War exclusion defeats Nord Stream’s claim for a €580m indemnity

Following a six-week trial, the High Court has dismissed Nord Stream’s €580m claim against its insurers arising from the sabotage of its gas pipelines, holding that a market-standard war exclusion applied. The decision is financially significant for London Market and international insurers and provides a useful re-examination of the English law approach to construing standard war exclusions, particularly where “directly or indirectly” causation language is used.

Key findings

• There were two separate limbs of exception in Exclusion 2.i, such that the exclusion would apply if either the ‘war exclusion’ or the ‘government act/order exclusion’ was met. The proviso “except as otherwise provided in Section I of the Policy” qualified only the ‘government act/order’ limb of the exclusion.
• GC9 was limited to pollution hazards and did not override Exclusion 2.i.
• A broad causal test applies to exclusions of an indirect nature. The legal test applied did not require that war was the “effective cause” or had a connection of “real significance”. The test does not impose any further legal requirement such that the act must have a “war like” aspect or create a “special danger”. It is sufficient that the cause is capable of being identified as having a contributing factor, such that the cause would be “significant”, “noticeable” or “specifically accountable”.
• For the exclusion to apply, it was not necessary in this case for the Court to determine who was the perpetrator. The Court needed to be satisfied that in each case, the perpetrator would have had the requisite causal connection to the war. If any of the four potential perpetrators (the US, Russia or the government of Ukraine or a substate actor of Ukraine) had carried out the act, the war would have been a “significant” cause.
• The exclusion applied to an act outside of the confines of the “theatre of war” (eg, Russia or Ukraine), against a non-military target in international waters, that was economically and strategically connected to the conflict.

Background

The Claimant, Nord Stream AG (owned by a Gazprom entity), is the operator of natural gas pipelines running through the Baltic Sea.

On or about 26 September 2022, the pipelines were damaged by explosions at locations approximately 6.5km apart, and three out of the four lines in the area were rendered inoperable. One line (Line 2) was also damaged (the timing of which was disputed) in the form of an indentation. For the purposes of Nord Stream’s claim Line 1 and Line 2 were relevant, and explosions on the ‘NS2 pipelines’ did not form part of the claim.

The German Federal Public Prosecutor General subsequently initiated an investigation into the explosions; it was agreed that documents disclosed to Nord Stream from that investigation’s file would be subject to confidentiality provisions.

Insurers’ primary defence was that Nord Stream’s claim was excluded by a war exclusion (Exclusion 2.i), being directly or indirectly occasioned by or a consequence of the war between Russia and Ukraine. Insurers had the burden of proving that the exclusion applied.

Construction of the policies

Exclusion 2.i

The relevant section of the exclusion provided that “…Notwithstanding anything to the contrary contained herein, this section does not cover loss or damage directly or indirectly occasioned by, happening through, or in consequence of war (whether war be declared or not), invasion, acts of foreign enemies, hostilities, civil war, rebellion, revolution, insurrection, military or usurped power or confiscation or nationalisation or requisition or destruction of or damage to property by or under the order of any government or public or local authority except as otherwise provided in Section I of the Policy.” [Emphasis added].

The Court was required to consider if there were two separate heads of exception in Exclusion 2.i, such that the exclusion would apply if either the ‘war exclusion’ or the ‘government act/order exclusion’ was met.

The Court confirmed that it should strive to reach a construction which does not find language superfluous, which is consistent with the approach in Hamilton Corporate Member Ltd v Afghan Global Insurance Ltd [2024] EWHC 1426 (Comm), where the Court gave meaning to all parts of the clause.

Nord Stream argued that damage “by or under the order of any government” only applied to damage which is “occasioned by, happening through or in consequence of” war and referred to a “monolithic list” of perils. This was considered selective and the use of the canon of construction of noscitur a sociis (each word takes its meaning from the words with which it is linked or surrounded) was rejected. Further, even on Nord Stream’s construction it appeared to acknowledge a natural break in the clause at “confiscation”, which would result in the clause being divided into two halves.

Insurers’ two-part construction was preferred. The Court found that there were two separate heads of exception in Exclusion 2.i, such that the exclusion would apply if on the facts, either the ‘war exclusion’ or the ‘government act/order exclusion’ was met.

As a result, the proviso “except as otherwise provided in Section I of the Policy” only applied to the second limb of Exclusion 2.i. Consequently, even if the Court was wrong about the application of GC9 being limited to pollution hazards, it would not have assisted Nord Stream in widening the first limb of Exclusion 2.i (the war exclusion). 

Institute Clauses

The policies incorporated a number of market-standard Institute Clauses intended to address offshore construction and operating risks, as well as specific exclusions including the war exclusion.

Nord Stream sought to initially argue that the ‘Institute Clauses’ effectively took precedence over the war exclusion, with reference to the means of providing the insuring clause. This position developed in oral submissions, to argue that the only thing resembling an insuring clause was found in the modified Institute Clauses.

The Court rejected this, holding that the policies were not limited to construction risks and had to be construed as a coherent whole. There was no basis in the wording for treating the Institute Clauses as overriding or displacing the express terms of the war exclusion.

General Condition 9

A further issue arose from the particular drafting of these policies, which incorporated General Condition 9 (GC9) alongside Exclusion 2.i introducing some confusion about which was to take precedence. This was a policy-specific point rather than a broader issue of market wording construction.

Insurers argued that GC9 was limited to pollution hazards. Nord Stream, by contrast, contended that GC9 had wider application and preserved cover notwithstanding Exclusion 2.i, relying on the wording: “Coverage provided by the above paragraph shall also extend to cover any other loss, damage, liability, cost or expense caused or inflicted by order of any governmental or regulatory body or agency.”

When dealing with these rival interpretations, the Court endorsed the approach taken in Hamilton v Afghan Global, namely that effect should be given to the clause as a whole wherever possible. It considered that Nord Stream's interpretation would leave little meaningful role for the surrounding pollution-related wording. By contrast, Insurers' construction gave GC9 a coherent function as a limited extension addressing losses arising from governmental measures taken in the context of pollution hazards.

The Court therefore held that GC9 was confined to pollution hazards and did not conflict with, qualify or override Exclusion 2.i.

Further, the Court held that even if it were wrong on GC9's limitation to pollution hazards, the clause would still not assist Nord Stream. On its proper construction, the phrase "governmental or regulatory body or agency" did not naturally extend to the army (or members of the army acting within the chain of command) or to a Ukrainian sub-state actor. Accordingly, GC9 could not be relied upon to circumvent the operation of the war exclusion.

Broad ‘indirect’ causation language in the war exclusion demonstrates its force

It was common ground that the conflict between Russia and Ukraine began on or around 24 February 2022 and satisfied the definition of ‘war’ under the policies.

Aside of the factual analysis, the question before the Court was what sort of causal connection was required under the policies by the wording “directly or indirectly occasioned by, happening through or in consequence of war”. It was accepted that this is a very broad phrase in issue.

Exclusions of an indirect nature employ a broad causation test, which denotes a looser causation connection than that of proximate or effective cause. As considered in Spinneys (1948) Ltd v Royal Insurance Ltd [1980] 1 Lloyd’s Rep 406, “there must be some limit on the application of the clause, for the chain of causation recedes indefinitely into the past. The draftsman must have intended to stop somewhere: and that place must be the point at which an event ceases to be a cause of the loss and becomes merely an item of history.”

In drawing this line, the Court confirmed that the legal test did not require that the cause must have a “war like” aspect or that of a “special danger”, it did not require that war was the “effective cause” or had a connection of “real significance”. It was sufficient that the cause was capable of being identified as having a contributing factor, such that the cause would be “significant”, “noticeable” or “specifically accountable”, a rather lower bar.

Causation – potential perpetrators and connections with the ‘war’

There were four possible perpetrators (the US, Russia or the government of Ukraine (or a substate actor of Ukraine)). Russia, the US and Ukraine have all publicly denied responsibility.

The issue that no one had admitted responsibility did not prevent the Court from making a finding on a causal connection. The Court did not have to determine which perpetrator carried out the act. The key question for the Court was whether it was satisfied that, in each case, that perpetrator would have had the requisite causal connection to the war (eg, the war was a “significant” cause of its actions).

The Court rejected arguments that the act was not linked to war as the pipelines were a non-military target in international waters, well outside the “theatre of war” and that there had been a longstanding animosity between Ukraine and Russia.

It was held that if any of the possible perpetrators carried out the act, the war would have been a “significant” cause of their actions.

Was the damage “by or under the order of any government”?

In light of the Court’s findings, this question was considered in the alternative, as the first limb of the exclusion was satisfied (any of the potential perpetrators on the factual conclusion were held to have a sufficient causal connection to the war).

Although no evidence of a direct order was before the Court, it was held on the evidence that in each case of the State perpetrators, if any of them were the perpetrator that the act would have been damage to property by the government. Further, were it necessary to decide on the issue of a Ukrainian substate actor, it was likely this was approved by the head of the Ukrainian army and would therefore, in this matter, fall within the phrase “by… any government”.

The Dent

One pipeline also sustained damage in the form of an indentation (the Dent). On the evidence, it was held more likely than not that the Dent was caused by an explosion as part of the same attack that caused the explosions/rupture damage.

Comment

Standing back, the conclusion of the case is unsurprising. Nord Stream sought to rely on evidence and submissions to suggest that the sabotage was not, as a matter of fact, sufficiently connected to the war. In our view, that argument always had an air of unreality. The attacks occurred against the immediate geopolitical backdrop of Russia’s invasion of Ukraine and targeted infrastructure of obvious strategic and economic significance. The Court’s approach reflects a commercially realistic assessment of that context: where the wording extends to loss “directly or indirectly” occasioned by war, it is difficult to divorce an act of this nature from the conflict which gave it its practical significance.

The judgment should not be seen as a radical expansion of English insurance law. It is better understood as an orthodox application of established principles of policy construction and causation to a major modern geopolitical loss.

The decision is nevertheless important for London Market insurers because it confirms the force of broad “directly or indirectly” causation wording in war exclusions. War need not be the dominant, proximate or effective cause of the loss. Nor must the relevant act itself be an act of conventional warfare. It is sufficient that the war is a significant, noticeable or specifically accountable cause of the act causing the damage.

The Court’s approach is also significant in rejecting attempts to confine the exclusion by reference to geography or target type. The pipelines were outside the immediate theatre of war, in international waters, and were not military assets. That did not prevent the exclusion from applying.

For insurers facing claims arising from Ukraine, the Middle East and other areas of geopolitical instability, the judgment reinforces the importance of close attention to the wording of war, terrorism, sabotage and political violence exclusions. It also underlines that government act/order wording may have a wider role than parties sometimes assume, although the result will remain highly wording-specific.

The decision sits comfortably with the Court’s approach in Hamilton v Afghan Global: market wordings should be construed as a coherent whole, with effect given to the language, structure and commercial purpose of the policy. The outcome in Nord Stream therefore strengthens, rather than unsettles, the existing direction of travel in English insurance coverage law.

Mills & Reeve’s political violence insurance team advises insurers and reinsurers on complex political violence, war and terrorism coverage issues, including policy construction, causation, exclusions and high-value cross-border disputes.

In re to Nord Stream AG v Lloyd’s Insurance Company S.A & Arch Insurance (EU) DAC [2026] EWHC 1685 (Comm).

Our content explained