Part I — Situation overview

The last still producing unit of the Paks Nuclear Power Plant came closest to shutting down at dawn on 4 August 2026. According to Portfolio’s report, around half past one in the morning the water level was already only millimetres from the critical value, but at that point the fall stopped, and by morning the river had risen by one and a half centimetres — the last turbine is thus for now operating safely. According to the data of the National Directorate of Water Management, the level of the Danube at Paks was −137 centimetres at that time. What made the situation really tight was the load side of the system: MAVIR, that is, the Hungarian Independent Transmission Operator Company Ltd, forecast a system peak of 7,282 megawatts for Monday evening. This fell short by barely 200 megawatts, about 3 per cent, of the summer record of 7,488 megawatts measured on 29 June. The difference is that in June Paks was still producing more than four times as much electricity.

The Monday evening factual data turned out more favourable than that. The Prime Minister, reporting in from the transmission operator’s control centre, announced that the load came in at around 6,300 megawatts, which fell more than 500 megawatts short of the expert forecast. He added that the total of the voluntary curtailment undertakings by large consumers had already exceeded 630 megawatts on Monday — on Saturday this value was still around 250 megawatts, on Sunday 400. The undertakings are tied to concrete players: according to the list published by Kapitány István — designated as the head of the economy and energy portfolio, with the title used by the press — MOL TVK undertook a reduction of 63 megawatts, Samsung SDI 32.8, the railway infrastructure operator 22.5, and SK and Audi Hungaria 22 each.

On Tuesday morning, however, the first analysis appeared that describes with figures what held the system. According to Mandiner’s account, the economist Sebestyén Géza broke down the data of the midday peak period: a mere 42.8 per cent utilisation of the domestic 8,600 megawatts of solar power plant capacity almost completely covered the country’s entire electricity demand, with the remainder made up of 501 megawatts of net imports and 330 megawatts of gas-based, 225 megawatts of coal-based, 176 megawatts of Paks and 141 megawatts of biomass-based production. According to MIAK’s reading this is the most important policy fact of the crisis’s ten days so far — but not in the direction in which the public debate immediately set off. The series of figures does not prove that there is sufficient generating capacity in the country; it proves that the bottleneck has moved between the times of day. At noon the system operates with a one-and-a-half-fold reserve, while in the evening — when the critical period, according to the official request too, falls between 5 and 10 p.m. — this same capacity gives zero megawatts. The risk of the crisis is not the lack of installed capacity, but that today nothing carries the midday surplus over into the evening.

Part II — Literature foundation

Two conceptual frames help this time-of-day bottleneck to be read as a planning rather than a communications question. The work The Innovator’s Dilemma by Clayton M. Christensen (American business thinker, a researcher of the organisational consequences of technological realignments) describes how market leaders systematically undervalue technologies that underperform on the established performance yardstick, because they measure them on their own yardstick. On the yardstick of traditional electricity planning (dispatchability, availability) the solar panel counted for decades as the worse product, and this is precisely what explains why storage never got to the centre of the planning agenda. The volume On Fire by Naomi Klein (Canadian author, one of the most widely read analysts of the social conditions of climate policy) gives the other side: according to her argument what is needed is not the green repainting of the old, centralised energy structure but distributed, and where possible community-owned, production — and the Hungarian midday peak on Monday was held by exactly such a distributed producer base consisting of tens of thousands of participants. The detailed treatment of the literature — author by author, with quotations — can be found in the 6.4 Literature in detail section.

📖 Source: Clayton M. Christensen: The Innovator’s Dilemma; Naomi Klein: On Fire — The Burning Case for a Green New Deal

Part III — MIAK’s concrete proposal

MIAK proposes three measurable measures. None of them is about building a new power plant: the first two can be carried out without amending legislation, from existing data and within the planning process, while the third does require legislative fixing.

3.1 An evening peak flexibility balance — a public indicator (within 30 days)

Today public discourse knows a single figure: installed capacity. This figure is informative in the midday hours, but misleading in the evening peak, because of the 8,600 megawatts of solar power plant capacity zero is available after 8 p.m. MIAK proposes that within 30 days the transmission operator publish a separate, daily updated balance exclusively for the evening peak hour: on one side the load forecast for that hour, on the other the components of the capacity actually available then — dispatchable domestic generation, contracted import capacity, storage discharge output, and callable demand-side curtailment. Solar production should feature in this balance with zero weight for the evening hour, because factually that is what it gives. This proposal follows from the G19 radical transparency and the D9 data-driven decision support programme points, and makes a single question answerable to which nobody can respond with a figure today: how large the reserve is in those four hours in which the system is genuinely tight.

3.2 A numerical target value for storage and flexibility capacity (within 90 days)

MIAK’s own energy transition programme point, K2, sets a target of 12 gigawatts of solar capacity by 2030 — in the light of the 8,600 megawatts now reported this target figure is practically on the verge of being met, while the second element belonging to it, grid storage and flexibility capacity, has been left without a numerical target value. This is the most important own lesson of the present crisis for MIAK too: the capacity target has been met, the storage target did not even exist. MIAK proposes that the next energy strategy document should within 90 days contain a stand-alone storage target value expressed in gigawatt-hours and a separate target figure for contracted demand-side flexibility, both given as a percentage of the prevailing evening peak load — thus the target is tied to the real bottleneck of the system, not to a technology. In Christensen’s frame (see 6.4.1) this step is exactly what is regularly omitted: raising the element that previously performed badly on the established yardstick into a stand-alone, measurable target, rather than treating it as a supplement to the old yardstick.

3.3 An annual security-of-supply stress test before Parliament (by the 2027 planning cycle)

Every component of the present situation was thinkable in advance: a long heatwave, a record low river water level, consequently reduced nuclear power plant output, and all of this close to the summer peak load. MIAK asks that by the start of the 2027 planning cycle an annually compulsory, public security-of-supply stress test be brought into the legal order, which calculates through at least three scenarios — a heatwave without Paks, a heatwave alongside low import capacity, and a lasting lull at the summer peak — and whose result is prepared by the transmission operator, with the head of the energy portfolio submitting it before Parliament in the form of a report. The template exists in MIAK’s own programme: the K1 measurable climate targets programme point describes exactly such an annual accountability report going before Parliament for emissions, while the G25 energy price shock preparedness plan covers the economic side. Of the two, it is the security-of-supply dimension that is missing today.

The three proposals can be strung on a single principle: let the measurement follow the bottleneck, not the technological debate. The public debate today runs along two equally mistaken axes — that the solar panel “saved” the system or that wind energy is “useless” — while both modes of production are weather-dependent, and the answer to both is the same: to carry the production over to where the consumption is. Whoever argues in favour of solar panels must also demand storage; so must whoever criticises wind energy. For MIAK this is not a compromise but the precise posing of the question.

Part IV — Expected effects and risks

Dimension Expected effect Risk
Economy Contracted flexibility gives companies predictable remuneration in place of today’s voluntary, uncompensated undertakings The storage target value creates an investment obligation at a time when the budget is under strain elsewhere too because of the energy crisis
Society The public reserve data for the evening peak make it understandable why saving is requested precisely in the evening In the short run the time-of-day balance shows a worse picture than installed capacity — this has to be accepted
Public administration The annual stress test takes security of supply out of the daily political debate and makes it a planning question If the stress test remains formal, the report becomes an annual routine without substantive consequence

The main trade-off is tense between measurement and communication. Installed capacity is a big figure, and comfortable for every government; the capacity available projected onto the evening peak is a small figure, and uncomfortable. Precisely for that reason the latter is the one that can underpin a decision. The proposal tips to the risk side if the storage target value narrows into a technological prescription — if the document prescribes a battery where a pumped storage reservoir or a contracted industrial disconnection gives the same service for less money. The target value should therefore be expressed in service, not in equipment: how much output, for how long, with what notice period it can be called.

The second trade-off lies between voluntariness and contract. The voluntary undertakings reached more than 630 megawatts fast, within a few days — a contractual system would not have produced this at such a pace. Voluntariness, however, is a once-usable resource: if a request arrives again next summer, that same group of companies will already be citing last year’s costs. Contractual flexibility is therefore not the opposite of voluntariness but its sustainable continuation.

Part V — Measurability and summary

5.1 What is worth tracking? (proposed KPIs)

Four performance indicators (KPIs, Key Performance Indicators) are worth tracking over the next twelve months:

  1. Evening peak reserve: the difference between the capacity available for the hours between 7 and 9 p.m. and the forecast load, broken down by day — proposed target: at least 10 per cent of the peak load, in publicly disclosed form.
  2. Storage discharge output: the aggregate output of domestic grid and behind-the-meter storage capacity callable in the evening hour — proposed target: a measurable, annually rising path, expressed in gigawatt-hours as well.
  3. Share of contracted demand-side flexibility: the contractually committed, remunerated disconnectable output as a percentage of the summer peak load — proposed target: the raising of today’s level of voluntary undertakings into contract within one year.
  4. Energy not supplied: the quantity of consumption not served, whether by plan or under duress, in megawatt-hours, broken down by area — proposed target: zero, and the measurement should be carried out even when it is zero.

5.2 Summary

MIAK’s key message in a single sentence: what follows from Monday is not that we have enough capacity, but that the bottleneck is a time-of-day one, and today we do not measure it. Concretely, it asks that within thirty days the transmission operator publish a separate evening peak flexibility balance, that within ninety days the next energy strategy document contain a stand-alone storage and flexibility target value, and that by the start of the 2027 planning cycle the annual security-of-supply stress test going before Parliament be brought into the legal order. Neither request is about whether the energy policy of the past sixteen years was good or bad — the solar build-out is a substantive achievement, and the storage layer that was not built alongside it is a substantive gap, and both come from the same period.

Two MIAK foundational values are in play in this matter. Data-drivenness, because today’s debate revolves around two claims — “the solar panel saved the system” and “wind energy is useless” — that can each be justified separately with a single selected hour, but together do not form a basis for decision; data-drivenness is not that we have data, but that the measurement is aimed at the hour and the quantity at which the decision is settled. And being ideology-free, because the mathematics of the evening peak is party-neutral: the same amount of storage is needed whether someone regards the solar panel as the future or nuclear energy — the missing four hours are in the same place in both visions.

Part VI — Justifications and further sources

6.1 The press framing by spectrum

The economic band was the only one that described the operation of the system as a technical balance. Portfolio brought both the news of the dawn turnaround (“Paks was within millimetres of a complete shutdown”) and the daily analysis of the system load, including the comparison of the near-record peak with the missing Paks output — that is, this band gave a numerical frame for the situation, without governmental evaluation. The same paper also published the news of the employee consultation started at Audi Hungaria, thereby bringing in the labour law dimension of the crisis, which no other band touched on.

On this day the conservative band turned the subject towards the production structure, and with that — independently of its intention — brought the most substantive policy material. Mandiner presented Sebestyén Géza’s full breakdown of the midday energy mix, and the same framing continued with a critique of the domestic role of wind energy; Magyar Nemzet quoted the nuclear energy expert Hárfás Zsolt, who, referring to the transmission operator’s data, recalled that last year there were several periods in which the combined output of the solar and wind power plants fell practically to zero. These two articles together describe exactly the fact from which MIAK derives today’s proposals — only in the service of the opposite political conclusion.

The liberal-left and the public affairs bands carried the political and personal thread of the crisis. Telex highlighted the Prime Minister’s report-in from the control centre and the call addressed to one of the large consumers, while 444.hu wrote that the energy crisis had brought the most active anti-government campaign of the opposition Fidesz so far, including the group leader’s claim that they had handed over the country in a better condition as regards energy supply. 24.hu and ATV carried the operational side: the list of corporate undertakings and the area-by-area blackout warnings. From the comparison of the bands one common gap is visible: not a single paper posed the question of what happens in the evening hours, when the capacity celebrated at noon does not produce — the public debate used the data of the midday hour to answer the evening problem.

6.2 Facts and data

Data Value Source
The water level of the Danube at Paks −137 cm National Directorate of Water Management, 24.hu, 3 August 2026
Water level rise from dawn to morning +1.5 cm Portfolio, 4 August 2026
Forecast Monday evening system peak 7,282 MW MAVIR, Portfolio, 3 August 2026
2026 summer system peak record 7,488 MW (29 June) MAVIR, Portfolio, 3 August 2026
Actual Monday evening load approx. 6,300 MW statement by Magyar Péter, Telex, 3 August 2026
Shortfall against the forecast more than 500 MW statement by Magyar Péter, Telex, 3 August 2026
Voluntary large-consumer undertakings (Monday) more than 630 MW Telex, 3 August 2026
Voluntary undertakings (Sunday / first tally) 400 MW / approx. 250 MW 24.hu, 2 August 2026
Domestic installed solar power plant capacity 8,600 MW analysis by Sebestyén Géza, Mandiner, 4 August 2026
Solar power plant utilisation in the midday peak 42.8% analysis by Sebestyén Géza, Mandiner, 4 August 2026
The supply breakdown of the midday hour (elements alongside solar) 501 MW import, 330 MW gas, 225 MW coal, 176 MW Paks, 141 MW biomass, 10 MW hydro, below 1 MW wind analysis by Sebestyén Géza, Mandiner, 4 August 2026
Domestic installed wind turbine capacity 325 MW analysis by Sebestyén Géza, Mandiner, 4 August 2026
The largest individual undertakings MOL TVK 63 MW, Samsung SDI 32.8 MW, railway infrastructure operator 22.5 MW, SK and Audi Hungaria 22 MW each statement by Kapitány István, 24.hu, 2 August 2026
Critical daily period 5–10 p.m., especially 6–9 p.m. Portfolio, 4 August 2026
Romanian evening import need 2,000–2,500 MW per day Transtelex, 3 August 2026

A single connection deserves highlighting. The 42.8 per cent midday utilisation and the 8,600 megawatts of installed capacity together mean that in the midday hour about 3,700 megawatts of solar output was available, and this order of magnitude covered the overwhelming part of the country’s consumption at that time. The same 8,600 megawatts gives zero megawatts at 9 p.m. The system is therefore not short of capacity but short of bridging in time — and it is this difference that settles what it is worth investing in over the next five years. Whoever regards installed capacity as the metric arrives at the conclusion that the task has been accomplished; whoever measures the evening hour arrives at the conclusion that the work is only beginning.

6.3 Policy dimensions

  • Environment and climate (programme points) — the storage pillar of the energy transition plan and the annual accountability of security of supply (programme point ID: K2, K1, K7);
  • Economy (programme points) — the radical transparency of system data and the energy price shock preparedness plan (programme point ID: G19, G25);
  • Transport and infrastructure (background material) — the exposure of the grid infrastructure and the significant undertaking of the railway infrastructure operator within the curtailment circle;
  • Public administration and e-government (programme points) — the audit of the organisational routines of crisis decision-making and of the delimitation of measurement responsibilities (programme point ID: KI11);
  • Digitalisation and AI regulation (programme points) — the machine-readable publication of hourly system data (programme point ID: D9, D2).

6.4 Literature in detail

6.4.1 Clayton M. Christensen: The Innovator’s Dilemma

Christensen’s fundamental thesis is that market-leading organisations do not fail because they manage themselves badly, but precisely because they manage themselves well: they weigh things according to their existing customers and the established performance yardsticks, and this yardstick systematically undervalues those new technologies that are worse on the old yardstick. In the author’s own formulation, disruptive technologies “result in worse product performance, at least in the near term”, and that is exactly what brought down the leading firms in the industries examined. The volume also deals separately with the fact that such technologies are typically simpler, smaller and cheaper, and at first are usable only in smaller markets remote from the mainstream — and then from there become competitive in the mainstream too.

From the point of view of Hungarian electricity planning this frame explains two things. One is why the solar panel remained for decades at the periphery of system planning: the traditional yardstick was dispatchability and availability, and on that the solar panel really does perform weakly. The other, and now the more important: this same yardstick distortion today affects storage. Storage does not even feature on the traditional production yardstick — it does not produce electricity, it merely relocates it in time — and therefore it is almost invisible in capacity-centred planning. In Christensen’s frame this is the typical moment: the most important missing element of the system is the one that the existing metric cannot show. Proposal 3.2 asks for a stand-alone target value for storage precisely for this reason, rather than a supplement to an existing target.

📖 Source: Clayton M. Christensen: The Innovator’s Dilemma

6.4.2 Naomi Klein: On Fire — The Burning Case for a Green New Deal

The central claim of Klein’s volume is that the answer to the climate crisis cannot be the technological repainting of the great state programmes of the 20th century. According to her argument what is needed is not a “green-painted New Deal” or a “solar Marshall Plan” but a change of a different kind: solar and wind production that is distributed and, where possible, in community ownership, as opposed to the highly centralised, monopolistic energy structure of the New Deal. To this the author also adds that individual action — after years of selective waste collection and bulb changing — will never be a sufficient answer: climate change is a collective problem that requires collective action.

The Hungarian Monday is an unusual, data-supported illustration of this thesis. In the midday peak the system was held not by a single large power plant but by tens of thousands of mutually independent producers — an ensemble of household-sized systems and of solar parks large and small — that is, exactly the distributed structure that Klein formulates as a normative argument. At the same time the Hungarian case also bears out Klein’s second thesis, and in an uncomfortable direction at that: the evening peak was bridged this time by individual voluntary going-without and by corporate goodwill, that is, precisely by the individual action of which the author writes that in itself it is not sufficient. MIAK derives from this not a devaluation of voluntariness but the conclusion that the present performance has to be institutionalised — into contract, target value and measurable reserve — otherwise it remains a one-off resource.

📖 Source: Naomi Klein: On Fire — The Burning Case for a Green New Deal

6.5 International comparison

The Romanian parallel shows this same time-of-day structure, only more sharply. In the neighbouring country Dacia and Ford halted their production until 19 August, the government consulted with more than eighty industrial players with consumption exceeding 50 thousand megawatt-hours a year, and asked the companies to hold back specifically for the period between 7 and 11 p.m. In the evening peak the Romanian system needs 2,000–2,500 megawatts of imports a day, while the low water flow also limits the hydropower plants, and one unit of the Cernavodă nuclear plant has already shut down — the two units together give close to a fifth of the country’s production. The Romanian government decision in preparation is ahead of the Hungarian one in two of its elements as well: it plans at least 24 hours of prior notice before compulsory curtailment, and it adjusts the permitted consumption to the company’s technology. Both are precisely the element of predictability that proposal 3.1 asks for in the Hungarian system on the measurement side.

And the offer of electricity exports arriving from the Slovak side is a reminder that regional assistance is at its scarcest in the evening hour: if the evening peaks of several neighbouring countries need imports at once in the same heat, the import capacity exists on paper, but in practice they compete for it. Central European mutual assistance therefore does not substitute for domestic storage — at most it supplements it.

Environment and climate

  • K2 — Energy transition plan
  • K1 — Measurable climate targets
  • K7 — Energy market shock resilience
  • K8 — Electric transport transition strategy

Economy

  • G19 — Radical transparency in economic decision-making
  • G25 — Energy price shock preparedness plan

Digitalisation and AI regulation

  • D9 — Data-driven public policy decision support system
  • D2 — Open data programme

Public administration and e-government

  • KI11 — Organisational behaviour audit

Proposed new programme point: An evening peak flexibility target value — for the Environment and climate area: the energy strategy should contain a storage target value expressed in gigawatt-hours and a contracted demand-side flexibility target value given as a percentage of the peak load, defined in service and not in equipment, with an annual review point.

6.7 List of sources

Press sources (MIAK press monitor, 4 August 2026 — topic 1):

Knowledge-base references (books):

  • 📖 Clayton M. Christensen: The Innovator’s Dilemma
  • 📖 Naomi Klein: On Fire — The Burning Case for a Green New Deal

Note: the local file path of the books does NOT appear in the visible text of the blog — only the author and the title. The file path is an internal matter of the generation process, not the reader’s.

MIAK internal materials:

  • MIAK policy area: Environment and climate (programme points; programme point ID: K2, K1, K7)
  • MIAK policy area: Economy (programme points; programme point ID: G19, G25)
  • MIAK policy area: Transport and infrastructure (background material)
  • MIAK policy area: Public administration and e-government (programme points; programme point ID: KI11)
  • MIAK press monitor, 4 August 2026 — topic 1, score: 95/100

Additional public data sources:

  • ENTSO-E Transparency Platform — hourly system load, generation mix and cross-border flows
  • MAVIR — daily system load and system peak time series
  • MEKH (Hungarian Energy and Public Utility Regulatory Authority) — electricity industry statistical publications
  • National Directorate of Water Management — Danube water level data

Generation metadata