Platinum and iridium have long been the bottleneck metals in green hydrogen production. These rare, expensive elements sit at the heart of today’s best water-splitting electrolyzers, and their cost has kept clean hydrogen too pricey to compete with fossil-fuel alternatives at scale. A cluster of peer-reviewed studies published between late 2024 and early 2025 is now drawing renewed attention from electrolyzer manufacturers and policy analysts as the hydrogen industry enters a critical scale-up phase in spring 2026. Those studies show that electrolyzers built without any platinum-group metals can run for more than 1,000 hours at industrial current densities, hitting durability benchmarks that the U.S. Department of Energy has set for next-generation systems. If these results survive the jump from lab cells to commercial stacks, they could strip out the single most expensive layer in the electrolyzer and fundamentally change the math on clean hydrogen.
Why platinum has been so hard to replace
Most commercial electrolyzers today use proton exchange membrane (PEM) technology, which relies on platinum for the hydrogen-producing cathode and iridium for the oxygen-producing anode. Together, these metals can account for a significant share of the electrolyzer stack’s material cost. As of early 2026, spot platinum trades near $950 per troy ounce and iridium near $4,700 per troy ounce, according to precious-metals market data. A single megawatt-scale PEM stack can require several grams of each metal, and at terawatt-scale deployment the cumulative demand would dwarf current mining output. Iridium is especially problematic: global production is only about 7 to 8 metric tons per year, nearly all of it as a byproduct of platinum mining in South Africa.
Anion exchange membrane water electrolysis (AEMWE) offers a workaround. It operates in alkaline conditions that allow cheaper, earth-abundant metals like nickel, iron, and cobalt to serve as catalysts. The catch has always been durability: early AEMWE prototypes degraded quickly, with membranes breaking down and catalysts losing activity after just tens or hundreds of hours. The DOE’s technical targets for AEM electrolyzers call for operation at current densities of 1 A per square centimeter or higher, with degradation rates low enough to support multi-year lifetimes. Until recently, no platinum-free system had come close.
Three lines of evidence, one conclusion
The strongest durability data come from a Nature Communications study on oxide-hybridized carbon catalyst supports. Researchers reported a degradation rate of roughly 0.17 millivolts per hour over 900 hours at 1 A per square centimeter. To put that in perspective, a rate that low, if sustained, would translate to less than 1.5 volts of total voltage increase over a 10,000-hour operating life, a range that falls within commercially viable territory. The same device reached a peak current density of 8.5 A per square centimeter at 2 volts, a high-rate benchmark that few non-precious-metal systems have matched in the published literature.
A second line of work, published in Nature Catalysis, took a different approach. Researchers used a seed-assisted growth method to build structured nickel-iron (NiFe) anode materials directly onto the membrane electrode assembly. According to a research briefing summarizing the effort, the NiFe anode sustained industrial-scale current density with stable performance over a test period described as 21 months. The briefing explicitly benchmarks the results against DOE alkaline and AEM electrolysis targets, placing the NiFe system within the government’s performance window for next-generation devices.
A third thread comes from a review paper in the journal Research, which synthesizes results from multiple groups. That review highlights more than 1,000 hours of stability at 1,000 milliamps per square centimeter for certain membrane electrode assembly designs. One configuration, described as a “3D interlocked interface strategy,” reportedly achieved 1,800 hours of stability at 1.0 A per square centimeter under pure-water operation, a condition far more demanding than the concentrated potassium hydroxide (KOH) solutions many labs use to boost their numbers.
The pure-water distinction matters. Running an electrolyzer on deionized water forces the membrane and electrode materials to handle ion transport with almost no help from the electrolyte. A separate Nature Communications study on local alkalinity effects reported reaching 3.0 A per square centimeter at 2.08 volts under pure-water feed, a result that approaches the current densities of platinum-based PEM systems under comparable conditions.
Open data and the replication question
Tohoku University researchers working on related AEM electrolyzer improvements have deposited experimental and computational data in the Digital Catalysis Platform, an open repository designed to let independent teams verify and build on published findings. That step matters in a field where durability claims can hinge on subtle differences in test conditions: cell temperature, membrane hydration, electrolyte purity, and whether the system ran continuously or was cycled on and off. Open data makes it harder for headline numbers to survive without scrutiny.
Replication by outside groups remains the critical next hurdle. All of the results described here come from the laboratories that developed the catalysts. Independent confirmation on different test stands, and eventually in multi-cell stacks rather than single-cell setups, will determine whether these numbers hold up outside controlled research environments.
Where the gaps are
The 21-month stability claim for the NiFe anode system is the most striking figure in the current literature, but it needs careful reading. The underlying Nature Catalysis paper is peer-reviewed, but the secondary summaries available for this analysis do not fully specify the operating profile behind that duration. Whether the cell ran continuously or in cycles, what temperature and pressure it maintained, and whether any membrane replacements occurred during the test are not detailed in the briefing. DOE durability targets are typically framed in terms of continuous operation at defined current densities. A system that logs 21 months of calendar time under intermittent use may experience far fewer total operating hours than a continuous test and may encounter different degradation pathways.
The 1,800-hour pure-water result from the Research journal review carries a similar caveat. Review articles synthesize many experiments and sometimes compress complex data into a single representative number. Until the primary paper behind the “3D interlocked interface” test is clearly identified and its protocol publicly accessible, that figure should be treated as promising rather than settled.
Cost projections remain the largest gap. None of the studies discussed here provide a full techno-economic assessment comparing platinum-free AEMWE systems to today’s commercial PEM electrolyzers at scale. Eliminating platinum and iridium from catalyst layers should reduce stack material costs, but stack hardware is only one component of total system expense. Membrane materials, bipolar plates, power electronics, gas purification, and water treatment all contribute meaningfully to the installed cost of a hydrogen plant. And if a platinum-free system requires higher cell voltage to reach the same current density as a PEM stack, the extra electricity consumption can erode the savings from cheaper catalysts. Industry estimates for PEM electrolyzer stack costs in 2025 ranged broadly from roughly $400 to $700 per kilowatt of capacity, with catalyst metals representing an estimated 10 to 20 percent of that figure depending on loading and design. Removing that precious-metal fraction would be meaningful but would not, on its own, halve the price of a complete system.
There is also ambiguity around which specific DOE targets are being cited. Federal benchmarks for hydrogen technologies evolve as costs fall and performance improves. A cell that meets an earlier set of targets might fall short of more recent guidance. Until the precise DOE documents used as reference points are identified, it is safest to say the reported devices appear consistent with at least one published set of DOE goals rather than definitively meeting the latest standards.
What the hydrogen industry is watching next
Green hydrogen matters because it is one of the few plausible routes to decarbonize sectors that electricity alone cannot easily reach: steelmaking, ammonia production, long-haul shipping, and seasonal energy storage. The International Energy Agency’s 2023 Global Hydrogen Review estimated that reaching net-zero emissions by 2050 would require roughly 150 to 200 million metric tons of low-emission hydrogen per year, up from less than 1 million metric tons in 2022. Scaling electrolyzer manufacturing to meet that demand with platinum-dependent technology would strain already tight supply chains for precious metals.
The convergence of multiple independent groups on similar performance levels for platinum-free systems is a meaningful signal, even with the gaps in cost analysis and some ambiguity around test protocols. The core technical barrier, demonstrating durable, high-rate operation without platinum-group metals, appears to be weakening. That does not guarantee rapid commercialization. Questions about membrane longevity under real-world cycling, stack engineering at megawatt scale, and full life-cycle cost remain open.
For policymakers and investors tracking green hydrogen pathways in 2026, the practical next markers to watch are replicated results across independent laboratories, stack-scale demonstrations with dozens or hundreds of cells rather than one, and techno-economic studies that fold these new catalysts into complete system models. Until those arrive, the latest AEMWE research represents a strong technical signal that a platinum-free route is viable in the lab. Proving it in the factory is the next test.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
