Is Google Cloud Down? The GCP Outage That Froze Revolut Cards and What CTOs Should Fix Now

Yes, Google Cloud went down in London. A major Google Cloud outage hit the europe-west2 region on 20 October 2026, starting at 06:08 UTC (07:08 UK time). 

The GCP outage took down Revolut’s app and cards, and disrupted Lloyds, HSBC and other UK businesses. At 11:36 UTC, Google traced the outage to a power event in zone europe-west2-b and said it had been mitigated.

This guide covers what failed, why one London zone caused so much damage and what UK regulators expect. It also gives CTOs five practical steps to protect their cloud infrastructure before the next outage. The GCP outage tracker at the end is updated after every new incident.

Google Cloud London Outage Timeline: 20 October 2026

Which Google Cloud Services Went Down in London?

Affected Google Cloud Services and What They Break

Google confirmed problems across seven services in europe-west2. Together, they span every layer of a typical application stack.

Revolut, Lloyds, HSBC: UK Firms Hit by the GCP Outage

Revolut was hit hardest, with its app, cards, transfers, ATM withdrawals, currency exchange and share trading all affected. Lloyds, HSBC, Bank of Scotland, Royal Mail and Screwfix also reported disruption. Lloyds said similar issues were affecting other organisations, which suggests a common cause. So far, Revolut is the only company whose outage has been publicly linked to Google Cloud.

Revolut Outage: Why Cards Were Declined at the Till

In many banking outages, the app goes down, but cards keep working. On 20 October 2026, Revolut customers in the UK, France and Spain were turned away at supermarket tills and coffee counters.

Every card payment needs live approval from the bank’s own systems. When those systems are unreachable, the terminal declines the card, regardless of the account balance.

Revolut has more than 80 million customers worldwide, including 13 million in the UK. It confirmed that customer money and data stayed safe. Customers still couldn’t get to their own money, though, and at a checkout that matters more than the balance.

The CTO takeaway: Card authorisation is a revenue-critical path. If it depends on a single zone or region, a single cloud fault can prevent all customers from paying at once.

Why One GCP Zone Failure Spread Across europe-west2

A Google Cloud region is made up of zones, each with its own power, cooling and network. London’s europe-west2 has three. If one zone fails, the other two should keep your services running.

In practice, three weak points break that promise.

1. Resources Locked to One Zone

Many GCP resources are created in a single zone by default. When that zone fails, they fail with it. The infographic below shows the most common ones and their regional alternatives.

2. Shared Dependencies

An app can run in all three zones and still rely on one DNS setup, message broker or secrets store. If that component goes down, every zone that depends on it goes down too.

3. Failover Nobody Has Tested

Failover that has only ever existed on an architecture diagram tends to fail the first time it is needed. Only regular testing under real load shows whether it works.

Google has traced the outage to a power event in a single zone, europe-west2-b. The errors and network failures it reported still spread across the whole europe-west2 region. Its full incident report should confirm which of these weak points played a part.

Why Status Pages Lag Behind Cloud Outages

On 20 October 2026, Revolut’s status page showed all systems as “operational”, while customers reported failed payments on social media. The page was later updated to show an incident affecting sign-ins, card payments, top-ups, and ATM withdrawals.

Status pages are usually updated by hand, once someone has confirmed the incident. Until then, customers fill the silence themselves, in public. AI providers face the same delay, as Deployflow’s Claude outage tracker shows.

Cloud Concentration Risk: FCA, PRA and DORA Rules Explained

For regulated firms, a cloud outage quickly becomes a compliance question. A hyperscaler region failing is exactly the kind of “severe but plausible” event regulators expect firms to have planned for.

UK: FCA Operational Resilience Rules

Under the FCA’s operational resilience rules (PS21/3), firms must:

• identify their important business services
• set an impact tolerance for each one
• prove they can stay within it during severe disruption

The deadline to comply was 31 March 2025. Blaming the cloud provider won’t satisfy a supervisor.

UK: Critical Third Parties Regime

Since 1 January 2025, HM Treasury has been able to designate key suppliers to the financial sector as critical third parties. Under the joint regulator rules (PS16/24), regulators can then oversee them directly. Designation doesn’t move responsibility for resilience away from the firms that rely on them.

EU: DORA

The Digital Operational Resilience Act (DORA) has applied since 17 January 2025. It requires firms to manage ICT third-party risk, test their resilience and keep a workable exit plan for critical providers. It also sets strict deadlines for reporting major ICT incidents. Revolut’s EU banking business is licensed in Lithuania, so an outage like this one can fall under DORA as well as UK rules.

Regulated teams need infrastructure designed around these rules from the start. Deployflow’s fintech DevOps services are built for FCA-regulated environments.

The same risk applies outside finance. Retailers, logistics firms and SaaS companies with everything in one region of one provider should have that dependency on the board’s risk register.

Five Fixes CTOs Should Make After the GCP Outage

Each fix below can start this week. Assign an owner and agree on what “done” looks like before the next incident tests it for you.

Three Questions for Your Next Board Risk Review

1. Which of your important business services would stop if one cloud zone failed?
2. When did your team last prove failover works, and how long did recovery take?
3. Which of your critical suppliers run in the same cloud region as you?

The third question matters more than most boards realise. A firm can spread its own workloads perfectly and still go down because of a payments, identity, or messaging supplier that relies on a single region.

Multi-Zone, Multi-Region or Multi-Cloud: Which Does Your Business Need?

Each extra layer of redundancy adds cost and complexity. Match the level to what each service is worth to the business.

Choose multi-zone for every production workload. It protects against a single zone failing at low extra cost. Treat it as the baseline across your whole estate.

Add multi-region for the journeys that earn or lose money. Payments, login and checkout justify the medium-to-high cost of surviving a full regional outage. Most estates need it for only a handful of services. During the London outage, Google advised affected customers to move workloads to other regions where feasible. That advice only helps teams that built for it in advance.

Reserve multi-cloud for regulatory or exit requirements. Running across two providers is the most expensive option and the hardest to operate. It makes sense when a regulator or contract demands a credible exit route. 

If you are comparing providers, Deployflow’s AWS, Azure and GCP security comparison is a great place to start.

GCP Outage Tracker: Major Google Cloud Incidents

Newest incidents first. Each new major outage is added here.

Protect Your Business From the Next Google Cloud Outage

The 20 October outage won’t be the last. The firms that recover fastest are the ones that prepared before it happened.

Deployflow helps CTOs and IT leaders:

• Find single-zone risks. Deployflow’s cloud management services team maps your workloads against zone and region failures and delivers a prioritised remediation plan.
• Prove failover works. Deployflow engineers design failover, run game days and set reliability targets aligned with your impact tolerances.
• Spot problems early. Sizif adds AI-driven monitoring and Slack or Teams alerting to your existing Zabbix and Wazuh setup.

Book a free 30-minute consultation and find out which of your services would stop if a single cloud zone failed tomorrow.

Google Cloud Outage: Frequently Asked Questions