balance sheet; an order-to-order vendor is a counterparty risk.

6. Keep capex in proportion. Cheaper stacks help, but power and utilisation decide the economics. Do not let a capex saving paper over a utilisation problem.

Electrolyzer Economics: Investor FAQ

Why does electricity, not capex, dominate green hydrogen cost?

Electricity is 60 to 75 percent of the levelised cost of hydrogen, because an electrolyzer consumes roughly 50 to 55 kWh per kilogram. A $10/MWh move in power changes the hydrogen cost by about $0.50/kg, a bigger swing than the entire gap between Chinese and Western stacks. Cheaper capex helps but moves the smaller lever.

What is the utilization paradox in electrolyzer economics?

The cheapest electricity is intermittent renewable power, which runs the electrolyzer at a low capacity factor. A low capacity factor spreads fixed capex over fewer kilograms, raising the per-kilogram cost. So the cheapest power can yield the dearest hydrogen: $18/MWh at 20 percent capacity factor gives ~$7.64/kg, while firmed $32/MWh power at 50 percent gives ~$4.55/kg.

When does the 45V hydrogen credit expire?

Under OBBBA, 45V terminates for any facility that has not begun construction before 1 January 2028, far earlier than the original schedule. It is worth up to $3/kg at the lowest carbon-intensity tier. Separately, hourly matching of clean power becomes mandatory from 2030.

How does 45V hourly matching affect the economics?

From 2030, every MWh consumed must be matched to clean generation in the same hour. Without overbuilding generation or adding storage, this caps the hours the electrolyzer can run on qualifying power, lowering its capacity factor, which raises LCOH. The credit’s own rule pushes the plant toward the high-cost end of the utilisation curve.

What is an electrolyzer worth without 45V?

Its durable value is as a flexible, integrated asset, not a merchant hydrogen machine. Reserve and ancillary-service participation can add ~16 percent to revenue; the oxygen co-product (~8 kg per kg H2) is saleable to nearby buyers; solid-oxide waste heat raises efficiency; and embedding the hydrogen in an e-fuel lets it inherit the RIN, LCFS and 45Z stack and mandated demand of the finished liquid.

Which electrolyzer technology has the most disruption potential?

Alkaline is the cheapest but the least disruptive, the cost floor. PEM adds flexibility but is capped by scarce iridium. The genuine disruption sits in SOEC (highest efficiency, and co-electrolysis straight to e-fuel syngas) and AEM (earth-abundant nickel/iron catalysts that sidestep iridium), with membrane-free designs a longer-dated radar item. Both hinge on proving stack durability under variable load, so watch durability data over nameplate efficiency.

Does the 2026 crude oil spike improve electrolyzer economics?

Directly, very little. Green hydrogen competes against grey hydrogen made from natural gas (about $1-2/kg in the US, $5-6/kg in Europe and Asia), so the gas price, not the oil price, sets the benchmark. Where gas is oil-indexed (Europe, Asia) a crude spike helps modestly; in the US it barely registers. The real benefit is indirect: high oil lifts e-fuel value, which rewards hydrogen embedded as a captive input to an e-fuel rather than sold merchant.

Methodology: LCOH built from public 2026 ranges (efficiency ~53 kWh/kg; capex alkaline ~$750-1,300/kW, PEM ~$1,600/kW, SOEC ~$2,500/kW; firmed power ~$32/MWh; 8% discount, 20-year life). Figures are illustrative and calibrated to disclosed anchors; assumptions are Trident’s framework and move with markets. Companies named (Enapter, Ionomr, Plug, Nel, ITM, H2Pro and others) are named as illustrative examples, not recommendations. Sources: DOE / NREL; Lazard LCOH; IEA; World Bank electrolyzer TEA; Treasury 45V final rules; OBBBA analyses; ancillary-service and membrane-free electrolysis literature. Analysis, not investment advice.