A 1 GW proton exchange membrane electrolyzer requires approximately 300–500 kg of iridium, and roughly 400 kg per gigawatt is a practical screening figure for anodes using iridium-based catalyst [S3][S4][S6].
The requirement applies to installed electrolyzer capacity rather than hydrogen output, so annual iridium consumption depends on capacity additions, catalyst loading, operating hours and the useful life of each stack.
Working estimate and unit conversion
For procurement and early-stage design, use 400 kg of iridium per 1 GW of PEM electrolyzer capacity as the base case, with 300–500 kg as the wider loading range [S3][S4][S6]. The figure converts to 0.4 g/W, 400 g/kW, 400,000 g/GW and 0.40 metric tons/GW.
One supply model uses 0.65 g of iridium per kW, equal to 650 kg/GW, and estimates that a 10% allocation of 7.1 metric tons of annual iridium supply supports 1.1 GW of new PEM capacity per year [S7]. This is a scenario allocation, not a universal physical requirement, but it demonstrates how quickly a loading assumption affects supply calculations.
Where the iridium is used
Iridium is the principal anode catalyst element in PEM electrolysis, while platinum is also present in the cathode; consequently, anode iridium represents much of the PEM stack’s platinum group metal inventory [S1]. The iridium estimate must therefore not be interpreted as the stack’s total PGM mass.
The U.S. Department of Energy target table reports combined anode and cathode PGM loading of 0.8 g/kW for the 2022 status, 0.1 g/kW for its 2026 target and 0.03 g/kW for the ultimate target [S1]. These are total PGM figures, not iridium-only values, so they cannot be inserted directly into an iridium calculation without a catalyst-specific iridium share.
The same DOE table pairs the 0.8 g/kW status value with stack operation at 2.0 A/cm² and 1.9 V per cell, a 40,000-hour life and a 4.8 mV/kH degradation rate. Its 2026 target calls for 3.0 A/cm², 1.8 V per cell, 80,000 hours and 2.3 mV/kH, with 0.1 g/kW combined PGM loading [S1]. Lower catalyst loading must therefore be evaluated alongside efficiency and durability rather than selected on mass alone.
Comparison of catalyst-loading cases

The practical choices can be screened by iridium mass, DOE PGM target alignment and supply exposure. S6 presents a broader commercial range of 300–500 kg of iridium per GW, while S7 gives 650 kg/GW for its supply-model case [S6][S7].
Those figures identify a development trajectory, not a guaranteed fleet-wide average, and require validation against current density, voltage degradation and stack lifetime.
Sizing iridium against annual capacity
At 400 kg/GW, every 1 GW added to a PEM fleet consumes 400 kg of iridium before accounting for replacement stocks. A 5 GW build consumes 2,000 kg, a 10 GW build consumes 4,000 kg and a 100 GW program consumes 40 metric tons at the same loading assumption [S3][S4].
That calculation rests on the supply and loading assumptions of the cited publication. It does not mean that every future PEM project will require 400 kg/GW, because lower loadings, recycling and competing industrial demand change the allocation.
Suitability for buyers and technology planners

The 300–500 kg/GW range is suitable for early-stage material-intensity screening, supplier discussions and comparison of PEM deployment scenarios [S3][S4][S6]. It is not sufficient for a final stack bill of materials, because a qualified stack supplier’s loading, catalyst inventory and replacement policy should govern the commercial commitment.
Iridium is a critical component of PEM electrolysers, while Enapter's AEM electrolysers do not use iridium [S3].
Procurement teams should request three vendor-specific items: iridium mass per GW at rated power, the operating point used for that loading, and the treatment of catalyst replacement during the stack life. Without those data, a 400 kg/GW planning factor can conceal material differences between products. For recovered material, iridium recovery from spent PEM electrolyzer stacks becomes a separate feedstock and supply-chain workstream.
Technical constraints behind lower loading
Iridium thrifting is constrained by the need to retain catalytic activity, operational voltage and service life. DOE performance targets require 80,000 operating hours at the 2026 and ultimate target levels, with the same degradation and efficiency framework applied at higher current density and lower cell voltage [S1].
The DOE status case contains a 2.5% efficiency allowance for hydrogen crossover, and stack life is defined around a 10% end-of-life voltage loss after conditioning [S1]. These design constraints mean that an iridium loading cannot be judged independently of membrane assembly, catalyst utilization, current distribution and degradation control.
Recovery improves material availability but does not eliminate replacement demand during operation. Circular supply models also combine recycling with substitution in other iridium-consuming applications and further catalyst thrift. WPIC estimated that substitution could free 45,000–67,000 ounces of annual iridium supply over the decade following its August 2023 assessment, subject to the stated substitution assumptions [S4].
Planning assumptions to document

Document the iridium estimate as loadings, not only stack or system totals. Record 400 kg/GW as the central screening case, 300–500 kg as the supporting range, 0.4 g/kW as the equivalent power-specific value, and whether the figure covers new-build catalyst inventory, initial stack charge or replacement demand [S3][S4][S6].
For bankability work, update the model with supplier-specific catalyst loading and the DOE performance constraints of 3.0 A/cm² at 1.8 V per cell, 80,000 hours and 2.3 mV/kH for the 2026 target [S1]. Track two signals: disclosed commercial iridium loading in g/kW and annual iridium allocated to PEM manufacturing. A project adding 1 GW at 400 kg/GW would require 400 kg, while 1 GW at 0.05 g/kW would require 50 kg [S3][S5].