When the cloud closes

A payment can fail without a customer becoming poorer on paper. The account may contain ample money, but the card terminal, transfer instruction or cash machine cannot reach the system needed to use it. The distinction between owning money and accessing it became uncomfortable for customers of Revolut on October 20th, when the digital-finance group’s status page showed partial unavailability across services including card payments and cash withdrawals.

Google Cloud separately reported problems in its London region, describing elevated errors and intermittent network failures affecting multiple services. Later that day Google attributed the hours-long disruption to a “power event” in its europe-west2 region and said its engineers had restored full service availability. Reuters linked Revolut’s disruption to the cloud incident, citing a source familiar with the matter; Revolut said its customers’ money and data remained safe. The public notices establish simultaneous operational trouble; they do not, by themselves, provide a complete technical explanation of the connection.

That qualification matters. A single incident, explained so far in one phrase, is a poor basis for declaring a particular engineering design defective. Nor does disruption establish insolvency, missing customer money or a cyberattack. The more useful financial question is how firms and their suppliers make an essential service available when something important underneath it stops working.

Money with an entrance

For a customer at a checkout, a balance is a promise of purchasing power only if the payment can be authorised. For a business, access may determine whether a supplier is paid or an employee receives money on time. A temporary loss of access can therefore have economic consequences even when the underlying claim on the institution is unchanged.

The scale of those consequences cannot be inferred from a status-page warning. Partial unavailability does not mean that every customer, transaction or country is affected. A failed attempt may succeed later or through another method. Without reliable transaction data, estimates of money stranded or purchases lost would be guesswork. Revolut’s notice is evidence of disruption, not a ready-made measure of its cost.

Still, this is a form of financial dependence that conventional balance-sheet measures capture poorly. Capital can absorb losses. Liquid assets can help meet withdrawals. Neither automatically restores an unavailable application or a broken network connection. Operational resilience belongs alongside those safeguards because it protects the route by which customers exercise their financial rights.

The shared supplier

Cloud computing offers financial firms a plausible bargain: sophisticated infrastructure without constructing and maintaining every component themselves. Outsourcing, however, changes where a dependency sits. It does not remove it. The Bank of England’s account of critical third parties identifies the systemic concern: institutions can rely on a small group of shared providers, so a disruption at one supplier can reach several firms.

Britain has already moved beyond treating this exclusively as a matter for individual procurement departments. Earlier this year, its financial regulators began overseeing designated critical third parties. The Financial Conduct Authority’s published list includes the relevant European entities of Amazon Web Services, Google Cloud and Microsoft, alongside Oracle’s British entity. The oversight concerns services whose disruption could affect the financial system.

This is not a regulatory warranty that a supplier will never suffer an outage. Nor does an incident alone prove a breach of the rules. The point of direct oversight is that an individual bank may lack the information or influence needed to assess a shared dependence with consequences extending beyond its own customers. Regulators can examine risks at the supplier as well as obligations at the financial firm.

The FCA is explicit that firms retain responsibility for managing their third-party relationships. Designation does not transfer a bank’s duties to the technology company or to its supervisors. That allocation matters: otherwise, an institution could treat the presence of regulation as a substitute for understanding how its own essential services work.

A backup that works

Google’s incident notice suggested moving workloads to another available zone or region where feasible. The qualification is doing real work. Recovery requires an organisation to know which components can move, what other services they need and whether the resulting system can process transactions correctly. A second location on a diagram is not evidence that a business service will continue under pressure.

The FCA’s operational-resilience framework takes the customer-facing service as its starting point. Firms must identify important services, understand their dependencies and test their ability to remain within tolerable disruption under severe but plausible scenarios. The purpose is practical: discover where a technically elaborate recovery plan fails to deliver a usable outcome.

For payments, an editorial implication is that testing should ask more than whether a screen comes back online. Can a customer complete the transaction? Are pending instructions reconciled? Can support staff give a reliable account of what has happened? These are questions to test, not allegations about Revolut’s handling of this incident; the public evidence does not settle them.

Communication is part of the service too. A useful notice distinguishes functions that are unavailable from those still operating and avoids implying certainty about an unresolved cause. Now that service has returned, a clear account of what failed and what will change would be more informative than an assertion that systems are back to normal.

The October disruption will need a fuller technical record before firm conclusions can be drawn about its cause or total impact. Its immediate lesson is narrower, but substantial. Financial technology can make money move with remarkable ease. The responsibility to keep that movement possible remains with the firms whose names customers see when it stops.