Skip to content
Briefing

Microsoft 365 outage: Insurance and reinsurance implications

Disruption to Exchange Online began on 12 November. Some users could not access Outlook mailboxes or receive emails promptly. Teams calendars and invitations were affected, although most calls, meetings and group chats remained available. Some Microsoft 365 Copilot users could not obtain responses or access work data. In its 09:03 GMT update on 13 November, Microsoft reported early signs of stabilisation, with capacity still below normal. Microsoft said it was not aware of malicious activity causing the incident and would publish its findings on the cause after recovery.

HKEX is reported as saying that its email service was affected while its markets continued to operate normally. In Australia, underwriting agency Coversure reported difficulty communicating with brokers and the London market about business due to be bound that day or the next. It used telephone calls and Teams to handle instructions.

This article considers the potential insurance and reinsurance implications of the outage.

When does an outage produce an insured loss?

For policyholders and claims teams alike, the first question is whether any available policy is triggered.

Under conventional property/business interruption cover, insured physical damage is commonly required. Supplier and utilities extensions may also be damage-based. Non-damage extensions must be checked against their specific triggers: a cloud-service outage does not automatically satisfy a denial-of-access or service-interruption extension.

For cyber policies, consider whether cover extends to a non-malicious system failure at a service provider or requires a security failure or cyberattack. Cover for the insured’s own systems may not extend to a dependent provider’s systems.

For each potentially relevant cover, one should check the insured systems and providers, any direct-contract or named-provider requirement, exclusions, waiting period, monetary retention and sublimit. Establish when any waiting period starts and whether, once it is exceeded, cover includes losses incurred during that period or only losses incurred afterwards. Check the applicable indemnity or restoration period, as well as cyber/data and infrastructure exclusions, including any exceptions that preserve cover.

The next question is what the business actually lost. For example, an online retailer whose cloud-hosted order-management system becomes unavailable may lose sales or incur additional overtime costs to process orders manually. Sales delayed should be distinguished from sales lost, taking account of later fulfilment, cancellations and costs saved under the policy’s loss formula and indemnity or restoration period. Reasonable and necessary mitigation costs may be covered even where the sale was preserved, but only where the relevant cover provides for them, subject to any requirement for prior consent, any cap linked to the loss avoided and any sublimit.

Did the instruction take effect in time?

Brokers, coverholders and underwriters should review instructions sent or received close to an inception, renewal or facultative placement deadline. Who had authority to bind? Had the parties agreed terms by telephone or on a placing platform? Were any conditions still outstanding? Under English law, a late written confirmation may record a contract already in force, subject to any applicable formalities. If cover was missed, an investigation into possible professional indemnity exposure should establish what cover could have been obtained, on what terms, whether the client would have accepted it and whether it would have responded to the particular loss.

Similar questions arise for claims notices. Check the prescribed recipient, permitted means of service, reporting deadline and any deemed-receipt provision, particularly under claims-made cover. Preserve the original message, delivery trace, platform timestamps and call notes, and seek acknowledgement through an agreed alternative channel. Microsoft said it was extending its email timeout window to 48 hours to preserve queued messages. That technical measure does not itself extend a contractual deadline or establish effective notice.

Cross-border programmes

For a cross-border programme, establish which entity used which Microsoft 365 environment, contracted with which provider and actually sustained the loss. For example, a mainland Chinese parent may procure the software, a Hong Kong office may receive orders and a UK subsidiary may bear the lost margin. Check each entity’s insured status, the local policies and any difference-in-conditions (DIC) or difference-in-limits (DIL) cover under the master policy. A group-level loss or accounting allocation alone does not establish a covered loss.

Third-party liability

A separate liability claim could follow if an IT provider had undertaken to maintain a back-up service and its customer suffered loss when that service failed. The service contract, including any liability cap and exclusions, helps determine the provider’s liability. Whether its technology errors and omissions policy responds is a separate question.

Insurers’ own operations

Insurers may themselves have had difficulty receiving notices or accessing claims files. They should keep a record of alternative reporting channels, affected customers and urgent matters awaiting action. A material disruption to policyholder service may also engage regulatory reporting duties. UK firms should promptly assess the impact under the applicable Financial Conduct Authority (FCA) and Prudential Regulation Authority (PRA) rules; some reporting requirements call for immediate notification.

Reinsurance

For cedants and reinsurers, recovery depends on the treaty’s scope, exclusions, attachment point and aggregation wording. An ‘event’ and an ‘originating cause’ can group losses differently. A common provider alone does not establish a single aggregating event or cause.

This remains a developing incident. Microsoft’s root-cause findings, the sequence of recovery across services and evidence of transactions ultimately lost or completed may change the analysis of individual claims and whether losses can be aggregated. Time-sensitive notifications and evidence preservation should proceed while those questions remain open.

(Update: This article was prepared based on publicly available information as at the morning of 13 November. Microsoft 365 services have since been restored. Subsequent developments do not affect the principal insurance and reinsurance issues discussed in this article.)

Published
16 November 2026
Reading Time
6 minutes