Global Portable Hydrogen Fuel Cell Market Report, Size, Share and Forecast 2026–2033: The Stack Got Cheap; the Cartridge Did Not
The portable hydrogen fuel cell market was worth USD 218.0 million in 2025 and reaches USD 1,046.0 million in 2033, a 21.66% CAGR. The value headline conceals the more important series: units grow from 21,500 to 149,800 at 27.46% while revenue per system falls 4.56% a year from USD 10,140 to USD 6,983. This is a market where volume compounds much faster than money.
The reason is that the stack has commoditised and the fuel has not. Chinese capacity set the price of membranes, catalysts, plates and balance-of-plant, so a PEM system is now a manufactured object. What a customer actually has to buy is the hydrogen, and hydrogen travels inside a regulatory envelope — 150 litres and 25 MPa for a metal hydride system under 49 CFR 173.311 and ISO 16111, requalified every five years, certified under IEC 62282-6-101:2024, which replaced its 2010 predecessor in 2024 and covers cartridges explicitly. That envelope is a barrier to entry for a new supplier and a moat for an incumbent, and it is why the surviving vendors sell supply contracts rather than boxes.
Demand is institutional, not consumer. The consumer electronics thesis died with MTI Micro in 2012 and Toshiba’s Dynario before it. Uncrewed aircraft are now 32.6% of the market in 2025, rising to 41.8% by 2033, and that forecast rests less on fuel cell engineering than on the FAA’s Part 108 beyond-visual-line-of-sight rulemaking, whose projected final rule is 2026. Defence and security hold a steady 27%; industrial, telecom and critical-infrastructure backup slip from 21.6% to 20.4%; outdoor recreation and construction, the closest thing to a consumer purchase, shrink to a combined 6.2%.
Concentration is real but not measurable to a share: one supplier with EUR 143.274 million of group sales and 90,000 fuel cells delivered cumulatively, a group of specialists with differentiated platforms and thin balance sheets, a Korean conglomerate subsidiary with the only integrated drone-plus-fuel offer, and a long tail of Asian system builders. Pheonix Research does not publish a vendor share split, because the disclosure does not support one.
A fuel cell is only portable if its fuel is
The portable hydrogen fuel cell has spent thirty years being described as imminent. It is now, quietly, a real market — but not the one the pitch decks promised. It did not replace the lithium-ion battery in a laptop or a phone. Toshiba's Dynario shipped in Japan in 2009 and vanished; MTI Micro suspended its micro fuel cell division in March 2012. The market that survived does not sell to consumers at all. It sells to militaries, drone operators, telecom tower owners, pipeline inspectors and emergency services — buyers who are not choosing between a fuel cell and a battery, but between a fuel cell and a diesel run, a generator, or a mission that does not happen.
Pheonix Research models that market at USD 218.0 million in 2025, rising to USD 1,046.0 million in 2033 — a 21.66% CAGR across the eight periods from 2025 to 2033. The headline is not the interesting part. The split inside it is. 21,500 systems in 2025 become 149,800 in 2033, a 27.46% CAGR, while revenue per unit falls from USD 10,140 to USD 6,983 — 4.56% a year. Units grow sevenfold; dollars grow 4.8-fold. Anyone modelling this market on a value CAGR alone will overstate the price umbrella and understate the volume, and will misread where the money actually is.
What the number covers
The scope is PEM fuel cell systems rated roughly 100 W to 10 kW that a person or a small vehicle can carry, together with the hydrogen storage and cartridges sold with them. It excludes vehicle traction stacks, stationary combined heat and power, and anything bolted to a foundation. It also excludes direct methanol fuel cells — the technology that dominates the off-grid market today through SFC Energy's EFOY line — except where a supplier's disclosures are used as a scale reference. That exclusion matters: methanol systems are the incumbent this market competes against, not a slice of it.

| Metric | 2025 | 2033 | CAGR |
|---|---|---|---|
| Market value, USD million | 218.0 | 1,046.0 | 21.66% |
| Systems shipped, units | 21,500 | 149,800 | 27.46% |
| Revenue per system, USD | 10,140 | 6,983 | −4.56% |
| Value per unit implied by scope: system plus carriage, not fuel consumed | — | — | — |
The constraint moved from the stack to the canister
A PEM stack is now a manufactured commodity. That is the single most important change in this market since 2020, and it is not a technology story — it is a capacity story. China's electrolyser builders and their fuel cell counterparts scaled the same membrane, catalyst and bipolar-plate supply chain, and the effect is visible in the price gap the IEA reports upstream of this market: electrolysers cost USD 2,000–2,600 per kW outside China and USD 600–1,200 per kW inside it, for the same duty. A portable fuel cell stack bought in Shenzhen is a line item. A portable hydrogen fuel cell product is not, because the product is the fuel.
Hydrogen travels with a portable system in one of two ways: compressed in a small cylinder, or bound into a metal hydride canister. Both are regulated as transport of dangerous goods before they are regulated as consumer products. A metal hydride storage system is capped at 150 litres water capacity and 25 MPa under 49 CFR 173.311 and its ISO 16111 counterpart, and requires periodic requalification — a five-year cycle that sits on the customer's asset register, not the vendor's. The safety standard governing the system and its cartridge was replaced in 2024: IEC 62282-6-101:2024 supersedes IEC 62282-6-100:2010 and, unlike its predecessor's framing, treats "fuel cartridges" as part of the object it certifies. Every one of those clauses is a certification cycle a lithium-ion pack does not have. Every one is also a moat, because the vendor who certifies the cartridge is the vendor who sells the cartridge.
The commercial consequence is visible in how the survivors describe themselves. Doosan Mobility Innovation sells a hydrogen supply system — 7 to 10.8 litre vessels and its own refilling equipment — as a product line standing beside the DP30 air-cooled powerpack, not as an accessory. SFC Energy lists fuel cartridges among the things it sells, in a portfolio it describes as both direct methanol and PEM. Neither company is in the business of selling a box that runs on a fuel the customer sources elsewhere. The customer does not buy a fuel cell and then look for hydrogen; they buy a supply contract with a warranty attached.
The buyer is an institution, and the purchase is a programme
This is why the market's demand profile looks nothing like a consumer electronics curve. Institutional buyers do not adopt; they qualify, trial, and then procure in lots. That produces a specific financial signature, and SFC Energy's 2025 accounts are the clearest example available: group sales of EUR 143.274 million, essentially flat at −1.0%, with the fuel cell segment at EUR 99.835 million and 69.7% of the group — but order intake down from EUR 167.762 million to EUR 118.427 million, with Germany described as stable and North America and India weak, the India softness attributed to a customer postponement. Flat sales on a 29% order-intake decline is what a programme-driven book looks like when one buyer moves a delivery by two quarters. The company also reports more than 90,000 fuel cells delivered cumulatively since 2003, 16.6 million kWh generated and 179 million operating hours — the numbers of an installed base that is maintained, not a product that is retailed.
The same signature appears in the failure case. Advent Technologies, the high-temperature PEM specialist, reported revenue of USD 0.293 million for the nine months to September 2025 against USD 3.52 million a year earlier, with USD 0.4 million of cash on the balance sheet and a going-concern qualification in its own filing. High-temperature PEM is a genuinely differentiated technology with real defence and aerospace applications. It went from a listing to a liquidity question in three years because the programmes it was waiting on did not convert at the rate its cost base required. That is the risk this market carries: it is not that the technology fails, it is that procurement cycles are longer than cash runways.

Drones are the only application growing faster than the market
Uncrewed aircraft are the reason this market re-accelerates, and they are the one segment whose growth is driven by a regulatory event rather than an energy-price event. A fuel cell does not make a drone better at hovering; it makes a drone able to stay up. Doosan's DJ25 VTOL is rated at 330 minutes of flight with a 4 kg payload, and its DT30X at 150 minutes with 5 kg — endurance numbers a battery multicopter cannot reach at those payloads. In November 2025 Intelligent Energy powered the UK's first hydrogen-fuelled beyond-visual-line-of-sight flight, a 25 kg six-blade hexacopter carrying its IE-SOAR system from Llanbedr in Wales into Eryri National Park and up to 10 km offshore, coordinated by BT with uAvionix and Skyfarer, with flight times quoted up to three hours.
But the flights above are trials, and trials do not scale a market. What scales it is the rule that lets a drone operator fly without a visual observer for the whole mission. The US Federal Aviation Administration's Part 108 rulemaking — formally "Normalizing Unmanned Aircraft Systems Beyond Visual Line of Sight Operations" — is the gate, and its projected final-rule date is 2026. Until it lands, endurance is a capability looking for a business model: an operator who must keep a visual observer within sight can use a 30-minute battery as easily as a three-hour fuel cell for most missions. Pheonix Research models uncrewed systems as USD 71.0 million of the USD 218.0 million 2025 market, 32.6%, rising to USD 437.0 million and 41.8% by 2033. That forecast is not a bet on fuel cell engineering. It is a bet on a docket in Washington.
| Application | 2025 | Share | 2033 | Share |
|---|---|---|---|---|
| Uncrewed aircraft and drone propulsion | 71.0 | 32.6% | 437.0 | 41.8% |
| Defence and security: dismounted, sensor and comms power | 59.0 | 27.1% | 289.0 | 27.6% |
| Industrial, telecom and critical-infrastructure backup | 47.0 | 21.6% | 213.0 | 20.4% |
| Emergency response and humanitarian | 14.0 | 6.4% | 43.0 | 4.1% |
| Outdoor recreation and marine auxiliary | 16.0 | 7.3% | 34.0 | 3.3% |
| Construction and portable tools | 11.0 | 5.0% | 30.0 | 2.9% |
| Total | 218.0 | 100% | 1,046.0 | 100% |
Note what does not grow. Outdoor recreation and marine auxiliary falls from 7.3% to 3.3% of the market; construction and portable tools from 5.0% to 2.9%. Those are the segments closest to a consumer purchase, and they are the ones where a battery, a small petrol generator or a methanol cartridge wins on price per kilowatt-hour. The market's growth is entirely concentrated in buyers who are paying for endurance or for the absence of a diesel supply line. That is a narrower, harder market than "portable power" — and it is the one that exists.

Geography: manufactured in Asia, bought in three blocs
Asia-Pacific is both the largest demand pool and the manufacturing base, at USD 76.0 million in 2025, 34.9% of the market, and USD 381.0 million by 2033, 36.4%. Korea contributes through Doosan's military and industrial drone programmes and its stationary PEMFC business; China contributes through drone airframes, backup power and an export supply of stacks and balance-of-plant that has compressed system prices everywhere. Europe follows at USD 68.0 million on the strength of a single dominant supplier and a dense set of hydrogen and aviation-decarbonisation funding programmes. North America is smaller commercially at USD 57.0 million, because the largest buyer there is a single ministry and its procurement is lumpy rather than market-like.
| Region | 2025 | 2033 | What sets the pace |
|---|---|---|---|
| Asia-Pacific | 76.0 | 381.0 | Korean defence drone programmes, Chinese manufacturing scale |
| Europe | 68.0 | 322.0 | Standards-led procurement, aviation and maritime decarbonisation funding |
| North America | 57.0 | 291.0 | Defence procurement cycles and the Part 108 outcome |
| Rest of world | 17.0 | 52.0 | Off-grid telecom and humanitarian deployment |
| Total | 218.0 | 1,046.0 |
One number should temper any reading of regional demand as a proxy for hydrogen infrastructure. The world had 1,160 hydrogen refuelling stations in operation at the end of 2024, 125 of them added that year, spread across 45 countries: 294 in Europe, 748 in Asia, 89 in the United States. Portable fuel cell customers barely use them. Doosan's customers refill from DMI equipment; military users take delivered gas; telecom sites use cartridges swapped by a service visit. Station count matters to this market only as a signal of national commitment, not as distribution. The distribution channel for a portable hydrogen fuel cell is a purchase order.

Where the value sits
The competitive structure follows from the same logic. There is one supplier with a disclosed, nine-figure fuel cell business and an installed base large enough to describe in operating hours; a small group of specialists with differentiated platforms and thin balance sheets; a Korean conglomerate subsidiary with the only integrated drone-plus-fuel offer at scale; and a long tail of Chinese and Japanese system builders whose names rarely appear on a Western procurement shortlist. Nobody in that list has a share worth quoting with confidence, and Pheonix Research does not publish one — the disclosure is too thin to support it, and a fabricated split would be worse than an honest absence.
| Company | Base | Portable position | Disclosed scale |
|---|---|---|---|
| SFC Energy | Brunnthal, Germany | DMFC and PEM off-grid systems, cartridges, defence and telecom | EUR 143.274m group sales 2025; 90,000+ fuel cells delivered cumulatively |
| Intelligent Energy | Loughborough, UK | IE-SOAR 800 W to 2.4 kW UAV systems; telecom and automotive platforms | Private; 600 patents; UK's first hydrogen BVLOS flight, November 2025 |
| Doosan Mobility Innovation | Yongin, Korea | DP30 air-cooled powerpack, hydrogen vessels and refilling, DS30W and DJ25 drones | 78 mobility fuel cell units cumulative; DJ25 rated 330 min endurance |
| Horizon Fuel Cell Technologies | Singapore and China | Air-cooled stacks, UAV systems, education and small-portable kits | Private; the principal source of low-cost air-cooled portable stacks |
| Bramble Energy | Crawley, UK | Printed circuit board fuel cell stacks for low-volume portable formats | Private, venture-backed; pre-scale manufacturing |
| Advent Technologies | Boston and Patras | High-temperature PEM membrane electrode assemblies and systems | USD 0.293m revenue, nine months to September 2025; going-concern qualification |
| Ballard Power Systems | Burnaby, Canada | Stack supply and heavy-duty modules; portable not a core line | Listed; volume concentrated in bus and truck programmes |
| Plug Power | Latham, New York | Material handling and stationary; a hydrogen supply business in its own right | Listed; the only participant operating a large hydrogen production and delivery network |
| Cummins (Accelera) | Columbus, Indiana | PEM stacks through the former Hydrogenics business; mobile and stationary | Segment disclosure only; portable sits inside a much larger power portfolio |
| Toshiba | Tokyo, Japan | Direct methanol micro fuel cell for consumer devices; the Dynario, 2009 | Discontinued; included as the reference case for consumer failure |
What would have to be true
Three things carry the forecast, and each is falsifiable. The first is the US beyond-visual-line-of-sight rule reaching a final form in 2026 and being usable by operators at commercial scale; a further delay of two years removes roughly a third of the modelled 2033 uncrewed-systems value. The second is that the cartridge certification burden stays with the vendor rather than being standardised into a commodity swap — if a multi-vendor cartridge format emerges under IEC 62282-6-101, the consumable margin that currently funds these small companies disappears. The third is defence budgets continuing to fund soldier and sensor power as a programme line rather than folding it into general electrification.
Against that, the honest counterweight is the hydrogen itself. Global hydrogen demand reached almost 100 Mt in 2024, and low-emissions supply is still around 1% of it. A customer buying a portable hydrogen fuel cell today is, in the overwhelming majority of cases, running it on hydrogen that was reformed from natural gas. That does not weaken the case for the product — an industrial customer buying a fuel cell is buying uptime, not a carbon claim — but it does mean the market's environmental proposition is contingent on a hydrogen supply chain that is still being built, and on policy that is still being argued. The IEA's Global Hydrogen Review 2025 is the cleanest statement of how far that build-out still has to go.
Related Pheonix Research
The upstream supply this market depends on is analysed in the Global Hydrogen Electrolyser Market report, and the segment context sits on the Hydrogen & Electrolysers segment page. The competing technologies a portable fuel cell is measured against are covered in Global Stationary Energy Storage Systems and Global Advanced Energy Storage & Fuel Cells. For the diesel generator alternative this market displaces at remote sites, see the India Diesel Genset Market and the Middle East and Africa Hybrid Diesel Genset Market.
Articles on this market
- The Fuel Cell Was Ready. The Shelf Was Not. — why the consumer micro fuel cell died of refuelling logistics rather than electrochemistry, and what that pattern still predicts.
- Three Hours of Endurance, Thirty Minutes of Permission — why hydrogen drone endurance is gated by a rulemaking rather than a membrane, and who benefits from the delay.
Methodology
There is no primary statistic for this market. No statistical agency collects portable hydrogen fuel cell shipments, the customs codes that carry these systems are shared with other converters, and the largest participant discloses a segment that mixes methanol and PEM. Pheonix Research therefore builds the market bottom-up rather than adopting a top-down figure: disclosed revenue and unit counts from listed participants are reconciled against programme-level evidence — published product ratings, endurance specifications and named deployments — and then scaled by an explicit long-tail allowance for the private Chinese, Japanese and Korean system builders that do not disclose. The 2025 base of USD 218.0 million is a modelled central estimate with a plausible range of USD 180 million to USD 265 million; the 2033 value and the unit series are generated from the same build rather than extrapolated from the base. Every arithmetic step is computed deterministically, and the unrounded outputs are retained in the dataset so that the tables here reconcile exactly.
Sources
- SFC Energy AG, Annual Report 2025 — group and segment sales, order intake, installed base and cartridge portfolio.
- Advent Technologies Holdings, Form 10-Q for the quarter ended 30 September 2025 — revenue and going-concern disclosure.
- Doosan Mobility Innovation — powerpack ratings, hydrogen vessel capacities, drone endurance specifications.
- Intelligent Energy, "IE powers UK's first long-range hydrogen drone trial", 25 November 2025 — IE-SOAR ratings, flight profile and partners.
- IEA, Global Hydrogen Review 2025, executive summary — demand, low-emissions share and electrolyser cost ranges.
- IEC 62282-6-101:2024, Micro fuel cell power systems — Safety — General requirements.
- 49 CFR 173.311, Metal hydride storage systems.
- Ludwig-Bölkow-Systemtechnik, H2stations.org, 2025 press release — global station count and regional split.
- US Office of Information and Regulatory Affairs, RIN 2120-AL82 — FAA beyond-visual-line-of-sight rulemaking status.
- EE Power, "MTI Suspends Operations of MTI Micro Fuel Cell Division", March 2012 — the consumer micro fuel cell exit.
Table of Contents
Competitive Landscape
Four layers, none of them holding a share worth quoting
The competitive structure of the portable hydrogen fuel cell market is layered rather than tiered, and the layers do not compete on the same axis.
The disclosed incumbent. SFC Energy is the only participant publishing a fuel cell business of consequence: group sales of EUR 143.274 million in 2025, essentially flat at minus 1.0%, with the fuel cell segment at EUR 99.835 million, or 69.7% of the group, against order intake down from EUR 167.762 million to EUR 118.427 million. More than 90,000 fuel cells delivered cumulatively since 2003, 16.6 million kWh generated and 179 million operating hours describe an installed base rather than a retail product. Its portfolio spans direct methanol and PEM, which is precisely why its segment disclosure cannot be used to size the hydrogen slice of this market.
The specialists. Intelligent Energy in Loughborough holds 600 patents and sells IE-SOAR at 800 W to 2.4 kW into UAV, telecom and automotive duty, and in November 2025 powered the UK’s first hydrogen-fuelled beyond-visual-line-of-sight flight with BT, uAvionix and Skyfarer. Horizon Fuel Cell Technologies supplies low-cost air-cooled stacks and small systems from Singapore and China and is the price reference for the entry tier. Bramble Energy is pre-scale. Advent Technologies is in a liquidity event rather than a competitive one.
The integrated conglomerate. Doosan Mobility Innovation is the only participant offering drone airframes, fuel cell powerpacks, hydrogen vessels and refilling equipment as one package — DS30W at 120 minutes and 3 kg, DT30X at 150 minutes and 5 kg, DJ25 VTOL at 330 minutes and 4 kg. That integration is a structural advantage in defence procurement, where a programme office would rather qualify one supplier than four. It also caps growth, because a customer who wants a competitor’s airframe cannot buy half the package.
The adjacent giants. Ballard, Plug Power and Cummins’ Accelera business own the stacks, the balance sheets and, in Plug Power’s case, a hydrogen supply network. None of them treats portable as a core line, and all of them could enter at scale by acquisition rather than development. Their absence from the segment is a choice, and it is the reason the specialists still exist.
The pricing dynamic that follows is unusual. Chinese stack supply has pushed system prices down at roughly the rate this report models — 4.56% a year in revenue per unit — while certification and programme qualification have kept the number of qualified suppliers small. The result is a market that is simultaneously deflationary in hardware and protected in access. That combination is comfortable for whoever already holds the certification, and brutal for whoever is trying to buy their way in.
Pheonix Research does not publish a vendor market-share table for this market. No participant discloses portable hydrogen fuel cell revenue as a separate line, the largest discloses a methanol-and-PEM segment, and several material suppliers are private or consolidated inside conglomerates. A split constructed from those inputs would be an artefact of the assumptions rather than a measurement, and would misinform a procurement decision more than the absence does.
Value Chain
The stack is a commodity; the fuel is the product
The value chain for a portable hydrogen fuel cell has five links, and only two of them carry defensible margin.
Link one: materials and stack. Membrane electrode assemblies, platinum-group catalyst, gas diffusion media, bipolar plates and the balance-of-plant. This is where Chinese and Japanese capacity has changed the market. The IEA’s evidence from the adjacent electrolyser chain is the clearest available proxy for the direction of travel: USD 2,000 to 2,600 per kW outside China against USD 600 to 1,200 inside it, for equipment performing the same duty. A stack is now sourced, not developed.
Link two: system integration. Packaging a stack into something that survives a soldier’s rucksack, a drone’s vibration profile or a telecom cabinet’s temperature swing. Thermal management, air filtration in dusty environments, water management in a sealed enclosure, and a control board that starts reliably at minus 20C and does not fail at plus 50C. This link carries engineering margin but not pricing power, because there are enough integrators that the customer can shop.
Link three: hydrogen storage and cartridges. The compressed cylinder or metal hydride canister, its valve, its regulator and its refilling interface. Doosan sells 7 to 10.8 litre vessels and multi-purpose refilling equipment as a named product line beside its DP30 powerpack. SFC Energy lists fuel cartridges among the items it sells, in a portfolio it describes as both direct methanol and PEM. This is the link where a customer becomes recurring rather than transactional.
Link four: certification. Every cartridge crosses a jurisdictional boundary with a dangerous-goods classification attached. A metal hydride storage system is limited to 150 litres water capacity and 25 MPa under 49 CFR 173.311 and ISO 16111, and needs requalification — a five-year cycle carried on the customer’s asset register. IEC 62282-6-101:2024 replaced IEC 62282-6-100:2010 in February 2024 and, unlike its predecessor, certifies the fuel cartridge as part of the object. Small suppliers treat this as compliance cost. The ones that survive treat it as the barrier that keeps them in business.
Link five: channel and service. Procurement, not retail. A fuel cell reaches a defence customer through a programme office, a telecom customer through a site-service contract, and a drone operator through an integrator. Aftermarket — cartridge resupply, stack replacement, requalification support — is where revenue becomes annuity.
Two consequences follow. The first is that gross margin migrates downstream: a supplier selling only a stack competes with Shenzhen on price, while a supplier selling a certified cartridge plus resupply competes with nobody in particular. The second is that working capital sits in the wrong place for the small players: they carry hydrogen inventory, vessel inventory and certification cost against a customer base that orders in lots. Advent Technologies’ filing is the cautionary version — USD 0.293 million of revenue for nine months of 2025 against USD 3.52 million a year earlier, and USD 0.4 million of cash at the end of that quarter.
Investment Activity
Capital is going into capacity and certification, not into new entrants
Investment in this market is small, concentrated and largely defensive, and the pattern is more informative than the totals.
Incumbent reinvestment. SFC Energy’s 2025 position — EUR 143.274 million of sales, order intake down to EUR 118.427 million from EUR 167.762 million, weakness attributed to North America and to an India customer postponement — describes a company funding certification, cartridge supply and regional service coverage out of a flat top line. That is the modal capital allocation in this segment: sustaining spend on qualification and consumables rather than capacity expansion.
Distress rather than entry. Advent Technologies’ trajectory is the clearest signal of what happens when capital and procurement cycles diverge. Revenue of USD 0.293 million for the nine months to September 2025 against USD 3.52 million a year earlier, USD 0.4 million of cash and a going-concern qualification. A differentiated HT-PEM platform did not produce a fundable business because the defence and aerospace programmes it targeted did not convert on its cash timeline. Investors reading this market should treat that as the base rate for platform-only specialists.
Public money as the real venture capital. The capital that de-risks this technology is programme funding, not equity: defence power requirements, national hydrogen strategies, EU aviation and maritime decarbonisation calls, and the demonstration budgets behind BVLOS trials like the BT-coordinated UK flight. A supplier’s growth is often a function of which national programme adopted it, which is why regional revenue shares in this report look more like procurement maps than adoption curves.
Strategic absence. There has been no consolidating acquisition of scale in portable fuel cells. Ballard, Plug Power and Cummins’ Accelera all own the capability and none has bought a portable specialist, which tells you how they price the segment’s near-term cash flows. If that changes, it will most likely be triggered by a regulatory event that converts trials into fleets rather than by a technology milestone.
Where the next money goes. Three places, in order of conviction: cartridge manufacturing and refilling infrastructure, because that is where recurring revenue sits; certification and test capacity, because that is the binding constraint on new product introduction; and hybrid power electronics, because the fuel cell rarely sells without a battery and a controller beside it.
Technology & Innovation
Innovation has moved from power density to the logistics of refuelling
The engineering frontier in this market is no longer the stack. Air-cooled PEM platforms in the 1 to 3 kW class are a mature product category: Doosan’s DP30M2S is rated 1.3 to 2.6 kW, and Intelligent Energy’s IE-SOAR family spans 800 W to 2.4 kW across three ratings. Both are manufactured at scale, both are proven on real aircraft and real sites. Incremental improvement here is measured in hundreds of hours of durability and a few percentage points of efficiency, not in a step change.
What is still moving is everything around the stack.
Storage format. A compressed cylinder at 350 or 700 bar needs a high-pressure vessel, a regulator and a compressor to fill it. A metal hydride canister binds hydrogen at low pressure, which is a genuine safety advantage in a carried application and a genuine procurement advantage under 49 CFR 173.311 — but it is heavier per kilogram of hydrogen and it heats on charging and cools on discharge, which shapes how fast it can be refilled. The choice between the two is not a technology preference; it is a function of which regulation the customer can operate inside.
Cartridge interfaces. Today each vendor’s cartridge talks to its own system. The moment IEC 62282-6-101 certification becomes common enough that a second-source cartridge can be certified against a competitor’s system, the consumable margin that funds these companies compresses. That is the single largest commercial risk embedded in the technology trajectory, and it is a standards event rather than an engineering one.
High-temperature PEM. Advent Technologies’ HT-PEM membranes run at 160C to 200C, tolerate reformate with carbon monoxide in it, and need no humidification — which makes them attractive for military and aerospace duty where the fuel is whatever is available. It is a real technical differentiation. It has not translated into a business at the scale the cost base required, which is itself the finding: in this market, differentiation without a procurement win is a liability.
Printed circuit board stacks. Bramble Energy’s PCBFC approach replaces machined plates with printed circuit board processes, aiming at low-volume formats where tooling cost normally dominates. If it works at scale it lowers the minimum efficient size of a fuel cell business, which matters in a market where annual unit volumes are still measured in tens of thousands.
System intelligence. The least visible change, and commercially the most useful: hybrid control between a fuel cell and a battery buffer, plus remote telemetry. A fuel cell that reports its own degradation and schedules its own cartridge swap is a service product rather than a hardware sale, and service products are what institutional buyers actually procure.
Market Risk
Five things that can break this forecast
One: substitution. This is the highest-scoring risk in the model. A portable fuel cell competes with a lithium-ion pack that improves every year, with a diesel generator that costs less per kilowatt-hour, and with a direct methanol cartridge that is already certified and already sold by the incumbent. The fuel cell wins only where endurance or the absence of a fuel supply line is the binding constraint. Any application where a battery is good enough will revert to a battery, and the segments closest to a consumer purchase — outdoor recreation and construction — are modelled to shrink from 12.3% of the market in 2025 to 6.2% in 2033 for exactly that reason.
Two: programme concentration. Demand is a handful of procurement decisions. SFC Energy’s 2025 order intake fell 29% with one customer’s India delivery moved by a couple of quarters; that is not a demand signal, it is a calendar effect, and it moved a nine-figure business. A market this concentrated has lumpy reported growth that says nothing about its trajectory in either direction.
Three: the US BVLOS rule. Uncrewed systems carry 41.8% of modelled 2033 value. If Part 108 slips materially beyond 2026, or if it is finalised with operational limits that keep missions inside visual range in practice, roughly a third of the 2033 uncrewed-systems value goes away and the market’s growth rate falls by several points. This is the single largest identifiable swing factor, and it is a legal one.
Four: the hydrogen itself. Nearly all hydrogen sold today is produced from fossil feedstock, and low-emissions supply is around 1% of a roughly 100 Mt market. Where a customer’s justification for buying a fuel cell is environmental rather than operational — a corporate decarbonisation commitment, for instance — that justification is contingent on a supply chain being built under policy that is still in contention. A change to a production credit or a certification scheme can remove the rationale for a purchase without touching the product.
Five: cartridge commoditisation. The commercial model of this market depends on the consumable being vendor-specific. A multi-vendor cartridge standard certified under IEC 62282-6-101 would be good for customers and bad for the margin that currently funds certification, service networks and the smaller suppliers’ cash flow. This is the risk the industry is least willing to discuss and the one most likely to arrive as a customer demand rather than a regulatory requirement.
Two further risks are worth naming without ranking. Currency: costs are Asian and European, revenue is global, and a 10% move in the euro-dollar rate moves the reported size of this market by more than its annual unit growth. And disclosure: because the largest participant mixes methanol and PEM in one segment, and several material suppliers are private, the published evidence base is thin enough that a single company’s restatement can move the modelled market by several percentage points without any underlying demand changing.
Regulatory Landscape
This market is regulated as dangerous goods before it is regulated as a product
No other constraint shapes the portable hydrogen fuel cell market as much as the rule set that governs moving hydrogen from one place to another.
Transport of the fuel. A metal hydride storage system is regulated under 49 CFR 173.311 in the United States and ISO 16111 internationally, with a water capacity limit of 150 litres and a pressure limit of 25 MPa, and requires periodic requalification on a five-year cycle. Compressed hydrogen cylinders carry their own regime. In practice this means a cartridge has a service life, an inspection record and a disposal obligation, and a fleet operator has to administer all three. It is the operational reason vendor-supplied refilling equipment persists: the vendor administers the compliance.
Certification of the system. IEC 62282-6-101:2024 supersedes IEC 62282-6-100:2010 and covers micro fuel cell power systems and fuel cartridges that are wearable or easily carried by hand. The 2024 revision matters commercially: an existing product certified against the 2010 text faces a transition, and a transition is a cost that falls disproportionately on the smallest suppliers.
Aviation. Drones are the growth engine and aviation is the strictest gate. In the United States, the FAA’s Part 108 rulemaking — formally ‘Normalizing Unmanned Aircraft Systems Beyond Visual Line of Sight Operations’, tracked as RIN 2120-AL82 — is the instrument that converts endurance from a demonstration into a service. Its projected final-rule date is 2026. In the UK, the Civil Aviation Authority’s visual-line-of-sight limit is what made Intelligent Energy’s November 2025 BVLOS trial notable, because it required a command-and-control chain combining protected-spectrum radio, cellular and satellite links to satisfy the regulator as well as the aircraft. Until BVLOS authority is routine, a three-hour fuel cell competes with a 30-minute battery for missions the operator is not permitted to fly anyway.
Hydrogen policy. The US Section 45V clean hydrogen production credit reached final regulations in January 2025, retaining the three-pillar architecture linking eligibility to new supply, regional deliverability and hourly matching. The EU hydrogen bank and the Clean Hydrogen Partnership operate a different mechanism toward the same objective. Both matter to this market indirectly: a portable fuel cell runs on hydrogen it did not produce, and a customer buying one in a jurisdiction without low-emissions supply is buying a device that will run on reformed natural gas for the foreseeable future. The IEA put global hydrogen demand at almost 100 Mt in 2024 with low-emissions supply still around 1% of it, so the environmental case for this product is a forward claim rather than a current fact.
What is not regulated. There is no international standard for swapping a cartridge between vendors, and no customs code that isolates portable fuel cell systems from other converters. Both absences favour incumbents and penalise anyone trying to build a market-wide channel.
