In Nord Stream AG v Lloyd’s and Arch [2026] EWHC 1685 (Comm), the Commercial Court has given helpful guidance on indirect causation in the context of the operation of war risks exclusions, confirming that broad wording in those exclusions will give insurers significant scope to exclude damage and loss that occurs far away from any battlefield. In this article, partner Alexandra Lyons and knowledge development lawyer Tom Cusworth examine the decision and the implications for insurers and policyholders.

For policyholders, the judgment highlights the need for proactive risk management, particularly those with substantial physical assets, critical infrastructure, complex supply-chain networks or international operations in or around war zones. Policies do not always meet expectations in terms of coverage, and policyholders must pay close attention to exclusions for war, political violence, terrorism and state-backed activity, in particular the causation language used.

The exclusion wording at issue is not uncommon in the commercial insurance market. Insurers may seek to rely on the judgment when arguing for a broader application of war exclusions covering losses caused “directly or indirectly” by war or hostilities. The decision may consequently have wider relevance for claims linked to geopolitical events across different regions.

Background

The proceedings arose out of Nord Stream’s claim for approximately €570m under primary and excess offshore operating all risks policies (the “Policies”), following damage to its two gas pipelines in the Baltic Sea caused by unknown saboteurs in September 2022. In addition to the explosion damage, Nord Stream also claimed in respect of a separate indentation found on NS1 Line 2 during post-explosion inspections, the “Dent”. Together, the explosion damage and the Dent formed the “Damage” for which Nord Stream sought an indemnity.

The policies responded to operational risks but were written on a modified version of the WELCAR 2001 form, which is more commonly associated with construction risks. The insurers denied cover, relying principally on Exclusion 2.i in the policy wording. That exclusion removed cover for loss or damage “directly or indirectly occasioned by, happening through, or in consequence of war”, and also for damage to property caused “by or under the order of any government or public or local authority”.

The insurers’ case was that the Damage fell within Exclusion 2.i because it was causally connected with the war between Russia and Ukraine. They also alleged that the Dent had been caused by an explosive device and formed part of the same coordinated sabotage operation as the rupture damage. Nord Stream disputed the insurers’ reliance on the exclusion and contended that, if the Damage had been deliberately caused by a government, it was covered under a separate provision in the policy wording, General Condition 9, which addressed deliberate government damage.

Issues and decision

The insuring clause

A central issue in the judgment was identifying the relevant insuring clause, which would provide the necessary starting point for the coverage analysis.

The policy contained no express insuring clause (an alarming yet surprisingly common phenomenon). Nord Stream therefore argued that the incorporated Institute Clauses (the Institute Clauses for Builders’ Risks and Institute War Clauses Builders’ Risks) provided the relevant cover for war risks, and that the narrower war exclusion in those clauses should apply. Insurers argued that those clauses were not intended to cover operating pipelines but were limited to property under construction and floating assets. The policies should therefore be read as containing an implied ”All Risks” insuring clause, consistent with their description as “Offshore Operating All Risks”. The court agreed with insurers. The Institute Clauses could not be stretched beyond their ordinary meaning without effectively rewriting the contract, whereas an implied term for all risks cover would not contradict any express term of the Policies.

Exclusion 2.i

Exclusion 2.i contained the following wording:

“2.i. 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.”

The court had to determine whether this should be read as a single list or should be divided into two limbs (as the insurers argued): a war limb, and a separate government act/order limb. Relying on the Divisional Court’s examination of the principle of “noscitur a sociis” (literally, “a word is known by the company it keeps”) in FCA v Arch [2020] EWHC 2448, the judge agreed with insurers, holding that these were two distinct categories and determining that the exclusion could apply if either limb were satisfied.

General Condition 9 (“GC9”)

Nord Stream also sought to rely on General Condition 9, a ”DELIBERATE DAMAGE” extension, for coverage of the damage. GC9 provided affirmative cover for damage directly caused by any act or order of any governmental or regulatory authority … to prevent or mitigate pollution hazard, or threat thereof, resulting directly from damages to the property insured for which the Insurers are liable under this Policy. Nord Stream argued that the second paragraph of GC9, extending this cover to “any other loss, damage, liability, cost or expense caused or inflicted by order of any governmental or regulatory body or agency”, provided cover for deliberate government damage, even if unrelated to pollution.

The court rejected that argument, holding that GC9 had to be read as a whole and could not be detached from its pollution-hazard context.

Causation test

The central point of principle in the judgment was the causal threshold for the exclusion, and what the insurers were required to establish in order to show that the damage was “directly or indirectly occasioned by, happening through, or in consequence of war”.

It was common ground that the inclusion of the word “indirectly” meant that something less than a proximate causal link was required. Nord Stream argued for an “effective cause” or “real significance” test of causation between the loss and the war, whereas the insurers argued for a broader indirect causation test requiring only that war “contributed to” the loss.

Applying the test set out in other cases (including Crowden v QBE Insurance [2017] EWHC 2597 (Comm)), the court rejected Nord Stream’s case, finding that an “effective cause” test was close to that of a proximate cause. The court held the correct test was whether war was a “significant” (ie noticeable or specifically accountable) contributing cause to the damage or loss, but that it did not need to be of “real significance” as argued by Nord Stream. The court also did not consider that the test required any comparative ranking against other causes.

On the facts, the court recognised that there was no conclusive information on whether Russia, the US or Ukraine had inflicted the damage. However, the court held that it was unnecessary to determine which actor had carried out the sabotage, because on analysing each potential scenario, the war was a sufficiently significant cause of the damage for Exclusion 2.i to apply.

The Dent

Although the Dent was an alternative and freestanding factual issue, the court found that it was more likely than not that the Dent, too, was caused by an explosive charge dislodged from the same sabotage operation. It was therefore also excluded under Exclusion 2.i.

Key takeaways

The judgment provides a helpful clarification of the “significant cause” formulation set out in Crowden in the context of a broadly worded indirect-causation exclusion clause: an indirect cause need not outweigh competing causes, but it must be “significant”, meaning noticeable or specifically accountable as a cause of the loss. It should be noted that the causal language in this case (“directly or indirectly occasioned by, happening through, or in consequence of”) is not necessarily the broadest available. Contrast, for example, the words “in connection with”, which were found in Stonegate v MS Amlin [2022] EWHC 2548 (Comm) to connote only a loose causal connection, with no requirement that the cause be significant.

This serves a useful reminder for policyholders and their representatives to pay close attention to (and, if necessary, negotiate) the drafting of their policies, if there is a risk that this could impact on their coverage. Causal language in exclusions is often overlooked, but small differences in drafting can lead to dramatic variations in the scope of coverage provided for significant losses (€579m in Nord Stream, over £1bn in Stonegate).

The decision is of obvious relevance to operators considering coverage of losses arising from the ongoing Middle East conflict, who should carefully consider any exclusions for war, political violence, terrorism and state-backed activity in their policies, including the causation language used. The analysis is, however, not limited to coverage under war and related risks policies, and has broader application across all commercial lines of business.

At the time of writing, Nord Stream had been refused permission to appeal by the Commercial Court but was granted additional time to seek permission directly from the Court of Appeal.