E.ON Waited Two Years for a Number and Got an Adverb: 'Gradually'
E.ON Waited Two Years for a Number and Got an Adverb: 'Gradually'
The headline numbers from E.ON's nine-month report are almost aggressively unremarkable: EUR 7.1 billion of adjusted EBITDA, broadly flat year on year, EUR 2.2 billion of adjusted net income, up 4%, and — by CFO Nadia Jakobi's own framing — roughly 75% and 79% of full-year guidance at the midpoint. That is a company executing a plan. The story of this call is not the plan. It is what the plan's ceiling now depends on, and how close the German regulator just moved to it.
The regulator blinks, but keeps the pen
Ten days before this call, BNetzA postponed its gas-network WACC determination to 2027 and publicly signalled the same treatment for power networks in 2028. This matters mechanically rather than cosmetically: the postponement replaces a low-interest year with a higher-interest year inside the historical reference window. Jakobi calls it “a reflection of the market movements that we have now seen, particularly in the last couple of months with bond yields rising significantly globally” — Nadia Jakobi, Chief Financial Officer · 2026-11-11. It is the same rates transmission the wider tape is wrestling with — the global keyword set for the quarter is dominated by rising-rate phrasing — but for a regulated network owner it lands as a revenue line, not a discount rate.
The second piece of positive news was the new network development plan, which lifts German CapEx requirements by roughly 20% versus the 2024 version and, crucially, adds demand categories the old plan barely modelled: data centres and batteries. That is the investment need story in one document, and it is why NDP 26 entered E.ON's keyword set almost immediately after the call. Management's number for the extra envelope at the back end of the plan — EUR 1.5–2 billion a year — is unchanged from February, which is itself the tell.
'Gradually' is doing a lot of work
Analysts have been chasing this CapEx decision for four straight quarters. In November 2025, the answer was that the drafts were a “closed shop exercise within the regulatory authorities” — Nadia Jakobi, Chief Financial Officer · 2025-11-12. In February 2026, Leo Birnbaum's line was that “we would have hoped, we would know more by this time, but we don't” — Nadia Jakobi, CFO or Finance Executive · 2026-02-25 — after which he added that the prior year's potential upside “continues to apply” — Leonhard Birnbaum, CEO · 2026-08-12. Today the language has genuinely shifted, though only from 'no' to 'maybe, in stages': “we have a gradual and also positive increase in visibility of regulatory parameters... and that could mean that we also gradually increase our CapEx envelope” — Nadia Jakobi, Chief Financial Officer · 2026-11-11.
And when a decision comes, it will arrive on a familiar calendar: “if we were to decide that we gradually increase our investment, we would usually inform you about that as part of our full year reporting” — Nadia Jakobi, Chief Financial Officer · 2026-11-11. Three analysts asked essentially the same question in three different ways and got the same answer three times. The CapEx envelope — a keyword that has now surfaced in five of the last six quarters — remains the single swing factor in the equity story, and it is still shut.
What is actually missing? Not the risk-free rate, and not, apparently, the arithmetic. It is the beta, where the regulator published a broad range; the redispatch treatment inside the efficiency benchmarking; and the OpEx adjustment factor, which is directionally positive but still unquantified. Jakobi was careful to insist the WACC postponement is not a quid pro quo: it is not to be traded against weaker outperformance allowances. That denial was necessary precisely because the market suspects otherwise.
What stopped being discussed
Here is the quieter signal. Twelve months ago, efficiency benchmark complaints and redispatch were E.ON's loudest regulatory grievances — Birnbaum's set piece was that E.ON owns a third of the networks but 70% of the wind and 50% of the solar, and then gets punished for redispatch it cannot control. This quarter, redispatch has largely fallen out of the curated keyword set, and benchmarking has drifted far down the list. The dispute has narrowed from 'the whole framework is wrong' to 'one parameter is wrong'.
That narrowing is genuinely constructive for visibility, but it comes with a subtle inconsistency. On Sweden, Jakobi twice described the preliminary RP5 model as “clearly disappointing and absolutely not satisfactory” — Nadia Jakobi, Chief Financial Officer · 2026-11-11, then told a later questioner that “the current proposal on WACC calculations is in line with our expectations... the methodology that has been used was a favorable methodology” — Nadia Jakobi, Chief Financial Officer · 2026-11-11. Both cannot be true in the same breath. What is consistent is the remedy: E.ON has a mid-triple-digit million euro under-collection to recover, a proven willingness to litigate the challenge in court, and no intention of cutting Swedish CapEx — “we are absolutely determined to spend money where we create value” — Nadia Jakobi, Chief Financial Officer · 2026-11-11.
In the past regulatory period, we have successfully challenged the regulatory decision in front of the court.
Strip out the legal theatre and the adequate return argument is now a timing argument: Germany delivers in 2027–28, Sweden in October 2027. The value-creating spread of 150–200 basis points over WACC is not in dispute. The date is.
Three genuinely new risks, one of them unquantified
The most striking keyword change is not regulatory at all. Bad debt now tops E.ON's own keyword table — a retail-credit theme displacing a decade of network-regulation vocabulary. Jakobi's disclosure is reassuring but hedged: “we have not observed any material deviations in payment behavior across the countries where we operate” — Nadia Jakobi, Chief Financial Officer · 2026-11-11, while conceding that the lagged pass-through of higher power and gas prices means the real test lands next winter. She called the U.K.'s proposed energy-debt circuit breaker directionally welcome. Notably, other reporters in this window flagged the same credit line item at the same moment — bad debt is a sector signal from higher power prices, not an E.ON-specific problem.
Second: agentic AI switching in retail. A new question, and the answer was instructive — E.ON leans on the fact that price comparison website transparency already exists, that seven in ten customers stay more than three years, and that scale procurement with a strong balance sheet is the moat. That is a brand-and-hedging defence against a technology risk, not a product answer. As the global keyword set fills with Personal AI agents and agentic commerce, this is the least quantified exposure in the portfolio.
Third, and faintly delightful: E.ON is reviewing former nuclear sites as data-centre locations, with Nvidia as an early counterparty. Jakobi was emphatic that this is “an initial review of whether former nuclear plant sites could be a suitable location for data center development” — Nadia Jakobi, Chief Financial Officer · 2026-11-11 and no more. data centers keep showing up — as grid-connection demand, as an NDP input, now as a land asset — which is a nice reminder that E.ON's exposure to the AI build-out is through copper, not compute.
The bottom line
Nothing in these nine-month numbers changed. What changed is the probability distribution around the CapEx envelope, which moved from 'binary and blocked' to 'staged and probable'. The next two decision points are the beta consultation and the full-year planning round; the bigger ones are 2027 and 2028. The market is being asked to underwrite a decade of grid investment on the strength of an adverb. E.ON has spent two years arguing it will not spend a euro until it can see the return — and to be fair, it has not. The bull case is that 'gradually' becomes a number in February. The bear case is that beta, benchmarking and OpEx stay in the same drawer Sweden's WACC just came out of.