TECHNOLOGY EXPLAINER
The Cost of Producing Hydrogen
— building and reading the LCOH (levelized cost of hydrogen) yourself
A hydrogen price quoted as "US$ X per kg" is not set by the electricity price alone. Capital cost, capacity factor, discount rate, efficiency and stack life all sit in the same equation. This article builds the LCOH formula, tests it on a base case with every assumption written out, and then reads the figures published by the IEA, IRENA, the U.S. Department of Energy (DOE) and Japan's Basic Hydrogen Strategy to see which assumptions each number rests on.
- What LCOH is (the short version)
- The LCOH formula — cost on top, hydrogen output below
- The components — six dials and how public sources treat them
- Our calculation (1): breakdown of a base case
- Our calculation (2): sensitivity — which dial matters
- Our calculation (3): capacity factor combined with electricity price
- A materials engineer's view (1): what capacity factor changes
- How to read published figures and targets
- Putting public bodies' assumptions into our formula
- A materials engineer's view (2): where do materials improvements show up in LCOH?
- Limits, and what we could not confirm
- Glossary / References / Claim-to-source audit
Sourced = a value stated in published material (link given)
Our calculation = a figure this article derived from stated assumptions (every assumption is in a table)
Not yet confirmed = a plan, target or scenario with no confirmed outcome
Readings from a materials point of view are marked separately as Commentary.
This article explains the economics of hydrogen production. It is not advice on operating equipment, handling hydrogen or making investment decisions. Hydrogen is a flammable gas and its handling is governed by laws and standards.
1. What LCOH is (the short version)
- What it measures: LCOH (levelized cost of hydrogen) takes the total cost of a plant over its life and divides it by the hydrogen produced over the same period, giving a "per kg" figure. The IEA defines the production cost as the total expenditure to produce one unit of hydrogen and says that, for electrolysis, it depends mainly on electricity price, CAPEX, cost of capital, efficiency and utilization rate. It is a theoretical value that does not include commercial margins Sourced [Reference 1]
- Not only electricity: In our base case electricity is about 40% of LCOH, and capital cost, fixed O&M and stack replacement together are about 60% Our calculation. Even with free electricity a non-electric cost remains (Chapter 4).
- A number comes with its conditions: DOE's "US$2.00/kg in 2026" is a target that assumes electricity at or below US$0.03/kWh and equipment availability above 90% Sourced [Reference 4]. The IEA's estimates, DOE's targets and Japan's "yen per Nm3" targets are different yardsticks (Chapter 8).
2. The LCOH formula — cost on top, hydrogen output below
LCOH is defined by bringing each year's costs and hydrogen output to present value with a discount rate, then dividing: the present value of all costs ÷ the present value of all hydrogen produced. Converting the up-front investment into a level annual amount (the capital recovery factor, CRF) gives the simpler form used in this article.
LCOH = [ CRF × CAPEX + fixed O&M + annualized stack replacement ] ÷ annual hydrogen output + electricity price × specific energy consumption
• CRF (capital recovery factor) = r(1+r)n ÷ ((1+r)n − 1), where r is the discount rate and n the plant life in years
• Annual hydrogen output (per 1 kW of plant) = 8,760 h × capacity factor ÷ specific energy consumption (kWh/kg)
• Specific energy consumption (kWh/kg) = lower heating value (LHV) of hydrogen ÷ efficiency (LHV basis). This article uses an LHV of 33.3 kWh/kg, a value back-calculated from the tables in IRENA ("65% efficiency = 51.2 kWh/kg") and DOE ("65% = 51 kWh/kg"), with which both are consistent.
First, what the simplified form leaves out: tax, depreciation, inflation, residual value, financing structure (the mix of equity and debt), the cost of water, revenue from selling oxygen, and the cost of compression, storage and transport. The discount rate is treated as a single real rate, with inflation excluded. The IEA also leaves out water costs and oxygen revenue in its analysis [Reference 2].
That simplicity is also a limit of the LCOH idea itself. The IEA notes that one European hydrogen price index (IBHYX) is calculated with a project-finance model that goes beyond the traditional LCOH approach, incorporating separate return expectations for equity and debt and key investment variables Sourced [Reference 1]. LCOH is a guide to cost, not the price at which a project breaks even.
Other studies analyze hydrogen from the standpoint of an investor who optimizes operation against fluctuating wind output and electricity prices. Glenk and Reichelstein (Nature Energy, 2019) studied Germany and Texas and concluded that renewable hydrogen was already cost competitive in niche applications (the abstract gives EUR 3.23/kg), though not yet for industrial-scale supply; they projected this would change within about a decade, at a level they put at EUR 2.50/kg Sourced [Reference 7]. That is a 2019 outlook built on assumptions different from ours. It is cited here as an example of the point that an estimate moves with its date and its assumptions.
Converting efficiency to specific energy consumption
Efficiency (LHV basis) and specific energy consumption (the electricity needed to make 1 kg of hydrogen) are related by 33.3 ÷ efficiency, as in the formula above Our calculation.
| Efficiency (LHV basis) | kWh per kg | Source and standing |
|---|---|---|
| 58% | 57.4 | Efficiency the IEA assumes for Chinese electrolyzers in 2024 [Reference 2]. Conversion is ours |
| 61% | 55 | DOE's 2022 status for the whole system, PEM and alkaline (55 kWh/kg, 61%) [References 4, 5] |
| 63% | 52.9 | Efficiency the IEA assumes as the 2024 global average [Reference 2]. Our base case |
| 65% | 51.2 | IRENA's 2020 "today" assumption [Reference 3]. DOE's 2022 status at stack level is 51 kWh/kg (65%) for both PEM and alkaline [References 4, 5] |
| 77% | 43 | DOE's ultimate target for PEM stacks (43 kWh/kg) [Reference 4]. Not yet confirmed |
Stack efficiency (of the stacked cells) and whole-system efficiency (including rectifiers, gas treatment and so on) are different numbers; DOE gives both separately [References 4, 5]. The basis for capital cost also differs by source: the IEA gives electrolyzers in US$/kWe (electrical input) and gas reforming in US$/kWH2 (hydrogen output) [Reference 2], and some literature that IRENA reviews quotes cost per kW of hydrogen output on an HHV basis [Reference 3].
3. The components — six dials and how public sources treat them
The main dials that move LCOH are electricity price, capital cost, capacity factor, discount rate, life and stack replacement, and efficiency. The table shows how public sources treat each.
| Dial | What it represents | How public sources treat it |
|---|---|---|
| Electricity price | Cost per kWh of purchased or generated power | IRENA says cheap electricity is essential for competitive hydrogen and that lower electrolyzer costs cannot make up for a high electricity price [Reference 3]. The IEA's 2024 analysis uses a levelized cost of solar PV electricity of US$20-110/MWh and onshore wind of US$25-100/MWh [Reference 1] |
| Capital cost (CAPEX) | Electrolyzer system, balance of plant (BoP), engineering-procurement-construction (EPC) and contingency, summed | In the IEA survey, the installed cost of electrolyzers made outside China in 2024 was US$2,000-2,600/kW (alkaline only; includes equipment, gas treatment, plant balancing, EPC and contingency). Made and installed in China: US$600-1,200/kW [Reference 1]. More than half of the total cost is EPC and contingency, which depends on location [Reference 1] |
| Capacity factor | The share of the year, expressed as hours at rated output | IRENA notes that an electrolyzer coupled to PV could typically operate fewer than 2,000 hours a year [Reference 3]. The IEA analysis sets the electrolyzer's capacity factor equal to that of the renewable power plant [Reference 1] |
| Discount rate (cost of capital) | The return required on investment; the higher it is, the heavier up-front capital cost becomes | The IEA analyzes a cost of capital of 6-20% [Reference 1]. IRENA's 2020 assumption is 10% for "today's average" and 6% for "best" [Reference 3] |
| Life and stack replacement | Years of use, and how often a degraded stack is replaced | The IEA assumes a stack life of 50,000 hours (it can reach 95,000 hours; 50,000 was chosen from the IEA's analysis of the economic optimum allowing for degradation) [Reference 2]. DOE's 2022 status is 40,000 hours for PEM and 60,000 hours for alkaline, with a target of 80,000 hours [References 4, 5]. The IEA puts the stack at 15-20% of total investment [Reference 1] |
| Efficiency | The share of input electricity that ends up as the energy content of hydrogen | The IEA assumes 63% (LHV) as the 2024 global average and 58% for China [Reference 2]. The IEA says the lower efficiency of Chinese electrolyzers partly cancels the benefit of their lower capital cost [Reference 1] |
| Other | Fixed O&M, water, oxygen sales | The IEA sets annual O&M at 3% of CAPEX and does not consider water cost or oxygen revenue [Reference 2] |
All Sourced (published material from the IEA, IRENA and DOE). The IEA figures are analytical assumptions, not results from any particular project.
Elsewhere the IEA says that electrolyzers account for 55-65% of the levelized cost of hydrogen production, while about 60% of the investment required is EPC and the stack is only 15-20% of total investment Sourced [Reference 1]. That sentence does not state its conditions, and we could not confirm whether they match our base case. The breakdown in the next chapter (capital cost, O&M and stack replacement together at about 60%) is of a similar order, but because the conditions differ we do not treat the two as the same number.
4. Our calculation (1): breakdown of a base case
The base case takes the values the IEA used in its 2024 analysis as its foundation, and adds our own assumptions for the items the IEA did not specify (capacity factor, electricity price, discount rate, life). Every assumption is in Table 1. It does not represent the conditions of any particular project or manufacturer.
| Item | Base value | Basis and standing |
|---|---|---|
| Installed cost | US$2,300/kW | The 2024 "outside China" value used in the IEA's Figures 3.8 and 3.10 [Reference 1]. The IEA annex gives a range of US$2,000-2,600/kWe [Reference 2] |
| Efficiency | 63% (LHV) = 52.9 kWh/kg | IEA annex, 2024 global average [Reference 2]. LHV taken as 33.3 kWh/kg (Chapter 2) |
| Capacity factor | 50% (4,380 h/yr) | Our assumption. The IEA gives capacity factors of 12-35% for solar PV and 30-53% for onshore wind [Reference 1]. A middle value, imagining solar and wind combined |
| Electricity price | US$50/MWh | Our assumption. Held constant regardless of capacity factor (limits discussed in Chapter 6) |
| Discount rate | 8% (real) | Our assumption. The IEA range is 6-20% [Reference 1] |
| Plant life | 20 years | Our assumption |
| Fixed O&M | 3% of CAPEX per year | IEA annex [Reference 2] |
| Stack life | 50,000 h | IEA annex [Reference 2]. At 4,380 h/yr in the base case, one replacement in year 11 |
| Stack replacement cost | 20% of CAPEX | The upper end of the IEA's "stack is 15-20% of total investment" [Reference 1]. No fall in price or reuse at replacement is assumed (ours) |
| Water and oxygen | Excluded | The IEA does not consider them either [Reference 2] |
Table 1 Assumptions of the base case (all are premises of Our calculation; the figures in the basis column are themselves Sourced).
With these assumptions the CRF (r = 8%, n = 20 years) is 0.1019, and 1 kW of plant makes 8,760 × 0.5 ÷ 52.9 = 82.9 kg a year Our calculation. The terms then work out as follows.
| Term | Calculation | US$/kg | Share |
|---|---|---|---|
| Annualized capital cost | 0.1019 × 2,300 ÷ 82.9 | 2.83 | 43% |
| Fixed O&M | 0.03 × 2,300 ÷ 82.9 | 0.83 | 13% |
| Stack replacement | US$460/kW (20% of 2,300) in year 11, discounted, annualized, then ÷ 82.9 | 0.24 | 4% |
| Electricity | US$0.05/kWh × 52.9 kWh/kg | 2.64 | 40% |
| Total (LCOH) | 6.55 | 100% |
Table 2 Breakdown of the base case (Our calculation). Shares are rounded and may not sum to 100%.
The base-case LCOH is about US$6.5/kg; electricity is about 40% and the remaining about 60% is non-electric cost Our calculation. Even with free electricity, this plant and capacity factor would still leave about US$3.9/kg. A handy conversion: a US$10/MWh (1 US cent per kWh) change in the electricity price moves LCOH by about US$0.53/kg (10 US$/MWh times 52.9 kWh/kg).
This falls within the range the IEA shows. For 2022, a period of high energy prices, the IEA put the cost of hydrogen from renewable electricity at slightly under US$4/kg in locations with strong renewable resources and above US$10/kg in locations with less potential Sourced [Reference 1]. But our base case is not a reproduction of the IEA's estimates.
5. Our calculation (2): sensitivity — which dial matters
Starting from the base case, we change one item at a time and see how LCOH moves. The ranges are our choice, based on values that appear in the public sources of Chapter 3. The ranking changes with how the ranges are chosen, so look less at the order than at which items matter by an order of magnitude.
| Item | Base value | Favorable value | LCOH (US$/kg) | Unfavorable value | LCOH (US$/kg) |
|---|---|---|---|---|---|
| Capacity factor | 50% | 90% | 5.06 | 30% | 8.96 |
| Discount rate | 8% | 6% | 6.15 | 20% | 9.33 |
| Electricity price (US$/MWh) | 50 | 20 | 4.96 | 80 | 8.13 |
| Installed cost (US$/kW) | 2,300 | 900 | 4.17 | 2,600 | 7.05 |
| Plant life | 20 yr | 30 yr | 6.24 | 10 yr | 7.61 |
| Efficiency (LHV) | 63% | 70% | 5.89 | 58% | 7.11 |
| Stack life | 50,000 h | 80,000 h | 6.44 | 30,000 h | 6.82 |
Table 3 Sensitivity (Our calculation). Base-case LCOH is 6.55 US$/kg. Electricity price in US$/MWh, installed cost in US$/kW. Rows are ordered by size of swing.
What the table shows Our calculation:
- Capacity factor and discount rate matter on the same scale as electricity price and capital cost. At a 30% capacity factor LCOH is about US$9.0/kg, and at 90% about US$5.1/kg. A 6% discount rate gives about 6.2 and a 20% rate about 9.3 US$/kg. The discount-rate range runs to the ends of the cost-of-capital range the IEA analyzes (6-20%) [Reference 1].
- Plant life of 10 versus 30 years gives about 6.2 to 7.6 US$/kg. The more the life shortens, the faster the annualized capital cost grows (through the CRF).
- Stack life, over the range tried (30,000-80,000 h), swings LCOH by only about US$0.4/kg, small compared with the other items. This depends on the stack replacement cost (20% of CAPEX) and the number of replacements we assumed.
- Five percentage points of efficiency (63% to 58%) make a difference of about US$0.6/kg. Even if capital cost is low, low efficiency gives some of it back through the electricity bill (the IEA makes the same point [Reference 1]).
6. Our calculation (3): capacity factor combined with electricity price
In real projects, capacity factor and electricity price are linked: cheap electricity tends to come from variable renewables, which are available for fewer hours a year. Holding the rest of the base case fixed, the table combines the two.
| Capacity factor | Power 20 | Power 30 | Power 50 | Power 80 |
|---|---|---|---|---|
| 20% (1,752 h/yr) | 10.21 | 10.73 | 11.79 | 13.38 |
| 30% (2,628 h/yr) | 7.37 | 7.90 | 8.96 | 10.55 |
| 40% (3,504 h/yr) | 5.87 | 6.40 | 7.46 | 9.04 |
| 50% (4,380 h/yr) | 4.96 | 5.49 | 6.55 | 8.13 |
| 60% (5,256 h/yr) | 4.43 | 4.96 | 6.02 | 7.61 |
| 75% (6,570 h/yr) | 3.82 | 4.35 | 5.41 | 6.99 |
| 90% (7,884 h/yr) | 3.48 | 4.01 | 5.06 | 6.65 |
Table 4 LCOH (US$/kg, Our calculation). Column headings are electricity prices in US$/MWh. Installed cost of US$2,300/kW, efficiency of 63%, a discount rate of 8% and so on are those of the base case.
Table 4 shows that at low capacity factors a cheaper electricity price lowers LCOH by less. At a 20% capacity factor, cutting power from 50 to 20 US$/MWh lowers LCOH by only about US$1.6/kg, and LCOH stays above US$10/kg. Conversely, at a 90% capacity factor electricity is about 52% of LCOH (about 29% at 30%) Our calculation.
We hold the electricity price constant regardless of capacity factor. In practice, raising the annual capacity factor can require extra procurement, storage or grid purchases, which can raise the average cost of electricity. The IEA's regional analysis assumes optimal oversizing of the renewable plant to minimize the levelized cost of hydrogen [Reference 1]. Do not read the table as "raise the capacity factor and power will be available at this price." The relationship between the grid and renewables is covered in our explainer on the power grid and renewable energy.
7. A materials engineer's view (1): what capacity factor changes
In a plant with a low capacity factor, most of LCOH is annualized capital cost and O&M. At a 30% capacity factor, non-electric cost is about US$6.3/kg, about 71% of LCOH. Here materials and structures that make "plant per unit of hydrogen" cheaper matter most.
In a plant with a high capacity factor, electricity grows to about 52% of LCOH (at 90%), and efficiency (overpotential) and degradation matter. One point of efficiency changes the electricity bill, and degradation changes how many times the stack is replaced.
IRENA says that where operating hours are short, as in a plant connected to solar alone, a short operating life still translates into a longer calendar life, so materials that are cheaper but degrade faster could be an option Sourced [Reference 3]. Conversely, where electricity is expensive, efficiency must come first. So the "best" material depends on the project's capacity factor and electricity price (Commentary).
IRENA also says that changing the operating strategy, such as bidding into the balancing market to add revenue, could speed membrane degradation and bring more frequent replacement, and that raising current density increases output at the expense of faster degradation [Reference 3]. Materials need to be assessed not only by steady-state life but by operating patterns that include start-stop cycles and load changes.
8. How to read published figures and targets
The hydrogen cost figures that public bodies publish mix estimates of actual costs, scenario outlooks, targets and illustrative assumptions. They also refer to different stages of cost (Fig. 7). Before citing a number, checking the following four points reduces mix-ups.
- Which cost is it? Production cost (LCOH), supply cost to the import port (CIF), a price index, or a market price
- Estimate or target? A 2024 estimate, a 2030 scenario, or a policy target
- What is in the capital cost? Does it include installation (EPC) and contingency? Is it per kW of electrical input or of hydrogen output? On an HHV or LHV basis?
- Assumptions on electricity price and utilization: Are there conditions attached for the target to hold?
| Source | Figure | What kind of figure | Assumptions and cautions |
|---|---|---|---|
| IEA Global Hydrogen Review 2025 | Natural gas without CCUS: US$0.8-4.6/kg (2024) | Range of production cost estimates Sourced | A regional range reflecting natural gas prices of US$1.7-25/MBtu, CO2 prices of US$0-70/t-CO2, cost of capital of 6-20% and more. Water cost excluded [Reference 1] |
| Same | Renewable electrolysis: just under US$4 to over US$10/kg (2022) | Range of estimates for a year of high energy prices Sourced | Just under 4 where renewable resources are strong, over 10 where they are weak. In 2024 the lower boundary for renewable hydrogen (China) was broadly aligned with the upper limit for natural gas [Reference 1] |
| Same | 2030, some regions: US$2-4/kg | Outlook in the Stated Policies Scenario (STEPS) Not yet confirmed | Applies to parts of China, the Middle East, Australia and parts of the United States. Cost of capital 6-20%. Electrolyzer installed cost set at US$675/kW in China and US$1,600/kW elsewhere in 2030 (notes to Figs. 3.8 and 3.10) [Reference 1] |
| IRENA Green hydrogen cost reduction (2020) | Shows "today's average" and "best" conditions in a chart | Conditions of an analysis Sourced | Average: US$770/kW, 65% (LHV), US$53/MWh, 3,200 h, WACC 10%. Best: US$130/kW, 76%, US$20/MWh, 4,200 h, WACC 6% [Reference 3]. We do not transcribe values read off the chart; only the assumptions are shown |
| DOE Technical targets for PEM and alkaline electrolysis | Production cost: US$2.00/kg in 2026, US$1.00/kg in 2031 | Target Not yet confirmed | Values assume electricity at or below US$0.03/kWh and availability above 90%. System capital cost targets (excluding installation) for PEM: US$250/kW in 2026, US$150/kW ultimate [Reference 4] |
| Same | 2022 status: PEM above US$3/kg, alkaline above US$2/kg | Status (DOE's own analysis) Sourced | This is the lower bound even under the low-cost-electricity, high-availability conditions above. The PEM value is based on analysis in the H2NEW consortium, to be published in a DOE Program Record [References 4, 5] |
| Japan Basic Hydrogen Strategy (6 June 2023) | Hydrogen supply cost (CIF): ¥30/Nm3 (about ¥334/kg) in 2030, ¥20/Nm3 (about ¥222/kg) in 2050 | Policy target Not yet confirmed | A target for supply cost including imports (CIF), not production cost (LCOH). The strategy says that converting the March 2023 LNG price to a hydrogen supply cost gives ¥24/Nm3, so the target is on a par with recent fossil fuel prices. The 2050 figure of ¥20/Nm3 is described as a level at which hydrogen power generation costs no more than gas-fired generation [Reference 6] |
| Same | Electrolyzer targets for 2030: alkaline ¥52,000/kW, PEM ¥65,000/kW | Policy target Not yet confirmed | The strategy cites targets from the 2019 Hydrogen and Fuel Cell Strategy Roadmap, says it will continue to support development including auxiliaries such as rectifiers, and says it will consider revising the targets [Reference 6] |
Table 5 Main figures of public bodies and their assumptions. This article does not convert between yen and US dollars.
From Table 5 we read the following (Commentary).
- DOE's US$2.00/kg target excludes installation from capital cost and assumes cheap electricity. The IEA's estimates include EPC and contingency in capital cost. Even the words "capital cost" cover different things.
- Even for the 2022 status, under cheap electricity and high availability, DOE puts PEM above US$3/kg and alkaline above US$2/kg. Those conditions differ greatly from operation tied directly to solar or wind, whose capacity factors the IEA gives as 12-53% [Reference 1].
- Japan's "¥30/Nm3" is a supply cost including imports, not a production cost. It cannot be used as is as a guide for hydrogen made by electrolysis at home.
- The IEA says that for low-emissions hydrogen there is not yet a liquid market, so strictly speaking there is no representative "market price" [Reference 1]. A published "hydrogen price" may be a price index or the value of an individual contract.
- The IEA says that even using Chinese electrolyzers with lower capital cost, the cost saving for renewable hydrogen in Europe is limited to 3-13%, and in some configurations LCOH can even rise slightly. The reasons are the large EPC share and lower efficiency [Reference 1].
9. Putting public bodies' assumptions into our formula
Here we put the assumptions of Chapter 8 into our formula as they stand. This is not a reproduction of any public body's estimate; it is a calculation to see how LCOH moves when the assumptions change. Each body uses more detailed assumptions on discount rate, installation cost, tax and so on that differ from ours, so the results will not match published figures.
| Case | Capital cost | kWh per kg | Electricity | Capacity factor | Discount rate | LCOH (US$/kg) |
|---|---|---|---|---|---|---|
| DOE 2026 PEM target conditions | US$250/kW (excluding installation) | 51 kWh/kg | US$30/MWh | 90% | 8% | 1.76 |
| Same, capital cost doubled (installation assumed on top) | US$500/kW | 51 kWh/kg | US$30/MWh | 90% | 8% | 1.99 |
| IRENA "today's average" conditions | US$770/kW | 51.2 kWh/kg | US$53/MWh | 3,200 h (36.5%) | 10% | 4.53 |
| IRENA "best" conditions | US$130/kW | 43.8 kWh/kg | US$20/MWh | 4,200 h (47.9%) | 6% | 1.04 |
| IEA 2024 "China" capital cost and efficiency (power etc. as in the base case) | US$900/kW | 57.4 kWh/kg | US$50/MWh | 50% | 8% | 4.53 |
Table 6 Results of putting public bodies' assumptions into our formula (Our calculation). Life 20 years, O&M 3%/yr, stack replacement at 20% of CAPEX (stack life 80,000 h in the DOE and IRENA rows, 50,000 h in the IEA row).
Three things can be read from this Our calculation.
- DOE's US$2.00/kg comes out at a similar order of magnitude in our formula. With US$250/kW (excluding installation) we get about US$1.8/kg, and doubling to allow for installation and the like gives about US$2.0/kg. We could not confirm the breakdown of DOE's figure (treatment of installation, discount rate, tax), so we do not identify which assumptions give 2.00.
- Under IRENA's conditions, "average" and "best" differ by about US$3.5/kg. The effects of capital cost, electricity, efficiency, operating hours and cost of capital all moving in the favorable direction at once are stacked together. IRENA's "best" is an illustration that puts several assumptions at their best at the same time.
- Even at a Chinese installed cost of US$900/kW, an efficiency of 58% makes LCOH about US$0.4/kg higher than lowering only the capital cost to 900 (about US$4.2/kg). This points the same way as the IEA's remark that low efficiency partly cancels the advantage of low capital cost [Reference 1].
10. A materials engineer's view (2): where do materials improvements show up in LCOH?
Improvements in materials (catalysts, membranes, porous transport layers, bipolar plates, seals) reach LCOH by three routes: efficiency (electricity cost), durability (stack replacement) and the cost of making the stack (capital cost). For the base case we calculated the effect of moving each.
| Materials-side change | Change from base case | LCOH (US$/kg) | Change |
|---|---|---|---|
| Lower overpotential (+1 point of efficiency) | Efficiency 63% → 64% (LHV) | 6.44 | −0.10 |
| Longer stack life | 50,000 h → 80,000 h (DOE's target value) | 6.44 | −0.10 |
| Extra electricity from degradation (the opposite direction) | End-of-life voltage up 10%, average energy consumption over the period up 5% | 6.68 | +0.13 |
| Stack manufacturing cost halved | Stack from 20% to 10% of capital cost, capital cost US$2,300 → 2,070/kW | 6.05 | −0.50 |
| (Reference) US$100/kW off total capital cost | US$2,300 → 2,200/kW | 6.38 | −0.17 |
Table 7 Effect of materials-side improvements (Our calculation). Other conditions as in the base case. The "degradation" row is a rough estimate that treats the base efficiency as the beginning-of-life value, with voltage rising linearly and current efficiency constant, based on DOE's definition (a 10% voltage rise at end of life) [Reference 4].
The IEA says the stack is only 15-20% of total investment and that about 60% of the investment is EPC [Reference 1]. In our base case, stack-related cost (the initial stack plus its replacement) comes to about US$0.97/kg, about 15% of LCOH Our calculation. Halving the cost of making the stack lowers LCOH by only about US$0.50/kg. Cheaper membranes and catalysts alone will not move most of LCOH.
Materials still matter because they move electricity cost and stack replacement through efficiency and durability. DOE's target table lines up, for PEM stacks, a reduction of platinum group metal loading from the 2022 status of 3.0 mg/cm² to 0.5 mg/cm² in 2026 and 0.125 mg/cm² ultimately; a cut in degradation rate from 4.8 mV/kh (0.25%/1,000 h) to 2.3 mV/kh (0.13%/1,000 h); and a lower voltage at a given current density (2.0 A/cm² at 1.9 V → 3.0 A/cm² at 1.8 V, and ultimately 1.6 V) Not yet confirmed [Reference 4]. DOE notes that the performance, durability and capital cost targets must be met simultaneously on the same stack or system [Reference 4].
DOE also writes that reducing platinum group metals should not compromise performance and durability [Reference 4]. IRENA likewise says that more active catalysts (ruthenium, for example) have long-term stability problems and that catalysts that are less robust are more prone to dissolution [Reference 3]. An improvement to a material can be checked against a single yardstick, LCOH, to see whether it improved one item at the expense of another (Commentary).
This article does not evaluate how platinum group metal prices affect LCOH, because we did not confirm primary sources on precious-metal prices.
11. Limits, and what we could not confirm
(1) The formula does not express a project's profitability
Tax, depreciation, financing structure, subsidies and residual value are not included. The IEA also says production cost is a theoretical value that excludes commercial margins [Reference 1]. The effect of policy on cost lies outside this article, whether the price-gap support under Japan's Hydrogen Society Promotion Act, EU schemes or U.S. tax credits.
(2) Electricity price is held constant
As Chapter 6 explains, the relationship between capacity factor and electricity price is not simple. The relationship between the grid and renewables is covered in our explainer on the power grid and renewable energy.
(3) Some public figures we could not read
- We could not confirm numerical values in the IRENA text for the chart showing LCOH under 2020 "today" and "best" conditions, so we do not transcribe them.
- We could not confirm the conditions behind the IEA statement that "electrolyzers are 55-65% of LCOH."
- We could not confirm the breakdown (installation, discount rate, tax and so on) of DOE's 2026 and 2031 targets (US$2.00 and US$1.00/kg). We did not consult the details of DOE's hydrogen production analysis model (H2A).
- In the IEA annex we consulted, we could not confirm a standard value for electrolyzer life (years) or discount rate, so a 20-year life and an 8% discount rate are our assumptions.
(4) We could not confirm actual domestic production costs in Japan
The targets in Japan's Basic Hydrogen Strategy are for supply cost including imports (CIF) [Reference 6]. We did not confirm from a primary source any actual production cost of electrolytic hydrogen in Japan.
(5) Other hydrogen sources are out of scope
For LCOH of natural gas reforming, with CCS and coal gasification we only cite the IEA's ranges. CCS is covered in our explainer on CCS.
- LCOH is "annual total cost ÷ annual hydrogen output." The IEA defines production cost as a theoretical value excluding commercial margins Sourced
- The base-case LCOH is about US$6.5/kg. Electricity is about 40%; the rest is annualized capital cost, O&M and stack replacement Our calculation
- Capacity factor, discount rate, electricity price and capital cost matter on the same scale. Capacity factors of 30% and 90% give about US$5.1-9.0/kg Our calculation
- Read a public figure by asking what the cost is, whether it is an estimate or a target, whether capital cost includes installation, and what electricity and utilization it assumes. DOE's target assumes cheap electricity and high availability; Japan's "¥30/Nm3" is a CIF target Sourced
- The stack is about 15% of LCOH. Improvements in materials act through efficiency and durability on electricity cost and stack replacement Our calculation
12. Glossary
- LCOH
- Levelized cost of hydrogen. The total cost over a plant's life divided by the hydrogen produced in the same period, in cost per kg.
- CAPEX / OPEX
- Capital expenditure / operating expenditure. CAPEX here means installed cost, including equipment, balance of plant, EPC and contingency.
- CRF (capital recovery factor)
- A factor that converts an up-front investment into a level annual amount, given life and discount rate: r(1+r)^n ÷ ((1+r)^n − 1).
- Discount rate (cost of capital, WACC)
- The return required on an investment. Used to bring future costs and hydrogen output to present value. WACC = weighted average cost of capital.
- EPC
- Engineering, procurement and construction. The IEA shows it as more than half of an electrolyzer's installed cost.
- BoP (balance of plant)
- Equipment other than the electrolyzer stack itself, such as rectifiers, gas treatment, cooling and water treatment.
- Stack
- The core of an electrolyzer, built by stacking cells. Made of membranes, catalysts, transport layers and bipolar plates.
- LHV / HHV
- Lower / higher heating value: whether the heat of condensation of water vapor is counted. The same equipment shows a different efficiency figure on each basis.
- Specific energy consumption
- Electricity needed to make 1 kg of hydrogen (kWh/kg). The inverse view of efficiency.
- Capacity factor
- The share of the year, expressed as hours at rated output.
- Degradation rate (mV/kh)
- Rise in cell voltage per 1,000 hours. DOE defines end of life as the point at which voltage has risen by 10%.
- Nm3
- Gas volume at standard conditions (0 °C, 1 atm). For hydrogen, 1 Nm3 is about 0.09 kg.
- CIF
- Cost, insurance and freight: a trade term for a price that includes freight and insurance. The basis of Japan's hydrogen supply cost target.
- STEPS
- The IEA's Stated Policies Scenario, an outlook that assumes only policies already announced.
- Sensitivity analysis
- Changing one input at a time, with the others held fixed, to see how the result moves.
13. References (primary sources)
- IEA "Global Hydrogen Review 2025" (full PDF consulted: Chapter 3 "Hydrogen production", "Five key questions", Box 3.2, Figs. 3.8-3.10 and 3.15-3.16; printed page numbers: Fig. 3.8 = p. 93, Box 3.2 = pp. 94-95, Fig. 3.9 = p. 98, EPC and capital cost survey = pp. 106-107, Chinese electrolyzers = pp. 109-111, "Five key questions" = p. 27) — iea.org
- IEA "Global Hydrogen Review 2025 Assumptions Annex" (PDF, p. 6 "Production pathways") — iea.blob.core.windows.net
- IRENA "Green hydrogen cost reduction: Scaling up electrolysers to meet the 1.5°C climate goal" (2020, PDF; Executive Summary Figs. ES1 and ES2, p. 43 "Lifetime aspects") — irena.org
- U.S. Department of Energy (DOE) "Technical Targets for Proton Exchange Membrane Electrolysis" — energy.gov
- U.S. Department of Energy (DOE) "Technical Targets for Liquid Alkaline Electrolysis" — energy.gov
- Ministerial Council on Renewable Energy, Hydrogen and Related Issues (published on the Cabinet Secretariat website) "Basic Hydrogen Strategy" (in Japanese), 6 June 2023 (PDF; "Reducing supply cost", printed pp. 11-12, and electrolyzer cost targets, printed pp. 26-27) — cas.go.jp
- Glenk, G. and Reichelstein, S. "Economics of converting renewable power to hydrogen", Nature Energy 4, 216-222 (2019), DOI 10.1038/s41560-019-0326-1 (abstract consulted) — nature.com
14. Claim-to-source audit
| Claim in the text | Basis | Label |
|---|---|---|
| That the IEA defines production cost as the total expenditure to produce one unit of hydrogen, depending for electrolysis on electricity price, CAPEX, cost of capital, efficiency and utilization, as a theoretical value excluding commercial margins; that a price index is an indicative theoretical market value and a market price is the value in an actual transaction (with no liquid market yet for low-emissions hydrogen); and that the IBHYX index is calculated with a project-finance model going beyond the traditional LCOH approach | IEA Global Hydrogen Review 2025 (Box 3.2, pp. 94-95) Reference 1 https://www.iea.org/reports/global-hydrogen-review-2025 | Sourced |
| The IEA analysis: production cost from natural gas without CCUS of US$0.8-4.6/kg in 2024 (gas price US$1.7-25/MBtu, CO2 price US$0-70/t-CO2); renewable electrolysis in 2022 from just under US$4/kg to over US$10/kg; in 2024 the lower boundary for renewable hydrogen (China) broadly aligned with the upper limit for natural gas; US$2-4/kg in some regions in 2030 (STEPS); electrolyzer installed cost of US$900 (China) and US$2,300/kW (elsewhere) in 2024 and US$675 and US$1,600/kW in 2030; cost of capital of 6-20%; water cost excluded; ranges of levelized electricity cost and capacity factor for solar PV and onshore wind; electrolyzer capacity factor equal to that of the renewable plant | IEA Global Hydrogen Review 2025 (notes to Figs. 3.8-3.10, pp. 92-98) Reference 1 https://www.iea.org/reports/global-hydrogen-review-2025 | Sourced |
| The IEA survey: 2024 installed cost of electrolyzers made outside China of US$2,000-2,600/kW (alkaline only; including equipment, gas treatment, plant balancing, EPC and contingency) and of US$600-1,200/kW for those made and installed in China; more than half of total cost is EPC and contingency; Chinese electrolyzers' lower efficiency partly offsets their lower capital cost, with a 3-13% saving for renewable hydrogen in Europe and, in some configurations, slightly higher LCOH | IEA Global Hydrogen Review 2025 (pp. 106-111, Figs. 3.15-3.16) Reference 1 https://www.iea.org/reports/global-hydrogen-review-2025 | Sourced |
| The IEA statement that electrolyzers account for 55-65% of the levelized cost of hydrogen production, about 60% of investment is EPC and the stack is 15-20% of total investment; the conditions of this statement are not given | IEA Global Hydrogen Review 2025 (Five key questions, p. 27) Reference 1 https://www.iea.org/reports/global-hydrogen-review-2025 | Sourced |
| The IEA annex: 2024 electrolysis capital cost (global average US$2,000-2,600/kWe, China US$600-1,200), efficiency (global average 63%, China 58%, LHV basis), annual O&M of 3% of CAPEX, stack life of 50,000 hours (up to 95,000 hours possible; 50,000 chosen from an economic optimum allowing for degradation); water cost and oxygen revenue not considered | IEA Global Hydrogen Review 2025 Assumptions Annex (p. 6) Reference 2 https://iea.blob.core.windows.net/assets/15673ab3-a86a-4434-bff4-490bb42d3563/GlobalHydrogenReview2025AssumptionsAnnex.pdf | Sourced |
| IRENA's 2020 report: electricity price is essential for competitive hydrogen and lower electrolyzer cost cannot make up for a high electricity price; the "today's average" (US$770/kW, 65% LHV, US$53/MWh, 3,200 h, WACC 10%) and "best" (US$130/kW, 76%, US$20/MWh, 4,200 h, WACC 6%) conditions; that PV-coupled electrolyzers could operate fewer than 2,000 hours a year, that with short operating hours cheaper but faster-degrading materials may be chosen, that participation in balancing markets or higher current density may speed degradation, and the long-term stability issue of ruthenium catalysts | IRENA Green hydrogen cost reduction (2020) Reference 3 https://www.irena.org/-/media/Files/IRENA/Agency/Publication/2020/Dec/IRENA_Green_hydrogen_cost_2020.pdf | Sourced |
| DOE targets for PEM electrolysis: US$2.00/kg in 2026 and US$1.00/kg in 2031; 2022 status (stack 51 kWh/kg and 65%, system 55 kWh/kg and 61%, degradation rate 4.8 mV/kh, lifetime 40,000 h, system capital cost excluding installation of US$1,000/kW, platinum group metals 3.0 mg/cm²) and the 2026 and ultimate targets; production cost premised on electricity at or below US$0.03/kWh and availability above 90%, with a status lower bound above US$3/kg; performance, durability and capital cost targets to be met simultaneously; end of life at a 10% voltage rise | DOE Technical Targets for PEM Electrolysis Reference 4 https://www.energy.gov/eere/fuelcells/technical-targets-proton-exchange-membrane-electrolysis | Sourced Not yet confirmed |
| DOE targets for liquid alkaline electrolysis: 2022 status (system 55 kWh/kg and 61%, degradation rate 3.2 mV/kh, lifetime 60,000 h, system capital cost of US$500/kW, production cost above US$2/kg) and the 2026 and ultimate targets | DOE Technical Targets for Liquid Alkaline Electrolysis Reference 5 https://www.energy.gov/eere/fuelcells/technical-targets-liquid-alkaline-electrolysis | Sourced Not yet confirmed |
| Japan's Basic Hydrogen Strategy (6 June 2023): hydrogen supply cost (CIF) targets of ¥30/Nm3 (about ¥334/kg) in 2030 and ¥20/Nm3 (about ¥222/kg, with hydrogen power generation cost at or below gas-fired) in 2050; ¥24/Nm3 when the March 2023 LNG price is converted; electrolyzer targets for 2030 of ¥52,000/kW (alkaline) and ¥65,000/kW (PEM), from the 2019 roadmap | Basic Hydrogen Strategy, 6 June 2023 (in Japanese) Reference 6 https://www.cas.go.jp/jp/seisaku/saisei_energy/pdf/hydrogen_basic_strategy_kaitei.pdf | Sourced Not yet confirmed |
| The study by Glenk and Reichelstein: for Germany and Texas, renewable hydrogen already competitive in niche applications (EUR 3.23/kg) and projected to become so for industrial-scale supply within about a decade (EUR 2.50/kg); a 2019 outlook | Glenk and Reichelstein, Nature Energy 4, 216-222 (2019) Reference 7 https://www.nature.com/articles/s41560-019-0326-1 | Sourced |
| The formula (CRF, annual hydrogen output, specific energy consumption); LHV = 33.3 kWh/kg (back-calculated to match IRENA's 65% = 51.2 kWh/kg and DOE's 65% = 51 kWh/kg); the base-case assumptions (Table 1); the breakdown (Table 2); sensitivity (Table 3, Fig. 4); the capacity factor and electricity table (Table 4, Fig. 5); calculations with public bodies' assumptions (Table 6); effects of materials-side improvements (Table 7); conversions (US$10/MWh = about US$0.53/kg and so on) | Our calculation. Assumptions are stated in Tables 1, 6 and 7. Capacity factor, electricity price, discount rate and life are our assumptions. The DOE, IRENA and IEA rows put only the published assumptions into our formula and are not reproductions of those bodies' estimates | Our calculation |
| Which materials matter depends on capacity factor; the reading that the stack is only part of LCOH; the four checks for reading public figures; the observation that DOE's target and the IEA's estimates differ in what capital cost covers; the four-stage layout of Fig. 7 | Our own arrangement and commentary based on the published material; not the views of any of the bodies | Commentary |
| The values read off IRENA's chart; the conditions of the IEA's "55-65%" statement; the breakdown of DOE's US$2.00 and US$1.00/kg and the details of the H2A model; official standard values for electrolyzer life and discount rate; actual domestic production costs of electrolytic hydrogen in Japan; the effect of platinum group metal prices; the effect of policy (subsidies, tax credits) | Could not be confirmed in published primary sources within the scope of this article (or not covered), so not stated | Commentary |
| That Figs. 2, 3, 4, 5 and 7 are explanatory drawings and not data from real projects, and that Figs. 1 and 6 are AI-generated images | A note by this article | Commentary |
Last updated October 2026. Sources are limited to primary material (published material from the IEA, IRENA and DOE, the Basic Hydrogen Strategy posted by the Cabinet Secretariat of Japan, and the abstract of a peer-reviewed paper). Estimates from market research firms are not used. Readings from a materials point of view are marked as Commentary and kept separate from sourced fact. The values read off IRENA's chart, the conditions of the IEA's "55-65%" statement, the breakdown of DOE's targets, actual domestic production costs of electrolytic hydrogen in Japan and the effect of platinum group metal prices are not stated here because no published primary source could be confirmed. The LCOH calculations in this article are derived from the formula and the stated assumptions; they are not measured values for any particular project or product. All figures are explanatory concept graphics or charts of our calculations. Figs. 2, 3, 4, 5 and 7 are vector drawings; Figs. 1 and 6 are AI-generated images, and none of them shows a real facility, device or product.