2025 could come to be characterised as the year the lights began to flicker across Europe as threats to energy security begin to manifest. We almost ran out of electricity in Britain in January. We did have a blackout at Heathrow Airport in March and Iberia suffered a full peninsular blackout in April. We also had warnings that accelerated decline of North Sea gas production might threaten the viability of key pipelines leading to winter gas shortages that would threaten the power sector.

2025 was also a landmark year for me personally, as some of my content went viral and I became a little bit famous in energy circles which has had some interesting consequences – being asked for selfies is strange, getting death threats is disturbing.

NESO demand forecasts remain a source of concern

The year began with a near blackout on 8 January. My analysis of the event went viral. NESO, claimed, implausibly to have had 3.5 GW in reserve despite issuing a slew of system warnings and writing a blank cheque to Energinet for bringing one bipole of the Viking interconnector back early from maintenance. Balancing the market that day cost some £23 million – 10 times more than normal – and exposed failings in NESO’s forecasting, leading it to institute an audit of its demand forecasting (announced in the Operational Transparency Forum on 22 January).

I had another look at the demand forecasts to see if anything has changed. NESO publishes a comparison of its half-hourly day-ahead forecast here. I compared it with BMRS data for the year 1 November 2024 – 1 December 2025 (the final date of the NESO forecast spreadsheet) and immediately found that NESO compares its forecast to INDO and the data do not match BMRS for 25 and 26 December 2024 for some reason. The difference in some settlement periods is just under 6 GW which is huge by system standards (18% of the actual demand in that settlement period!).

During the one year in question, the average error of the NESO forecast vs NESO version of INDO was 675 MW (679 MW vs BMRS INDO) and the maximum error was 4,931 MW (5,712 MW vs BMRS INDO). These are really significant errors – the average demand per BMRS which ought to be the official record, was 26,422 MW, so the average error was 2.6% of demand and the maximum error was 12.5% of maximum demand.

When comparing the error in any given settlement period with the demand in that period, the errors are up to 22.7% of demand. These errors are very significant. An error of 5.7 GW is more than 4 times higher than the SQSS – the amount of reserves NESO is required to hold to protect against the loss of the largest infeed to the grid. This means that a forecast error of this size on a tight day would overwhelm the system. (A forecast error isn’t an SQSS event, but if there were such a large error on a tight day the reserves held against an SQSS event wouldn’t be enough to cover it meaning the system could be short even without the loss of a large infeed.)

The real problem could be even worse – NESO does not balance the grid in half-hourly intervals, it balances it from moment to moment. If the average error over 30-minutes is so large, the 5-minute or 1-minute error could be even larger, which poses a real threat to the grid. Unless its forecasting improves, NESO might just run out of luck one day, triggering a blackout. Plus, NESO should benchmark its forecasting performance against actual final outturn rather than INDO which is the initial outturn.

A comparison of NESO’s published day-ahead half-hourly demand forecasts with outturn data shows no evidence of improvement over the year following the January near-miss and the announcement of the INDO forecast audit. Forecast errors remain volatile and fat-tailed throughout the period, with frequent deviations of ±10–20% of demand and absolute errors of several gigawatts occurring well beyond winter stress periods. The envelope of errors does not narrow over time, suggesting that either the audit has not resulted in material changes to forecast performance, or any improvements have been overwhelmed by growing system complexity.

Heathrow blackout exposed risks of aging equipment and poor maintenance

The drama continued in March when a transformer failure in West London caused a localised blackout affecting Heathrow airport. While I remain critical of Heathrow’s approach to its electricity supply risks – CEO Thomas Woldbye insisted the event was “unprecedented” despite an near-identical outage at Atlanta airport in 2017 – the real subject of my concern is the poor management of retirement risk on the power grid.

The subsequent report into the incident uncovered a litany of maintenance failures at National Grid Electricity Transmission (“NGET”). However, while NGET can be blamed for failing to act on identified faults in a timely manner, Ofgem also shares some blame in placing too much emphasis on investing in “load” – ie new connections, rather than “non-load” ie legacy infrastructure. Ofgem says it has never told any network operator to neglect legacy infrastructure and has never refused funding for essential upgrades, but by making its focus on connections so explicit, it is not surprising that network operators skew their internal efforts accordingly.

The North Hyde transformer failure also exposed problems with NGET’s assessment of asset life – a transformer that was 57 years old had a very low “end-of-life” (“EoL”) score, which for assets that are expected to last 50-65 years is difficult to justify. NGET had assigned an EoL score of just 12.7 out of 100 for the defective transformer (0 is for new assets and 100 is for ones that have reached the end of their lives). Ofgem uses NGET’s EoL score as a marker for non-load capex, saying in the RIIO-3 Final Determination that “we have directed NLRE (non-load related expenditure) funding for NGET at those assets with an EoL score of 75 out of 100 or higher”. This is clearly problematic if elderly assets are wrongly assigned low EoL scores by NGET.

Ofgem announced an official enforcement investigation in July that has yet to report back on its conclusions, but, unfortunately, NGET’s RIIO-3 Final Determination has been issued that relies on potentially inadequate EoL scores for the allocation of non-load capex. Ofgem has explicitly said it is front-loading load-capex in the expectation that faster delivery of net zero will lower bills for consumers. While this is unlikely, it means that if it emerges that the EoL scores are too optimistic, non-load capex will be forced up through re-openers, and consumers will face even higher bills.

This year Ofgem commissioned updated asset life analysis from consultants Cambridge Economic Policy Associates (“CEPA”), which found that 30% of switchgear, almost two thirds of overhead lines and just over a third of transformers were built before 1980 and are more than 45 years old. Well maintained switchgear should last 35-50 years, overhead lines last 40-60 years, and transformers can be expected to last 50-70 years. This means that a significant portion of NGET’s assets are now in the end of life window.

The Iberian blackout exposed vulnerabilities caused by widespread use of inverters

In April, we had an even more dramatic event: the full blackout of the Iberian peninsula. There has been much debate about the causes, with vested interests shouting abut how it was or wasn’t the fault of renewables. I have written a number of posts analysing the event, based on official reports, and concluded that the main problem was widespread non-compliance with the grid code by inverter-based generators.

• Voltage, inertia and the Iberian blackout part 1: the theory
• Voltage, inertia and the Iberian blackout part 2: faulty PV inverter crashed the grid
• Ghosts on the grid: why the phantom concept of vars risks our energy security
• Location, location, location: managing voltage in weak grids

The event was triggered by a faulty solar inverter, but this is almost irrelevant to what happened. Faults will always occur on the grid. Whether they cause blackouts depends on what happens next. In Spain, there were multiple failures – grid operator Red Eléctrica de España (“REE”) adopted static responses to a dynamic voltage disturbance, and had not instructed sufficient conventional generation to be ready to provide voltage support. By the time it realised the need, gas plants required several hours to warm up and synchronise with the grid, collapsing before this was able to happen.

Both conventional and inverter-based generators failed to deliver voltage support in line with grid code obligations. However, the real trigger for the blackout was the response to a large disconnection of embedded solar in response to within-day price signals – when prices went negative at the same time as the disturbance was ongoing, the grid frequency fell, but large amounts of wind and solar tripped off, failing to ride though a relatively modest reduction, in violation of grid code requirements. It was this failure and the consequent loss of generation that caused the cascade of trips that led to the full collapse of the Spanish and Portuguese grids and full islanding of the Iberian peninsular.

A few months later, REE requested emergency powers to prevent further blackouts as it had concerns over voltage management in the south of the country. Like many grid operators, and indeed many market participants, REE appears to have forgotten that “reactive power” and “vars” have no physical reality – they are a mathematical construct that represents complex numbers on the energy system.

Grid operators, regulators, and everyone else need to drop this language because it has led to a commoditisation of something that cannot really be commoditised – not all provision of voltage support is equal. Synchronous machines do it automatically be virtue of their electromagnetic coupling with the grid in the same way their mechanical inertia supports frequency. Inverter-based generators can only support voltage by diverting current away from powering loads, which in the limit, could cause a catastrophic collapse in grid frequency.

So far, grid-forming power electronics have not been deployed outside a few test cases and are not actually forming the grid anywhere in the world. In Australia, arguably the country that has done the most work on this, it has been acknowledged that trying to use inverters in this way presents significant dangers to grid stability and that the economic frameworks have yet to be developed. For inverters to provide grid support, they have to reduce the electricity they sell to customers, reducing their income, so they need much higher remuneration for this than synchronous generators do, which provide significant grid support automatically, for free, simply because of their physical and electromagnetic characteristics.

I have been criticised in the past for scaremongering when warning of the risks of blackouts and that blackouts will lead to fatalities. Ten people lost their lives in Spain and one in Portugal due to the direct effects of the blackout (one death in a house fire linked to the use of candles, seven fatalities when backup generators powering ventilators failed, and three deaths when a backup generator powering a ventilator malfunctioned, giving off carbon monoxide, poisoning the patient and two relatives). However, later research found that “in Spain, mortality rose over the next two days (+167 deaths; 95% Credible Intervals: +28 to +300), particularly among women aged 85+. No clear excess was found in Portugal.”

In addition, the economic cost was significant , with one estimate putting it between €2.25 – 4.5 billion.

Warnings about gas grid infrastructure are the icing on the cake

The year ended with NESO warning that an accelerated decline in North Sea oil and gas infrastructure might force premature shutdown of sections of offshore pipeline infrastructure, leading to potential gas shortages on cold winter days. This built on earlier work by Westwood Global Energy Group for Offshore Energies UK. It is possible that issues with offshore gas pipeline viability could emerge as early as the winter of 2026/27.

My last post of 2025 covered this risk. We have only two independent pipelines bringing Norwegian gas to the beach in the UK – the rest all connect into the UK’s offshore pipeline infrastructure, so this shutdown risk impacts not only the UK’s own gas supplies but also supplies from Norway. It also makes the UK much more exposed to the consequences of bad actors attacking these pipelines, since if parts of the UK offshore network close, these pipes (Langeled and Vesterled) will acquire even greater strategic importance.

In early December, Ofgem published its final RIIO-3 determinations. National Gas had the largest proportional increase in costs since the draft determinations, by 30%, but still has the largest gap between business plan and allowed costs at 21%. National Gas operates Britain’s primary energy system with nearly 5,000 miles of high-pressure pipeline running from Scotland to Devon. It serves half a million businesses, over 30 power stations, and more than 24 million homes.

“We are also concerned that significant reductions have been made based on partial reviews. Around just half of our IT projects have been reviewed, for example. Reductions to our gas quality, metering and telemetry submitted costs were apparently made before Ofgem had undertaken a detailed technical review – i.e. before it had a sufficient evidential basis to make these reductions.”
– National Gas response to Ofgem’s draft RIIO-3 determination

National Gas had been highly critical of the draft determination, saying that many of the cost elements had been reduced versus the business plan in an arbitrary way. While non-load capex and opex have both been increased in the final determination, versus the draft determination, they are still well down on the business plan. Indirect costs, which include IT and cyber resilience costs as well as system operation and planning support costs, have also been slashed.

National Gas may well appeal to the Competition and Market Authority, although Ofgem has expanded the scope for re-openers so National Gas may feel it can secure additional funding as the price control period progresses.

We’re going to need gas both for home heating and for power generation for decades to come. Playing fast and loose with critical gas infrastructure is a major policy blunder which presents serious risks to energy security.

Political and financial headwinds are likely to drive what happens next

What happens next depends on whether policymakers, regulators and grid operators recognise these risks and act on them appropriately. There are some encouraging signs from Ofgem, which, at least privately, is weighing these risks. Unfortunately, neither NESO nor DESNZ seem to be taking them particularly seriously, continuing to rush towards Clean Power 2030 despite the costs and risks to energy security. The Labour Government remains committed to killing off the North Sea despite widespread criticism from everyone including normal allies such as the unions. Privately, Labour MPs express concerns, but so far, Energy Secretary Ed Miliband is ignoring them.

But the wider political environment for the Government is challenging. There is a general expectation that Labour will do very badly in this year’s local elections, which may well lead to a change in Prime Minister, with key Cabinet members also being replaced. The rumour mill suggests Miliband will not be safe from this pressure.

The economic environment will also be key. The UK’s public finances remain under strain. Public sector net debt excluding banks is near 96% of GDP, a level last seen in the early 1960s, and public sector net financial liabilities are similarly elevated. Borrowing has continued to outpace receipts, with deficit figures for 2025-26 well above typical pre-pandemic levels and broad annual borrowing in the tens of billions of pounds. The Office for Budget Responsibility (“OBR”) in 2025 described the UK’s fiscal outlook as “increasingly vulnerable by both historical and international standards”.

Bond markets have reflected this vulnerability. In September, the cost of new 10-year government borrowing was 4.6% in Britain, compared with 4.0% in the US, around 3.5% in France and Italy, 3.2% in Canada, 2.7% in Germany, and just 1.6% in Japan. 30-year gilt yields reached their highest level since 1998. Recent increases in the cost of government borrowing have been described as being “consistent with an emerging fiscal crisis.

“We’re in an environment where Western democracies seem incapable or unable to address the underlying issues that are causing debt to keep on rising. So, a debt crisis, I would say, is likely by the end of this decade, but we might start to see that, as we started to see with France a few months ago, factored into market thinking much sooner.”
– Gerald Lyons, chief economic strategist at Netwealth

Some economists believe the UK and other major developed economies are heading towards a sovereign debt crisis as they struggle to control spending. “Six of the G7 are heading for a debt trap by the end of the decade,” Gerald Lyons, chief economic strategist at Netwealth, told the Edelman Smithfield Investor Summit this month.

France’s fiscal backdrop is more acute than in the UK. Public debt is well above 100 % of GDP – in some estimates, climbing past 115% of GDP – and the budget deficit in 2025 is one of the highest in the eurozone at around 5–5.4 % of GDP. Even the French finance minister has said he could rule out the need for an IMF bailout. Political gridlock has exacerbated the problem. In late 2025, French lawmakers had to resort to an emergency rollover budget law to keep the government funded into 2026 after failing to agree on a full budget due to deep divisions in parliament.

France’s deep welfare commitments and structurally high public spending make debt consolidation politically difficult, and the absence of a politically credible challenge to the welfare model leaves its fiscal trajectory unresolved. With debt high, deficits persistent and budget passage increasingly fraught, 2026 is likely to see France move closer to a sovereign debt stress scenario, even if formal default is avoided through ECB intervention.

Germany’s fiscal situation in 2025 sits somewhere between the UK and France: not crisis-level but in transition. The country has fundamentally reformed its constitutional debt brake to allow higher borrowing for defence and infrastructure, and the government is expanding public investment significantly, with new funds of the order of hundreds of billions of euros being discussed.

The debt-to-GDP ratio is projected to rise through the decade as a result. According to Germany’s Stability Council, debt could reach above 80% of GDP by 2029, exceeding standard EU thresholds and necessitating an “escape clause” to avoid disciplinary action under EU fiscal rules. While some macroeconomic forecasts suggest deficits will moderate later in this decade, these forecasts may be overly optimistic.

Germany’s industrial base is still exceptionally energy-intensive. High energy prices have eroded competitiveness, shifted cost structures relative to peers, and accelerated the offshoring of heavy industry. These pressures are structural rather than cyclical, meaning they are not easily reversed by short-term fiscal stimulus. Even if fiscal measures support demand domestically, they do not fundamentally fix energy cost disadvantages in global markets (which remain a primary driver of German industrial strength).

US tariffs and global trade frictions continue to weigh on Germany’s export performance, making a healthy recovery reliant on external rather than internal demand, which is a significant structural vulnerability. Mechanisms such as the Carbon Border Adjustment Mechanism (“CBAM”), intended to level the playing field on carbon costs, also have inflationary implications for German industry as they raise costs for carbon-intensive imports, may not fully protect domestic producers and could further dampen domestic demand if manufacturing margins are squeezed.

The IMF and other observers warn that higher debt levels must be paired with structural reforms or Germany risks weak growth and entrenched underperformance relative to peers. Germany has now experienced four years with no economic growth, with the IMF saying it needs to slash “excessive red tape” and “boost productivity and entrepreneurship”.

The EU’s CBAM (also being adopted in the UK) risks compounding Europe’s economic problems rather than solving them. By raising the cost of foundational inputs such as steel, cement, fertiliser and electricity, the CBAM is inherently inflationary and feeds through into construction, infrastructure and food prices at a time when growth is already weak.

While framed as a tool to protect European industry, the CBAM does nothing to address the core competitiveness issue of high energy costs and instead risks accelerating deindustrialisation through demand destruction, higher compliance burdens and trade retaliation. In practice, it looks less like a climate solution and more like a mechanism for exporting Europe’s policy costs, only for them to return via higher prices and weaker growth.

And key trading partners are not responding by lowering the carbon intensity of the goods they export to Europe. India has rejected EU proposals to increase its domestic carbon tax in response to the CBAM and has indicated it may pursue World Trade Organisation (“WTO”) action. South Africa is also considering making a formal complaint at the WTO. China is vocally critical of the CBAM and is calling for multilateral discussion on its compliance with WTO rules.

For large, advanced, non-EU economies, CBAM is not strong enough on its own to force industrial transformation, nor is it credible as an external constraint in the way EU policymakers often assume. Even where countries appear to be adapting, they are motivated by their own internal climate goals rather than compliance with EU policy objectives.

Trouble is brewing

My theory about what we are seeing: after the end of the Cold War, we began to take security for granted, believing that existential threats had receded. Despite the emergence of terrorist threats post 9/11, this line of thinking does not appear to have changed the driving philosophies of Western governments in the post Cold War era. The lack of existential threat meant that attention could be diverted to “doing things nicely”, leading to an explosion in regulation and a debt-funded expansion of the state to execute these nice things (huge growth regulators, major expansion of benefits and welfare and so on).

After decades, the affordability of this debt is starting to erode, while the red tape has strangled growth and created economic stagnation. On top of this, decision-making has been delegated to non-governmental bodies, reducing accountability and making policy change harder. This post cold war model is running now out of steam – bond markets are signalling the end of this particular road. Political elites are responding by increasingly trying to control speech, a typical sign of political models in decline.

The question is whether the reset can be peaceful, avoiding extreme financial hardship. I guess that in some countries there will be unrest (e.g., France) and in some there will be real hardship (could be any of us).

What does this mean for energy? Meaningful changes in the direction of policy will be hard without governance reform, but maintaining the current direction of travel will be financially constrained as well as limited by the physical realities of our energy systems.

Original article   l   KeyFacts Energy Industry Directory: Watt-Logic