For three years the European regulatory conversation about hydrogen ran in one direction. The RFNBO rules — additionality, temporal correlation, geographic correlation — were treated as a cost imposed on producers, the industry asked for them to be relaxed, and the risk a developer carried was that Brussels would not relax them in time.
Something reversed in the middle of September 2026, and it has not been priced into most project models yet.
What leaked
A draft Commission impact assessment for a revision of the Renewable Energy Directive circulated in mid-September. The clearest public characterisation of it comes from a law firm alert, which is careful about what the document is:[1]
This leaked RED IV impact assessment is unofficial and incomplete.
It is not a legislative proposal and has no legal weight.
That caveat belongs at the front, not the end. Nothing described below has been proposed, adopted, or even formally published. What makes it worth a project analyst's attention is not that it will happen, but that it is now a possibility that nobody was modelling a month earlier.
What the draft is reported to recommend:[1]
abolishing the existing mandatory national sub-quotas for green hydrogen and e-fuels
replacing them:[1]
with non-binding targets at EU level
What the leaked text did contain, and what was read in it directly, is a different and more useful kind of sentence — the Commission's own arithmetic on the gap:[2]
falling short of the estimated amount of 3.3 Mt of renewable hydrogen required to meet the 2030 targets
and its own count of what exists:[2]
the 2025 projected electrolysers capacity installed in Europe ranged between 514 MWel and 800 MWel
A target measured in millions of tonnes, a fleet measured in hundreds of megawatts.
Why abolishing a quota is a demand event, not a compliance event
Figure 1
What is currently binding, and what the leaked draft would make indicative
Renewable fuels of non-biological origin as a share of their sector, under RED III as it stands.
The smallest number on this chart and the one with the largest e-fuel projects attached to it.
The obligation that makes a refinery a buyer.
The leaked impact assessment states the industry target would have to increase to 60% by 2035.
These are the obligations a European offtaker currently faces, and therefore the reason an offtaker signs. The leaked draft is reported to recommend abolishing the mandatory national sub-quotas behind them and replacing them with non-binding targets at EU level. Germany, Spain and Belgium have already written national RFNBO mandates into law running to 2040.
Source: RED III targets as characterised in King & Spalding's analysis of the leaked draft · Data as of 2026-09-24
Show the data ▸Hide the data ▾
| Obligation | Year | Share |
|---|---|---|
| RFNBO in transport fuel supply | 2030 | 1% |
| Renewable hydrogen in industry | 2030 | 42% |
| Renewable hydrogen in industry | 2035 | 60% |
This is the part that reframes the risk, and it is worth stating slowly because the industry itself took a while to say it out loud.
The RFNBO rules do two separate things. They impose conditions on how hydrogen must be produced to count — that is a cost, borne by the producer. And they impose obligations on how much renewable hydrogen certain buyers must consume — that is a market, created for the producer.
Everyone spent two years arguing about the first. The leak is about the second.
A refinery that must source 42% of its hydrogen from renewable sources by 2030 is a buyer whether or not the price works, which is the entire reason a European electrolyser project can contemplate a long-term contract at all. Make that target indicative and the refinery becomes a buyer only if the price works. For most projects at most European electricity prices, the price does not work — the ceiling a buyer can pay without losing competitiveness is below USD 2/kg for most sector and region combinations, and for steel it is negative.
The obligation was not a tax on the project. It was the project's demand.
The industry noticed, and changed sides in a week
On 23 September 2026, 170 companies signed an open letter to the President of the Commission, the Executive Vice-President for the clean transition and the Energy Commissioner.[3][4] Its title is the argument: Keep the targets.
The figures in it are the signatories' own, and are reproduced here as their claims about their own exposure rather than as independent measurements:[3]
With over €15 billion committed to this day, Europe holds the largest volume of investments in end-use hydrogen applications
Investment in Europe grew 35% in the past year
over 4 GW of electrolysis capacity under construction
Meeting 2030 REDIII targets represents an investment opportunity of €50-60 billion
The reversal is the point. In July, the ask was to loosen the rules. In September, the ask is to keep them. Nothing about the economics changed in between — what changed is which half of the regulation came under threat, and the industry discovered which half it actually depended on.
King & Spalding adds the legal dimension that a developer should take seriously: Germany, Spain and Belgium have already implemented national legislation with RFNBO mandates running to 2040, and changing them would create regulatory uncertainty — with the alert raising legitimate expectations and investment treaty claims as live questions.[1]
And the other track has not moved at all
Figure 2
Two regulatory tracks, both unresolved, pointing in opposite directions
What has been promised, what exists, and when each is expected, as at 4 October 2026.
Article 6 of the delegated act. Unchanged: no revision has been published.
What the German associations' letter of 21 September asks to postpone to 2035.
Article 11. The same letter asks for full application to be postponed to 2035.
Committed for before June 2026, then autumn. Not published as at 4 October 2026.
King & Spalding: the Commission is scheduled to publish the proposed text of RED IV by the end of 2026.
The two tracks do not cancel out. If the delegated act is relaxed, compliance gets cheaper. If RED IV drops the binding sub-quotas, the buyer's obligation becomes optional. A model that assumes the first without pricing the second is valuing a project on the good half of an unresolved question.
Source: Delegated Regulation (EU) 2023/1184; industry letters of 21 and 23 September 2026; King & Spalding · Data as of 2026-10-04
Show the data ▸Hide the data ▾
| Instrument | Status at 4 October 2026 | What it would change |
|---|---|---|
| Delegated Regulation 2023/1184 | in force, unrevised | the cost of compliance |
| Targeted review of that act | promised before June 2026, not published | additionality and hourly matching |
| RED IV | proposal expected end of 2026 | whether the buyer's obligation is binding at all |
Here is the easiest thing to miss while watching the leak. The RFNBO delegated act — the one that actually governs how a project must be built — is unchanged.
A targeted review was committed for before June 2026, then slipped to autumn. On 21 September 2026 the German hydrogen associations wrote asking the Commission to:[5]
publish its proposal for the targeted review without further delay
and to:[5]
establish a clear and reliable timetable this September for the remaining process
Their specific asks, which are what relaxation would look like if it arrives: postpone full application of additionality to 2035, keep monthly temporal correlation instead of hourly or postpone hourly to 2035, and lower the bidding-zone exemption threshold from 90% to 80% renewable electricity.[5]
None of that has been adopted. Which means the deadline this blog wrote about in July still stands exactly as written: a plant in operation before 1 January 2028 is exempt from the additionality tests until 1 January 2038, and that door has not moved.[6]
What to do with this before your next committee
The temptation is to pick a side — either Brussels will relax the rules and projects get cheaper, or Brussels will drop the mandates and projects lose their buyers. Both are forecasts, and neither is available as a fact today.
What is available is the shape of the exposure, and it can be carried in a model without predicting anything.
- Separate the two regulatory variables in your model. One line for compliance cost, one line for contracted demand. They have been moving together in people's heads and they are now moving apart.
- Ask what your offtake contract says about a change of law. A take-or-pay signed against a national mandate running to 2040 is a different instrument if that mandate becomes indicative. The clause that matters is the one nobody read at signature.
- Name which national mandate your buyer is complying with. Germany, Spain and Belgium have legislated to 2040. A buyer in one of those jurisdictions has a different exposure to a RED IV change than one relying on the EU-level target alone.
- Do not move your additionality or matching assumptions yet. No revision has been published. Building a case on an expected relaxation is building on a document that does not exist.
- Treat the 1 January 2028 date as unchanged until something official says otherwise. It remains the only date on the European timeline that a project currently in construction can still act on.
- Discount any figure from the leak that you have not seen in a document. The partial PDF in circulation does not contain the headline targets being attributed to it.
The uncomfortable summary is short. For three years the regulation was the obstacle. The month it started looking like it might be removed, 170 companies wrote to say it was the asset — and both the removal and the relaxation are still unwritten.