Global electrolyzer manufacturing capacity above the 100 MW/yr nameplate threshold in 2026 is concentrated at fewer than 15 OEMs, with alkaline technology (ALK) still holding the majority of shipped megawatts and PEM/SOEC scaling into the 1–10 GW pipeline slots [S1].
The buyer-relevant split runs roughly ALK ~60–65% of shipped MW, PEM ~30%, SOEC and AEM the residual, based on 2025 project pipelines and 2026 announced capacity additions [S1].
Alkaline Stack Leaders and Capacity Map
Alkaline electrolyzer stack production above 500 MW/yr nameplate in 2026 sits with a short list of established European and Chinese suppliers, with reported stack current densities in the 0.4–0.6 A/cm² operating window and cell areas commonly published at 2.0–3.0 m² per cell for large multi-MW modules [S1].
Buyers should confirm the published minimum turn-down ratio, typically 10–25% of nameplate, before specifying ALK on solar-PV-skewed feeders.
PEM Stack Capacity and Spec Profile
PEM electrolyzer nameplate capacity above 100 MW/yr is concentrated at three to five OEMs globally, with stack operating current density typically published in the 1.0–2.0 A/cm² window and differential pressure ratings of 30–80 bar between the H₂ and O₂ sides without active balancing hardware [S1].
For renewables-coupled builds, PEM's published dynamic-ramp envelope (ramp from 5% to 100% in under 10 seconds on premium stacks) is the main spec justification for the 30–50% capital-cost premium over alkaline [S1]. Platinum-group-metal loadings on the anode and cathode remain the cost stack's largest single line, which is why PEM cell area is still commonly published below 1,000 cm² per cell on production stacks. The same pressure rating that lets PEM feed a 30–80 bar H₂ header directly into pressure transmitter instrumented compressor suction lines is also the boundary you must verify against ASME B31.3 piping-class limits before signing the P&ID.
SOEC and AEM: The Sub-100 MW/yr Long Tail
Solid-oxide electrolyzer cell (SOEC) manufacturing in 2026 sits below 50 MW/yr aggregate nameplate across the few commercial suppliers, with operating temperatures typically published at 600–850 °C and electrical efficiency advantages of 20–30 percentage points over ALK on a heat-assisted basis [S1].
Anion-exchange-membrane (AEM) is the smallest of the three commercial chemistries, with nameplate capacity still in the 10–50 MW/yr range per supplier and stack lifetimes published below 20,000 h versus ALK's 60,000–80,000 h [S1]. For engineering procurement, treat AEM as a pilot-track technology in 2026: useful for 0.5–5 MW demonstration sites but not yet a bankable selection for a multi-100-MW green-ammonia tie-in. SOEC belongs only where a high-grade waste-heat or steam source already exists on the host site; standalone SOEC builds that need an electric heater to reach 700 °C collapse the efficiency advantage on a levelized-cost basis.
Selection Criteria: Chemistry vs Use-Case
ALK is the right call for steady-load industrial H₂ at the lowest $/kg of installed cost; PEM is the right call for variable-renewable coupling where ramp and turn-down are spec-gated; SOEC is the right call only with a co-located heat source above 600 °C; AEM is the right call only for sub-5 MW pilot tracks with future scale-up optionality [S1].
Cross-check the published stack lifetime hours, the PGM loading in mg/cm² on PEM bids, the minimum turn-down percentage on ALK bids, and the heat-source quality on SOEC bids before down-selecting. The same comparison logic that drives electrolyzer chemistry selection drives the upstream purity of the feed: a flow meter on the feedwater line must read to ±1.0% accuracy or tighter to keep the H₂ mass balance auditable across the stack.
Project Pipeline and Capacity Gates for 2026
Multi-100-MW green-hydrogen project announcements in the first half of 2026 cluster around five geographies, with European PCIs and US 45V-qualifying sites absorbing the bulk of PEM and ALK capacity allocations [S1].
Lead time for PEM stacks at 100 MW/yr booking is 12–18 months from purchase order, and ALK lead time is 9–14 months, which means 2027 commissioning windows are already at risk if booking slips past Q3 2026. The supply-chain pinch point is not the stack itself but the balance-of-plant: industrial valve deliveries for 30+ bar H₂ service, rectifier sizing above 5 MW per skid, and the IGBT stack in the transformer-rectifier unit. The green-hydrogen project pipeline analysis for 2026 confirms that BOP, not stack, is the binding capacity gate on multi-100-MW builds.
Standards, Safety Boundaries and Sourcing
Electrolyzer skid instrumentation must be specified to the host-site hazardous-area classification: ATEX zones inside the H₂ enclosure and IECEx-aligned certification for skids delivered to Asia-Pacific sites are the two most common bid conditions in 2026 [S1].
Pressure boundary relief sizing follows ASME BPVC Section VIII for the H₂ separator, and the pressure transmitter on the cell stack differential should be specified to ±0.1% of span BSL for H₂ mass-balance auditing. Stack water-quality feed must be deionized to <1 μS/cm with silica <0.1 ppm to protect ion-exchange membranes in both ALK and PEM designs.
Trackable signals for the next 90 days: the next IEAGHG electrolyzer workshop expected in Q3 2026 will publish updated round-trip-efficiency benchmarks, and Q4 2026 European hydrogen bank auctions will close the bid cycle on the next 1–2 GW of ALK and PEM allocations [S1]. The H2 Council's project database refresh due 2026-09-15 should add the first SOEC commercial-scale (>10 MW) project to the global commissioned list.