Decarbonization is no longer optional. But financing the transition remains conditional.
On February 10, 2026, ArcelorMittal confirmed the investment at Dunkerque: €1.3 billion for an electric arc furnace with a capacity of 2 million tonnes per year, scheduled to come online in 2029.
The announcement ends two years of waiting and negotiations. It validates a rational industrial choice, but more importantly, it opens a more tense phase: execution under constraint.
What the announcement says
The figures are clear:
€1.3 billion investment
2 Mt/year capacity (one electric furnace)
2029: commissioning
Emissions cut by two-thirds compared to the current blast furnace
This is consistent with what was anticipated. The chosen project is the hybrid electric furnace model: majority scrap (60-70%), supplemented by imported DRI (30-40%). The integrated H₂-DRI model, initially mentioned, is not adopted at this stage.
ArcelorMittal publicly assumes this choice. At the publication of the European Steel and Metals Action Plan, Aditya Mittal (Group CEO) stated:
“It remains essential to act quickly to end unfair trade and dumping, and to address the high energy costs that make decarbonization projects difficult to progress.”
The message is clear: decarbonization is underway, but it only holds if the European framework follows. Electricity prices, CBAM, trade protection: everything is connected.
What the announcement doesn’t say (but structures everything)
The €1.3 billion investment is confirmed, but it rests on three implicit conditions that are not yet fully secured.
Sustainably competitive electricity prices
Dunkerque will consume 600-800 GWh/year. At €100/MWh (current high range), the annual cost is €60-80 million. At €60/MWh (Scandinavian level), it drops to €36-48 million.
The structural France-Sweden gap represents €20-40 million/year. Over the furnace’s 20-year lifespan, this amounts to €400-800 million in potential additional costs.
The announcement therefore assumes a long-term mechanism (carbon contracts for difference, regulated tariff, nuclear PPA) that is not yet finalized. State-ArcelorMittal negotiations are ongoing, but nothing has been publicly confirmed.
Access to progressively decarbonized DRI
The hybrid model relies on 1.5-2 Mt of imported DRI per year. If this DRI continues to be produced from natural gas (1.0-1.3 t CO₂/t), CBAM will tax it from 2029 onwards.
At €100/t CO₂ (target carbon price for 2030), the tax will be €100-156/t of DRI, representing €150-230 million/year in additional costs.
This estimate is based on an ETS carbon price of €100/t CO₂, the median assumption for 2030. At €70/t (low range), the additional cost drops to €126-168 million/year. At €120/t (high range), it rises to €216-288 million/year. In all cases, the trend is the same: the hybrid model progressively degrades after 2029.
The announcement therefore assumes that by 2030-2035, low-carbon DRI pathways (H₂ or gas with CCS) will be available on the global market. This is not guaranteed.
An electricity grid ready on time
Connecting 300-400 MW of peak power requires reinforcements of 400/225 kV substations and high-voltage lines. Estimated cost: €100-150 million (RTE scope). Timeline: 4-5 years.
If work does not start in 2025-2026, the 2029 schedule becomes tight. RTE has identified Dunkerque as a “priority node” in its ten-year development plan, but local coordination (offshore wind, other industries) remains critical.
Why this announcement now?
Two factors explain the timing.
CBAM forces action
The carbon border adjustment mechanism enters its progressive taxation phase in 2026-2029. Waiting until 2035 to decarbonize would have exposed Dunkerque to an immediate competitive penalty on its sales.
ArcelorMittal had to announce before 2026 to secure its regulatory timeline. Done.
Public support is conditional on timing
France 2030 grants (estimated €500-800 million) and European mechanisms (IPCEI, innovation funds) require firm commitment before mid-2026 to benefit from current allocations.
Further delay would have meant losing hundreds of millions of euros in public co-financing. The project was therefore under budgetary calendar constraint.
What really begins now: execution under constraint
The announcement closes the strategic phase. It opens the execution phase. And that’s where the real tensions will appear.
Meeting the 2029 deadline
3 years to build an electric furnace is technically feasible. But it requires:
Administrative permits finalized within 6-9 months
RTE grid work started before end of 2026
Scrap and DRI supplies secured by 2028
Team training and progressive transition from blast furnaces
Beyond physical infrastructure, the transition also involves skills transformation. Moving from blast furnace to electric furnace means closing certain jobs (coking plant, ore management) and creating others (electric control, scrap management, DRI quality control). ArcelorMittal plans to maintain 2,700 direct jobs, but the social transition (training, reassignments, local acceptance) is a calendar risk factor rarely quantified in business plans.
Any delay cascades. If the furnace starts in 2030 instead of 2029, ArcelorMittal pays 12 additional months of full-price ETS carbon allowances, potentially €80-120 million (assuming €100/t CO₂ x 1 Mt/month emitted by the blast furnace).
Avoiding the CBAM double penalty
From 2029, gas-produced imported DRI will be taxed. If the site hasn’t secured an alternative (H₂ DRI, DRI with CCS, carbon-indexed long-term contracts), it gets stuck:
Structural need for DRI for steel quality (scrap limited to 85-90%)
Increasing CBAM taxation on this input
No immediate fallback solution
The current hybrid model is viable until 2030. After that, it degrades without access to decarbonized DRI.
Not falling behind Nordic competitors
H2 Green Steel (Sweden) targets 5 Mt/year in 2030 with an integrated DRI-H₂ model and electricity at €50-60/MWh.
If Dunkerque remains in hybrid EAF (gas DRI + scrap) with electricity at €100/MWh, it’s competitive on standard steel thanks to CBAM. But on premium segments (automotive, demanding carbon certification), H2GS will set a new standard that Dunkerque will struggle to match without later conversion.
The 2029-2035 window will be decisive: either Dunkerque shifts to H₂ DRI, or it accepts a lasting “mid-market” positioning.
The macro risk: will European politics follow through?
The three identified conditions (electricity price, low-carbon DRI, timely grid) are not just technical. They are political.
The European electricity market reform has been under discussion since 2022, but blocked by structural divergences between member states. The sustainability of carbon CfDs depends on French and European budgetary capacity over 15-20 years. Low-carbon DRI availability assumes global hydrogen pathway coordination that doesn’t yet exist. Trade protection against Chinese dumping remains subject to Brussels arbitration.
Dunkerque will only be competitive in 2035 if Europe fulfills its energy and trade coordination commitments. Yet recent European history (2022 gas crisis, Covid fragmentation, Brexit) shows that coordination is never guaranteed.
That’s the real macro risk. Not technical. Political.
What this announcement reveals about European industrial decarbonization
Dunkerque is not an isolated case. It’s a test case for the systemic tensions of European industrial transition.
Decarbonization, yes. But at what cost, and funded by whom?
ArcelorMittal invests €1.3 billion. But this figure assumes:
€500-800 million in public grants (France 2030, IPCEI)
Recurring support mechanisms over 15-20 years (carbon CfDs, electricity tariffs)
Trade protection (CBAM, anti-dumping)
Without this triple crutch, the project doesn’t hold. Decarbonization then becomes a disguised quasi-nationalization: the state finances the investment, guarantees operating costs, and protects markets.
This may be necessary. But it must be said clearly.
Europe protects… but doesn’t coordinate enough yet
CBAM protects decarbonized European production against non-EU imports. That’s good.
But it doesn’t solve:
Intra-European energy cost gaps (France €100/MWh, Sweden €60/MWh)
Supply chain fragmentation (DRI, scrap, H₂ infrastructure)
Competition between member states to attract green investments (national subsidies in disarray)
CBAM protects Europe. It doesn’t make it competitive.
Industrial sovereignty remains conditional
Dunkerque reduces its dependence on coal (100% imported). But it creates a new dependence on imported DRI (30-40% of inputs).
Energy sovereignty improves (French electricity). Material sovereignty remains partial.
Conclusion: a necessary announcement, but not sufficient
The February 10, 2026 announcement is good news. It validates a massive investment, secures employment, and commits to real decarbonization.
But it also opens an execution phase under constraint, where every delay, every cost overrun, every administrative or logistical blockage can shift the project from industrial success to precarious balance.
Dunkerque will be decarbonized in 2029. That’s confirmed.
Will Dunkerque be competitive in 2035? That depends on what happens between now and 2030 on:
Industrial electricity prices
Low-carbon DRI availability
Infrastructure scaling
European electricity market reform
The announcement closes a strategic chapter. It opens the political execution phase.
And that’s where everything is now at stake.