Google Cloud restored services in its London region on October 20, 2026, after a power-related incident disrupted banking, public-sector and commercial systems across the United Kingdom for several hours. Google’s public status dashboard recorded the incident in the europe-west2 region as closed after seven hours and 52 minutes. The company said full service availability had been restored and that it would publish an analysis after completing its investigation.
The disruption demonstrated how a failure inside one cloud region can spread far beyond the technology provider itself. Reuters reported problems affecting financial institutions and public services, while user reports also pointed to interruptions involving Lloyds Bank, HSBC, Revolut, Rightmove, DPD, NHS England and Royal Mail. Not every reported problem was necessarily caused by Google Cloud, and the scale of impact varied between organisations.
What happened in London?
Google attributed the event to a “power event” at a third-party facility serving its europe-west2 cloud region. That region is made up of multiple data centres and availability zones designed to serve customers in Britain and elsewhere in Europe. Google’s status page listed 38 affected products, including AlloyDB for PostgreSQL, Apigee and BigQuery Data Transfer Service.
The official dashboard did not initially provide a detailed root-cause report. A power event can affect electricity supply, cooling, networking or equipment availability, but no specific technical chain was confirmed in the public material available on October 20. There was also no indication in the cited reports that the outage resulted from a cyberattack. The most accurate description remains the one supplied by Google: a disruption associated with a power event at a third-party facility.
Banks and public services reported problems
Revolut said several services were disrupted, including card payments and cash withdrawals at ATMs. The digital bank said customers’ money and data remained safe. Lloyds expected its services to return to normal after the root issue was resolved, while HSBC acknowledged that some customers had difficulty accessing services.
The disruption also reached systems outside banking. HM Courts and Tribunals Service reported an issue with the Cloud Video Platform used for remote hearings and said it was working with technology partners to restore service. The Guardian reported temporary problems with the NHS app, while Royal Mail said it was aware of technology issues but that letter and parcel deliveries were not affected. These distinctions matter: an online service interruption does not automatically mean underlying records, funds or physical operations were lost.
Why the outage matters
The incident arrived as British and European regulators increased scrutiny of financial institutions’ dependence on a small group of large cloud providers. In July, the United Kingdom placed Google Cloud, Microsoft, Amazon Web Services and Oracle under direct oversight as critical third-party suppliers to the financial sector. The goal is to understand and reduce the risk that a single provider failure could affect many regulated firms simultaneously.
Google Cloud has also been designated by European regulators as one of the critical computing providers used by the finance industry. The Financial Conduct Authority said it was engaging with Google and affected companies to understand the impact. That review may clarify whether contingency plans, regional redundancy and communication procedures worked as intended.
For businesses, the outage is a reminder that availability zones and backup systems only reduce risk when applications are designed to use them effectively. Important services may require tested failover between zones or regions, independent communications and procedures for operating during cloud interruptions. Google’s forthcoming incident analysis will be important for determining the exact cause and which resilience measures could prevent a repeat. Until that report is published, claims about a deeper technical failure would be speculative.
