The sorry tale of new Drax OCGTs highlights growing capacity market irrationality

The GB Capacity Market was introduced with a clear and repeatedly stated objective: to ensure security of electricity supply by bringing forward new capacity where the market would not do so on its own. In particular, it was meant to underwrite the construction of new large-scale dispatchable plant, particularly large gas plant, to endure security of supply during the transition to a lower-carbon system. Unfortunately it has been an abysmal failure in this regard with almost no new large gas plant delivered through the scheme (Carrington got an early capacity contract but AFTER it took FID). The market has been dominated by existing generators, with the new assets being primarily small gas and diesel engines until tighter emissions rules limited their participation.

.

This week I was at the Powergen conference in San Antonio, Texas, and I met with the US representatives of a UK based engineering company, SSS Clutch, which sells clutches that can enable generators to operate as synchronous condensers. When I met them at last year’s conference they mentioned that their equipment was installed at the three new Drax OCGTs which had capacity contracts starting October 2024-25. These are all 299 MW units, coming in just below the carbon capture readiness threshold.

As last year’s conference was in February 2025 I expected them to be running, it being four months after the start of the first delivery year of the capacity contracts. But they weren’t. Nor are they running now although one of the three, Hirwaun is currently commissioning. Neither of the other two is yet on the system (Millbrook and Progress Power). The plants were all built on time but are not running. The reason: no grid connection. But here’s the kicker – under Capacity Market rules, this does not only delay receipt of capacity payment, it triggers non-delivery penalties, potentially wiping out the entire capacity revenue for the year.

“Drax expects to take commercial control of the first of the three OCGTs (Hirwaun Power) in Q1 2026. The unit is now commissioning and receiving capacity market payments. The second and third sites are expected to commence commissioning in 2026.This is later than originally planned, primarily due to delays in grid connection by the relevant authorities,”
– Drax trading statement, December 2025

This produces a result that offends ordinary legal instinct, violating the principles of fairness that usually underpin English law. A company builds what it was asked to build, yet is penalised because a third party – the transmission owner – has not delivered the infrastructure required for it to operate. And Drax has no contractual remedy since the Capacity Market was deliberately structured to place this risk with the generator.

This is not simply a contractual dispute

Capacity Agreements are, in a formal sense, contracts under English law, which leads people to seek the familiar doctrines of prevention, frustration, change of circumstances, or even penalties. However, none of these applies in this case, because Capacity Agreements are not freely negotiated bilateral contracts. They are statutory instruments, created under secondary legislation, incorporating the Capacity Market Rules as amended from time to time, and explicitly subordinated to regulatory change. Public rather than private law applies.

The contractual promise is not “build a power station and you will be paid”. It’s: “if, during the Delivery Year, you are available to provide capacity in accordance with the Rules, you will receive capacity payments”. Availability, in turn, assumes a working grid connection – if the connection is missing, the Capacity Market does not recognise partial performance. From the scheme’s point of view, a fully built but unconnected plant is indistinguishable from one that does not exist.

English contract law is unforgiving in this case – there is no general doctrine of “market change” or “economic hardship”. A risk that was foreseeable in type if not in scale remains where the contract places it, and connection delay has always been a known category of risk, even if its current severity was not anticipated in 2014. Nor does frustration help, as the Capacity Market rules explicitly contemplate non-delivery, penalties and termination. Where a contract allocates the consequences of failure, frustration does not apply.

So, this is not a bad contract that a court can fix, it’s a regulatory scheme doing exactly what it was designed to do, even if it now appears grossly unfair.

Is there any legal recourse for Drax?

The Capacity Market rests on a set of factual assumptions:

• that grid connections, while slow, are ultimately delivered
• that new-build plant is financeable under the Capacity Market risk allocation
• that penalties incentivise performance rather than simply punishing the unlucky

Those assumptions are now under strain. Connection delays are no longer isolated, they are systemic, long and widely recognised to be a major problem with multiple initiatives designed to address the issue across Ofgem, NESO and the network operators. In some cases, connection dates are not only uncertain but effectively open-ended. Meanwhile, developers report that new large gas plant is no longer bidding in the capacity auctions, not because it isn’t needed, but because the Capacity Market represents risks that are unfinanceable once connection risk and penalty asymmetry are priced properly.

This matters because public authorities are not entitled to operate statutory schemes in a state of denial. If the factual premises on which a policy rests have collapsed, continuing to apply it unchanged can become irrational in terms of public-law. Irrationality is a grounds for the High Court to intervene if a Judicial Review is sought.

Irrationality does not mean unfairness, it means incoherence. For example, if the Government made possession of cheese illegal but later introduced regulations requiring every adult to consume 100 g of cheese per day, that second requirement would be irrational if the first rule remained in force. It would be impossible to comply legally with the cheese-eating obligation if having cheese in one’s possession were a crime.

In such a case, a Judicial Review (“JR”) could be sought and the High Court would be likely to rule that the cheese-eating requirement was incompatible with the existing legal framework, and require the Government to reconsider its position – either by abandoning the consumption requirement or repealing the criminalisation of cheese possession. A requirement to eat cheese every day would not be irrational in itself. The irrationality is conditional: given the continued ban on cheese ownership, it is irrational to insist that people must eat it.

Judicial Review is not about fairness in the everyday sense – courts do not intervene simply because a scheme produces harsh outcomes, but they will where a decision-maker acts in a way that is no longer rationally connected to the statutory purpose.

The Capacity Market’s purpose is to ensure security of supply through the procurement of deliverable capacity. If:

• new dispatchable plant is deterred from bidding
• existing Capacity Market Units (“CMUs”) cannot connect through no fault of their own, and
• the State continues to procure “capacity” that it knows may never be deliverable

then the scheme risks becoming self-defeating.

At that point, the argument is not “this is unfair to Drax”, but “it is irrational to continue running capacity auctions on unchanged terms when the mechanism no longer induces the capacity it was designed to procure”. Which is a very different claim.

The reductio ad absurdum is, what if the connection is never built…the most powerful way to expose the problem is to push it to its logical extreme. In this case, that is to suppose the transmission owner never delivers the grid connection, and the generating equipment is forever stranded, gathering dust until the developer dismantles it and sells it for parts.

Under current rules, the generator will accrue penalties until the annual cap is reached. After sustained non-delivery, the Capacity Agreement will be terminated without compensation.

From a policy perspective, this is perverse. Not only will the Capacity Market have procured nothing, but an otherwise perfectly usable plant would go to waste. From a legal perspective, it exposes a contradiction that a scheme intended to procure capacity is indifferent to whether the infrastructure required to make that capacity real is ever delivered. Once non-delivery ceases to be a risk and becomes a known impossibility, continuing to auction on the same basis looks increasingly irrational.

The Capacity Market Register shows that Long Stop Dates for the Millbrook and Progress units have been extended into 2026 and 2027 respectively. Such extensions prevent automatic termination of the Capacity Agreements, but they do not make the capacity deliverable, nor do they solve the underlying problem. The Capacity Market is still relying on capacity that cannot yet appear on the system, and whose delivery depends on grid connections that remain outside the scheme’s control.

The Capacity Market Register contains no indication that these units are subject to any bespoke treatment beyond extended Long Stop Date – they appear to be administered under the standard Capacity Market rules despite the absence of a grid connection.

Treatment of interconnectors provides evidence of internal inconsistency         

Interconnectors are allowed to participate in the Capacity Market,with the owners of the interconnectors being the Capacity Market participants. However, the owners of interconnectors do not operate them – they sell the capacity to the market. And the buyers of that capacity can choose to flow electricity in either direction with the owner having no control except if there is a question of safety.

The availability test for an interconnector is simply whether it is available for use. There’s no way to ensure that it’s importing. This means during a system stress event, they could actually be exporting, exacerbating the market shortage. In reality, the interconnector would likely be suspended in this situation to avoid a cross-border bidding war for energy, but there is no mechanism to require the buyer of capacity on an interconnector to flow electricity into the GB market. Neither the asset owner, the system operator, the regulator nor the government has the powers to compel imports of anything into the UK.

So the Capacity Market accepts probabilistic deliverability for interconnectors, justified by averages and modelling, while demanding absolute physical deliverability from domestic generation. That asymmetry was defensible when domestic capacity was assumed to be firm and interconnectors contributed a marginal amount, but it becomes much harder to justify when domestic firm capacity is being deterred and interconnectors make up a growing share of the market.

This does not prove illegality on its own, but it is strong evidence that the Capacity Market no longer has a coherent or consistent concept of what “deliverable capacity” means.

What might a court decide to do?

A successful JR would not lead to damages or retrospective compensation as courts dislike unwinding the past, but they could:

• quash the decision to hold or clear auctions on current terms
• declare that procuring new-build CMUs without credible connection deliverability is unlawful
• set aside existing Capacity Agreements prospectively, bringing penalties to an early end
• require ministers to reconsider Capacity Market design in light of current grid reality

Any such relief would almost certainly be prospective rather than retrospective, but that is not trivial – ending future penalties and forcing market redesign would materially change the economics of affected projects.

This matters beyond its potential impact on Drax. If the Capacity Market cannot bring forward new dispatchable capacity because connection risk and penalty exposure make participation irrational, then it’s failing in its core function. Continuing to run it regardless is not simply inefficient, it may be unlawful. English public law will accept harshness, but will not tolerate a scheme that has lost any rational relationship with its statutory purpose. At some point, the courts may be asked to decide whether that status has been reached.

Drax appears to be confident its grid connections will be completed in the next year, and has not suggested any plans for a judicial challenge. But that may be a bad thing for the market as a whole if the penalties Drax is exposed to deter other developers from pursuing projects. As my recent electrification report highlights, the GB market is in dire need of new gas-fired generation capacity, so hamstringing the very mechanism through which such capacity ought to be procured is hardly helpful.

Perhaps if Progress and Millbrook are still not running when next year’s Powergen conference comes round, they may take a different view.