November 2026 - Hydrogen's Fortnight in Review: Mid-Sized Plants Switch On, Mega-Projects Wobble, and Fuel Cells Find a Data-Centre Home
Across 37 developments in two weeks, hydrogen's centre of gravity moved from headline-grabbing pledges to buildable, bankable assets, and the deciding variable was demand, not technology.
One regulatory filing erased $16 billion of ambition in a single stroke. Elsewhere, a much smaller scheme was reported to be about to produce its first molecules. That contrast has defined the last fortnight in hydrogen news. The industry is sorting itself into projects with a buyer and projects with a brochure. Electrolysers — the machines that split water into hydrogen and oxygen — are being switched on across Europe. Pipelines are being approved and capacity booked. Fuel cells are finding commercial traction in sectors that have little to do with hydrogen cars. And where long-term offtake (a contracted buyer for the output) is missing, even very large projects are retreating.
Below we group the period's developments by theme, highlight what each cluster signals, and close with what we are watching next. Our reading is flagged as such; the facts are attributed to the published reports.
1. Production: the mid-sized plants are arriving, the giants are blinking
Europe's first wave reaches the switch-on stage
The most concrete green hydrogen production news came from Europe. Air Liquide said on its Q3 2026 earnings call on 29 October that its Normand'Hy project in Normandy is poised to start production imminently. The plant is designed for 28,000 tonnes a year of renewable and low-carbon hydrogen from a 200 MW PEM electrolyser supplied by Siemens Energy. (PEM, proton exchange membrane, is one of the main electrolyser technologies.) The French State is providing €190 million in support, and TotalEnergies has a long-term offtake agreement. The project is expected to cut CO2 emissions by an estimated 250,000 tonnes annually. ↗
In Germany, Energiepark Bad Lauchstädt has produced green hydrogen for the first time. It uses a 30 MW electrolysis plant powered by a nearby 50 MW wind farm, developed by Uniper, VNG Handel & Vertrieb and Sunfire. The output goes to TotalEnergies' refinery in Leuna through a repurposed natural gas pipeline operated by ONTRAS. Integrated commissioning is planned for 2027. ↗
In southern France, Qair and AREC Occitanie inaugurated Hyd'Occ at Port-la-Nouvelle. It has a 20 MW alkaline electrolyser rated at 2,700 tonnes a year and links to refuelling stations in Béziers and Narbonne, targeting heavy mobility and maritime uses. The reporting is candid about the weak point: success depends on converting planned customer agreements, including with Plug Power, into actual sales. ↗
Further up the pipeline, Scottish councillors granted planning permission for the 50 MW Creca Hydrogen Facility on the former Chapelcross nuclear site. It would serve local industrial users, but financing and operational details remain uncertain. ↗ In the US, Delaware's governor announced that Aternium has proposed a $380 million facility in Dover to produce electrolytic hydrogen and heavy water. The company has received a $1 million state investment and entry into the JobsFirst Permitting Accelerator, but the project is still in planning. ↗
The pattern is worth stating plainly. The assets reaching or nearing operation are 20 to 200 MW. They are tied to a named industrial customer or a regional cluster. In the French and German cases they sit inside state-supported frameworks. Normand'Hy has a long-term TotalEnergies offtake. Bad Lauchstädt has a refinery at the end of an existing pipe. Hyd'Occ, by contrast, is still waiting for its customer agreements to harden. Our reading: an offtaker with a pipeline connection is now a better predictor of delivery than an electrolyser order.
The gigaproject problem
Then there is Chile. TE H2 Chile, the joint venture between TotalEnergies and EREN Groupe, withdrew from environmental evaluation of the H2 Magallanes project. It was a $16 billion green hydrogen and ammonia scheme meant to use Patagonian wind to produce over 343,000 tonnes of hydrogen a year. The company cited difficulty securing long-term purchase agreements and equipment costs. ↗ The same report says Chile's commitment to becoming a green hydrogen leader remains intact, and it points to the need for further market maturation and technological advances.
Set that beside Sahamco Holding's Dakhla project in Morocco. It targets an initial one million tonnes a year of renewable ammonia, expected investment above 254 billion dirhams (over USD 25 billion), partners including NetZeroWorx and INEOS, and a Final Investment Decision (FID, the formal commitment to build) by 2027, with operations aimed at 2031. ↗ It is backed by the Moroccan government's 'Morocco Offer' framework.
Nothing in the Dakhla report suggests it faces the same issue as Magallanes, and we are not claiming it does. But the two sit on the same spectrum of scale and timeline, and Magallanes is a reminder of what the gap between announcement and FID can hold. The reason given there, buyers who will not sign long-term contracts at current equipment costs, is the central commercial question for every export-oriented project. Dakhla's 2027 FID window is where that question will be tested.
An export proof point from China
One project already answers the question in practice. China's Da'an Green Ammonia Project, backed by SPIC and Jilin Electric Power, renewed its ISCC EU RFNBO certification by passing its annual audit. (RFNBO means renewable fuels of non-biological origin, the EU's category for renewable hydrogen-derived fuels.) The project produces around 180,000 tonnes of ammonia a year from wind and solar power and has already exported 3,750 tonnes of green ammonia to South Korea. ↗ That is a modest volume against the capacity, but it shows that certified, traded green molecules already move between Asian economies. Separately, a policy overview reports that China accounts for 65% of installed electrolysis capacity, with Sinopec's Kuqa and Ordos facilities among its flagship projects. ↗ The same overview cites a goal of up to 200,000 tonnes of renewable hydrogen a year and a target of 50,000 fuel-cell vehicles by 2025. It does not say whether that vehicle target was met.
Offshore: useful lessons, not a business case
The PosHYdon pilot on the Q13a-A platform in the Netherlands confirmed that offshore hydrogen production can operate in a real marine environment. It used a 1 MW electrolyser, Emerson's DeltaV automation, and advanced filtration. The main lessons were about system integration and the need for robust seawater treatment. ↗ At 1 MW it is a learning exercise rather than a commercial template, and it should be read as one.
Our take: the production story this fortnight is not that hydrogen is getting cheaper. It is that projects with a contracted customer and a way to move the product are getting built, while projects relying on a future market are being re-examined.
2. Infrastructure: the network is being booked before the molecules exist
Hydrogen infrastructure produced the period's best evidence of demand intent. Germany's hydrogen core network operators secured over 6 GW in capacity reservations. The reservation process opened on 19 March 2026, and the Bundesnetzagentur-approved network is expected to span 9,040 kilometres, 60% of it converted natural gas pipelines. A standardised reservation fee is also being introduced. ↗ The report is careful to note that these are contractual commitments that signal demand but do not guarantee the associated production projects get built.
That caveat matters. A reservation with a fee attached is a more costly signal than a letter of intent, because it forces a would-be user to put money behind a planning assumption. Even so, it sits upstream of the harder questions: who produces, at what price, and who buys. Bad Lauchstädt illustrates the converted-pipeline logic at small scale, with a repurposed gas line carrying hydrogen to a refinery.
Across the border, the Luxembourg Government Council approved HY4Link Phase 1, a 40-kilometre pipeline between Thionville in France and Pétange in Luxembourg. Partners are Creos Luxembourg Hydrogen, NaTran and Fluxys Hydrogen. It forms part of the European Hydrogen Backbone initiative. Environmental assessments and route planning come before construction, and a Project of Common Interest designation from the European Commission may open further funding. ↗
Regional framing reinforced the point. At HyDéO Day in Toulouse on 5 November, organised by AD'OCC, participants discussed hydrogen as a geopolitical and energy-sovereignty issue, alongside technology dependencies and workforce development. ↗ Our reading: in Europe the sovereignty argument is now carrying as much weight as the climate argument, and it tends to attract more durable political money.
3. Heavy industry and power: where hydrogen's demand is being engineered
Steel
Steel is where hydrogen news turns from supply to demand. Nippon Steel completed more than 300 hours of testing with 100% hydrogen at its Hasaki R&D Center in Kamisu, Ibaraki, producing over 1,000 tonnes of direct reduced iron (DRI, iron made without a blast furnace) from lower-grade iron ore pellets. The test addressed the endothermic reaction (it absorbs heat) of hydrogen reduction and the powdering of lower-grade ore. The effort is supported by NEDO, and the company aims for commercial scale-up by 2040. ↗ That 2040 date is a useful reality check on how long this route takes.
In Australia, ARENA is granting AUD 15 million towards a AUD 30 million pilot by Hadean Energy in Hastings, Victoria. The pilot uses tubular solid oxide electrolysis developed by CSIRO to produce up to 7 kg of hydrogen an hour for a steel annealing process, where it could supply roughly 85% of the energy needed. It builds on earlier BlueScope Steel trials. ↗ The scale is small, but it targets a real industrial heat application.
Policy is trying to build the market around steel. The European Parliament's Environment Committee backed a proposal to cut the 2035 passenger-car CO2 reduction target from 100% to 87.5%. Manufacturers could offset up to 7.5% through credits linked to qualifying EU-made low-carbon steel. The proposal, led by Massimiliano Salini, would tie vehicle rules to industrial decarbonisation and create demand for processes such as hydrogen-based steelmaking, though negotiations with the Council are still required. ↗
Our reading is that this is a notable piece of demand engineering, and also a sign of how contested the underlying rules are. It rewards green steel through a mechanism outside the hydrogen sector, which is arguably smarter than another subsidy round. But it is a committee vote on a draft, not law.
Power generation
GE Vernova said its LM6000VELOX gas turbine ran on 100% hydrogen at its Global Technology Center in Greenville, South Carolina. It also validated the LM2500XPRESS package for blends up to 50% hydrogen. GE's commitment includes over $160 million in investment and job creation. ↗ In Lingen, Germany, RWE Generation and Kawasaki Heavy Industries began trial operation of a 34 MW hydrogen-compatible gas turbine at the Emsland gas-fired station. It can burn natural gas, hydrogen or any blend, and the trial will test scenarios up to 100% hydrogen. It is linked to the GET H2 Nukleus project and sits in the context of German policy requiring hydrogen-readiness for new gas plants. ↗
The signal: turbine makers are de-risking the machine, while the fuel supply question remains open. Hydrogen-ready is a design specification, not a fuel contract. The German requirement does create a regulated reason to build compatibility in. Whether any meaningful hydrogen actually flows depends on the infrastructure and pricing questions in sections 1 and 2.
4. Fuel cells: the money is in power for AI and defence, not in hydrogen cars
Data centres emerge as the surprise customer
The most commercially forceful fuel cell news this fortnight came from data centres. NTT DATA partnered with Bloom Energy to deploy 900 MW of onsite power generation across global data centres, using Bloom's solid-oxide fuel cells so capacity need not wait on grid connections. ↗ FuelCell Energy joined the StabilityMax project, led by the US Department of Energy's National Laboratory of the Rockies, to evaluate fuel cells, batteries and advanced controls on a Connecticut testbed for AI-driven load profiles. ↗ An investor note framed Bloom and Solaris Energy Infrastructure as beneficiaries of a net power shortfall that Morgan Stanley estimates rising from 33 GW to 38 GW between 2026 and 2028. ↗
A caution is in order. The summaries do not specify what fuel these systems run on, and none describes them as hydrogen-fuelled. They belong in a hydrogen digest because they pull fuel cell manufacturing, supply chains and investor attention into the sector, not because they prove hydrogen demand. We would not book any of the 900 MW as hydrogen consumption.
Still, the commercial logic is clear. Speed to power is the scarce asset for data centres, and fuel cell vendors that can offer it are being treated as infrastructure suppliers rather than as hydrogen-economy bets. For suppliers, that is a more reliable order book than hydrogen vehicles ever promised.
Defence and niche industrial demand
SFC Energy raised its 2026 earnings guidance on a EUR 42.7 million order for hybrid fuel-cell systems for military and civilian applications in Ukraine, aligned with the German government's Enablement Initiative. Adjusted EBITDA is now forecast at EUR 36 to 38 million, helped by demand from the semiconductor equipment industry. ↗ Defence and semiconductors pay for reliability and not for the lowest cost per kilowatt-hour. Our reading is that fuel cells are winning where the buyer is not price-led.
Transport: hybrid thinking, not hydrogen-only
Ballard Power Systems confirmed series production of its FCmove-SC fuel-cell engine in the Solaris Urbino 12 hydrogen transit bus. The engine is rated up to 80 kW with 57% peak efficiency, in a hybrid architecture where the fuel cell provides continuous power and a battery handles peaks. The report acknowledges competitive pressure from battery-electric buses. ↗ Ljubljana's LPP shows what that looks like in a fleet: of 35 new Mercedes-Benz buses, 25 are battery-electric eCitaro and ten are hydrogen fuel-cell eCitaro G, chosen by route needs as part of a 2030 zero-emission target. ↗
The ratio, 25 battery to ten hydrogen, says more than the press release does. For hydrogen vehicles in public transport, the realistic niche is the demanding routes, not the whole fleet.
In shipping, Hanwha Aerospace obtained type approval from the Korean Register for a 200 kW-class marine hydrogen fuel-cell system, described as a first in South Korea, with efficiency above 50% and a link to the IMO's 2050 net-zero goal. ↗ Type approval is a gate that must be passed before commercial deployments, and the report refers to future deployments rather than orders.
5. Natural hydrogen: exploration money is moving faster than evidence
Four reports in two weeks put natural (geological) hydrogen on the agenda. JERA invested in Koloma, a Denver-based company that has raised over US$400 million from investors including Amazon and Khosla Ventures. The report stresses that commercial viability still needs rigorous geological assessment. ↗ In Australia, Gold Hydrogen reported natural hydrogen flowing to the surface at its Ramsay project in South Australia, with concentrations up to 97% at Ramsay 3 and modelled production potential of 100 kg a day per well. It is collaborating with Mitsubishi Gas Chemical on exploring a green-methanol facility. ↗ Nearby, Prominence Energy reported that a CSIRO study found hydrogen in all 28 rock samples from the Eyre Peninsula. This does not confirm commercially recoverable resources, and multi-well drilling is planned for mid-2027. ↗ Mantle8 launched a 24-month exploration programme in Block 7, Ras Al Khaimah, with over 1,000 passive seismic sensors, targeting serpentinisation potential in the Semail Ophiolite. ↗
Some perspective helps. Normand'Hy is designed for 28,000 tonnes a year. On our own back-of-envelope arithmetic, that is roughly 77 tonnes a day, or several hundred wells at Ramsay's modelled 100 kg a day each. That is an illustration of scale, not a forecast, and the modelled figure is not a result. The point is that natural hydrogen is a credible option with real capital behind it, but it is still at the stage where finding out whether it works is the product. The usefulness for readers is as a watch-list: the mid-2027 drilling and the 24-month Ras Al Khaimah programme will produce the first hard data.
6. Emerging markets, finance and the R&D pipeline
Strategy documents versus signed contracts
Emerging-market activity this fortnight was heavy on frameworks. A commentary urged Colombia to look to South Korea's hydrogen economy and build the whole value chain, not only production. Colombia's roadmap aims for 3 GW of electrolyser capacity by 2030. Ecopetrol is involved, a Cartagena pilot illustrates localised production, and experts caution against copying Korean targets directly. ↗ Sonatrach signed an MoU on 3 November in Algiers with the IBRD and the IFC to explore reducing gas flaring and methane emissions while advancing green hydrogen and petrochemical initiatives. The MoU names no specific project or investment. ↗ The same applies to Asahi Kasei and Power to Hydrogen, whose MoU on commercialising anion exchange membrane (AEM) electrolysis includes no production commitments or investment details. ↗
Our reading: these are early-stage positioning moves, and the useful question for each is what converts them. For Algeria, the draw is a World Bank-linked path to private investment. For AEM, a lower-cost electrolyser alternative, it is whether anyone commits to manufacturing.
The listed-company reality check
Plug Power was heading into Q3 2026 earnings on 9 November. Analysts' consensus revenue was around $198.36 million, over 10% growth year on year, and the company is still dealing with net losses and negative cash flow under CEO Jose Luis Crespo. ↗ That figure is a consensus, not guidance, and we have not seen the result. Plug also appears in the Hyd'Occ story as a prospective customer, so its balance sheet and the French project's demand are indirectly linked.
Technology and storage research
The R&D items are small and early, but they show where the technology frontier is. DGIST and Kookmin University developed a TiO2–AgBiS2 photoelectrode for solar hydrogen, reaching 3.0 mA/cm² at 0.6 V, with durability and scalability unresolved. ↗ UCLA and Mote are piloting biomass gasification of roughly 7 tons a day of campus waste into hydrogen, with CO2 intended for geological storage. ↗ On hydrogen storage, Mirian Elizabeth Casco of Universidad Católica del Uruguay won a national L'Oréal-UNESCO award for work on solid-state storage through gas hydrates in biomass-derived carbon, funded with US$22,000. ↗ None of this changes the market this year. It does show that storage and low-energy production routes are still being worked on at university scale.
The through-line: demand certainty is the scarce resource
Taken together, the fortnight shows three things.
- Delivery tracks offtake, not ambition. Normand'Hy, Bad Lauchstädt and Da'an each have an identified buyer or route to market. Magallanes did not secure one at the price needed. Hyd'Occ remains exposed to its own conversion risk.
- Policy is shifting from supply support to demand construction. Pipeline capacity reservations, hydrogen-readiness rules for gas plants and the proposed steel credit all try to create buyers rather than only subsidise producers. The weakness is that each remains incomplete: reservations are not projects, a draft is not law.
- Fuel cells are decoupling from the hydrogen economy's slow build-out. Their strongest commercial signals this period came from data centres and defence, sectors with urgent needs and high willingness to pay. Buses and ships advance, but cautiously and alongside batteries.
The tension is between a market that wants to be global and a financing reality that is still regional and subsidised. Giga-scale export projects such as Dakhla need long-term buyers abroad. Mid-scale European projects can draw on state support and captive industrial demand. Our view is that the next 12 to 18 months will favour the second model, and that the first will be proved or disproved by FIDs, not by announcements.
Snapshot: scale versus status
| Project | Scale (as reported) | Status | Open question |
|---|---|---|---|
| Normand'Hy (France) | 200 MW PEM; 28,000 t/yr | Poised to start production | Ramp-up and regional market |
| Bad Lauchstädt (Germany) | 30 MW electrolysis; 50 MW wind | First hydrogen produced; integrated commissioning planned 2027 | Full integration with Leuna refinery supply |
| Hyd'Occ (France) | 20 MW; 2,700 t/yr | Inaugurated | Converting planned customer agreements into sales |
| Creca (Scotland) | 50 MW | Planning permission granted | Financing and operational details |
| H2 Magallanes (Chile) | $16bn; over 343,000 t/yr | Withdrawn from environmental evaluation | Long-term offtake and equipment costs |
| Dakhla (Morocco) | 1 Mt/yr ammonia; over USD 25bn | FID targeted by 2027; operations by 2031 | Securing buyers and financing |
What to watch
- Normand'Hy's first production and offtake flows. Whether the plant starts as indicated and how quickly output reaches industrial users will test the state-supported, anchor-offtaker model.
- Hyd'Occ's customer conversion. Watch for announced sales rather than intentions, including any firming of the Plug Power relationship.
- The EU vehicle and steel proposal in Council negotiations. The Environment Committee has voted, but the credit linking cars to low-carbon steel needs Council agreement.
- The Dakhla FID window. A 2027 decision target will show whether very large export projects can secure long-term purchasers, the problem cited at Magallanes.
- The natural hydrogen drilling calendar. Prominence's multi-well campaign is planned for mid-2027, and Mantle8's programme runs 24 months. These will produce the first evidence on whether the early signals add up to recoverable volumes.
For anyone tracking hydrogen energy news for decisions rather than headlines, the filter is simple: who has signed to buy, how will the molecule travel, and what happens if the subsidy stops? The projects that can answer those three questions are the ones now reaching the switch-on stage.