Comau positions itself as a turnkey integrator for electrolyzer and fuel-cell plants, with its MATE-XR exoskeleton and a hydrogen-specific pallet of welding, adhesive dispensing and stack-assembly cells listed as production-ready on the company's hydrogen automation page [S2].
Stack makers in 2026 split cleanly into two output bands: 1-10 kW mobility and portable power (HyZero, 1-10 kW) and 100 W to multi-kW aerospace/UAV/drone platforms (Intelligent Energy IE-SOAR 800, IE-SOAR 1.2, IE-SOAR 2.4, IE-DRIVE 100) [S3][S4].
Defining Hydrogen Smart Manufacturing: Scope and Cell Boundaries
A hydrogen fuel-cell "smart" line is conventionally scoped from Membrane Electrode Assembly (MEA) lamination through bipolar-plate stacking, cell compression torquing, EOL leak test and dispatch — Comau describes its offering as covering the full bipolar-plate-to-stack flow plus integrated testing [S2]. For electrolyzers (PEM and alkaline), the equivalent scope runs from electrode coating and frame welding through stack compression, balance-of-plant skid assembly and H₂ purity verification [S2].
Cell-boundary hardware is dominated by industrial robots with 6-220 kg payloads handling plate loading, laser welding of metallic plates and adhesive/gasket dispensing. The reference Chinese event for the sector — the 10th International Hydrogen Energy & Fuel Cell Technology and Product Expo in Foshan — booked 20,000+ m² of exhibition floor and 350+ pre-invited enterprises for its 2026-10-21 to 2026-10-23 window, signalling the scale at which Chinese stack builders are now buying capital equipment [S5]. In 2026, vision-guided plate-picking cells and trace-and-trace MES (per-piece laser marking on each plate) are treated as table stakes rather than optional add-ons by Tier-1 integrators.
Selection Criteria: What Buyers Actually Specify
For a related reference on reuse of brownfield process tools and AI-sensing retrofit patterns, see this industrial robot manufacturing process map — the same cell-control templates are being cloned from EV body shops into fuel-cell lines. Buyers in 2026 also demand deterministic cycle time, not just average throughput, because the heat-seal and compression stages of MEA lamination are sensitive to dwell-time jitter of more than ±2 %.
Who It's For — and Who Should Not Spec It
This automation stack fits three buyer profiles: (a) Tier-1 stack makers scaling 50 MW+/yr PEM or alkaline electrolyzer capacity, (b) automotive OEMs and contract manufacturers building fuel-cell powertrain lines, and (c) Tier-2 component suppliers running ≥ 50,000 plate units per year. Intelligent Energy's product line — IE-SOAR for UAVs, IE-DRIVE 100 for light automotive, IE-LIFT and IE-POWER for stationary and material handling — covers the typical output bands from 100 W class through multi-kW [S3].
It is NOT a fit for: (i) research labs doing fewer than 500 stacks per year (manual or semi-auto bench stations are cheaper and faster to reconfigure), (ii) producers targeting < 1 kW drone or eVTOL markets where the BOM dominates and a 1 MW robotic line cannot be amortized — HyZero's 1-10 kW mobility range is the practical floor for a dedicated stack-assembly cell [S4], and (iii) any program that cannot commit to a minimum 24-month production horizon, because robotic cell integration and tooling lead time on hydrogen lines is running 9-14 months as of 2026 [S2].
Main Options Compared: Robotic, Cylinder-Index and Linear-Flow
Three line architectures dominate 2026 hydrogen stack assembly. Option A — 6-axis robotic cells with rotary indexers: highest flexibility for mixed stack SKUs (passenger car vs heavy-duty), 6-15 stations per cell, cycle time 25-45 s, capex €8-15 M per 100 MW/yr equivalent [S2]. Option B — dedicated cylinder-index transfer lines: fastest at ≥ 60 ppm for one SKU, lowest per-piece labour, but changeover to a new plate geometry typically takes 4-8 hours, so this is the choice only for single-product, high-volume plants. Option C — linear-flow pallet systems with AGV/AMR handoff between cells: best brownfield fit because each cell can be added or replaced without stopping the line, but requires stricter MES orchestration and a unified pallet format across the factory.
Real Use Cases and Reference Programmes
Comau's published hydrogen automation portfolio lists three live programme families in 2026: bipolar-plate laser welding cells, stack compression and torquing stations, and EOL test benches integrating both H₂-snoop leak detection at 5-10 bar and high-pot insulation testing [S2]. Intelligent Energy operates the Chelveston Test Facility in the UK as a development and endurance-test site, with IE-SOAR and IE-DRIVE products documented for automotive, aviation/eVTOL, data-centre backup, and material handling deployments [S3].
HyZero positions itself as a 1-10 kW complete-system supplier for small mobility vehicles and portable power, with refillable cartridge or low-pressure storage modules integrated into the pack [S4]. For comparison with how discrete-power-electronics buyers weigh supplier tiers, see this IGBT supplier tier map — the IGBT content inside a 100 kW+ fuel-cell inverter is now a parallel make-or-buy decision for stack makers.
Limitations, Failure Modes and Standards Watch
The dominant failure mode on 2026 hydrogen lines is leak-test false-reject: residual coolant or NMP from MEA coating migrates into the leak-test fixture and triggers a false fail rate of 0.5-2 %, which at 30 s cycle time is the line's throughput bottleneck [S2]. Mitigations are forced-air purge stations and helium-sniffer pre-screens, both of which add 4-8 s per part. The second failure mode is bipolar-plate surface contamination from glove-oil in manual-loading fallback mode; a well-engineered line holds that mode to < 3 % of cycles and never runs unattended.
Standards governing the EOL and safety envelope include ISO 19880-1 for hydrogen fuelling stations, IEC 60079 series for explosive atmospheres on the welding and leak-test cells, and ATEX 2014/34/EU for equipment placed on the European market. On the OEM platform side, the BMW 5 Series GT Hydrogen Fuel Cell variant documented in the Autohome spec database remains a reference powertrain for the passenger-FCEV conversion community even though production volumes are low [S1]. Reference architectures for control and sensor retrofits in adjacent process industries are detailed in this nuclear I&C and AI-sensing roundup, and the perception-and-actuation logic of vision-guided stack picking maps onto the sensor and camera choices covered in this smart camera encyclopaedia entry.
Capex Bands, Cost-Down Levers and 2026 Sourcing
Published reference pricing for a 100 MW/yr PEM stack-assembly cell sits in the €10-18 M band (equipment only) in 2026, with stack compression and EOL test together representing 25-35 % of the cell cost [S2].
Chinese supply remains dominant for cell frames, end plates and graphite/metal bipolar plates; the Foshan expo in October 2026 is the largest concentration of hydrogen-specific Tier-2 suppliers in Asia, with 350+ pre-invited enterprises and a 20,000+ m² floor footprint [S5]. Western buyers increasingly dual-source critical SKUs (PTL, GDL, MEA) to keep stack cost in the $40-50/kW range targeted by most 2026 passenger-FCEV programmes [S2].
Sourcing Checklist and Trackable 2026 Signals
Track these three signals to keep a hydrogen line spec current through the rest of 2026: (a) Comau MATE-XR and hydrogen-cell order-book disclosures on the Comau hydrogen automation page, which is the public reference for line architectures and cycle-time benchmarks [S2]; (b) Intelligent Energy product roadmap updates on the IE-SOAR and IE-DRIVE families, since each generation change typically triggers a 3-6 month supply-chain requalification at Tier-1 customers [S3]; and (c) the Foshan CHFE 2026-10-21 to 2026-10-23 floor map and exhibitor list, which is the single highest-density source of new Chinese Tier-2 component suppliers in the 1-10 kW stack range [S4][S5]. Buyers should anchor any 2026 capex memo to a dated point-of-record on those three pages before releasing RFQs.
For component-level specifications, see additive manufacturing material, and oxy fuel cutter.