KOHEN AVOCATS FIRM · PARIS 16
Is Revolut down? Google Cloud outage, blocked money and missed payments: what Irish customers can claim and recover (2026)

Hassan KOHEN acts in urgent matters, from the police station to the assize court. Initial strategic analysis: €80 incl. VAT, personal reply within 24 hours.

100% confidential · Legal professional privilege · No obligation

Article generated by artificial intelligence, through a process designed and supervised by the firm

Source: French Court of Cassation – Judilibre and Légifrance open data

Paris Bar Police custody, investigation, assize court CNB listing avocat.fr
Maître Hassan KOHEN, avocat au Barreau de Paris
Hassan KOHEN
Attorney at the Paris Bar

Is Revolut down? Google Cloud outage, blocked money and missed payments: what Irish customers can claim and recover (2026)

Speak to a lawyer 06 89 11 34 45

On the morning of Tuesday 20 October 2026, thousands of people across Ireland opened their Revolut app and found a dead screen: cards declined at the till, cash withdrawals impossible, transfers frozen, and sometimes no way even to sign into the account. Ireland’s national broadcaster RTE reported that Revolut’s 3.4 million Irish customers were affected, with card payments, transfers, cash withdrawals and signing into accounts all disrupted. The cause was not an incident at the bank itself. Earlier that morning, Google Cloud had reported an issue with multiple services at its London data centre complex, region europe-west2; by Tuesday evening, Google said a power event at a third-party facility had been resolved and services restored. The BBC described customers unable to access funds abroad and at least one buyer struggling to complete a property purchase, while a small food business in England explained that it could neither process customer orders nor book parcels for collection, losing a full day of sales of perishable produce. Revolut stated that services were fully restored and that customers’ money and data remained safe throughout. Regulators reacted immediately: the UK Financial Conduct Authority said it was engaging with Google and affected firms, and RTE recalled that Google Cloud was among the technology companies designated by EU regulators as critical third-party providers on which the bloc’s finance industry depends. The French press covered the same outage on the same day for French users. Services are back, but the legal questions remain: when your money is unreachable for a day, who is responsible, does a power cut at an American cloud giant excuse a European payment provider, and how do you claim for a missed payment or a proven loss? This article answers each question with the EU and French rules that protect account holders, because those French provisions transpose the European directive that binds payment providers across the Union.

I. Your payment was blocked or never arrived: the payment provider is the first port of call

A. Card refused, cash unreachable, transfer frozen: the duty of proper execution

When a card payment is declined, an ATM refuses a withdrawal, or an online transfer stays pending for hours, the starting point is simple: the institution that holds your account owes you proper execution of your payment orders. European payment law, transposed in France in the Monetary and Financial Code, places that responsibility squarely on the payer’s provider until the money reaches the beneficiary’s provider. Article L133-22 of the Code monétaire et financier makes the payer’s provider, subject to the reserves stated in that article, «responsable de la bonne exécution de l’opération de paiement à l’égard du payeur jusqu’à réception du montant de l’opération de paiement» (i.e. its payment service provider is responsible to the payer for the proper execution of the payment transaction until the amount is received by the beneficiary’s payment service provider). Where the provider is responsible for a badly executed transaction, the same article continues: «il restitue sans tarder son montant au payeur. Si besoin est, il rétablit le compte débité dans la situation qui aurait prévalu si l’opération de paiement mal exécutée n’avait pas eu lieu» (i.e. it refunds the amount to the payer without delay and, where needed, restores the debited account to the position that would have prevailed if the badly executed transaction had not taken place). In plain terms, a frozen transfer is not merely a technical incident for which nobody answers: the payer’s provider must put the money back and repair the account.

Speed is part of the obligation. For ordinary transfers, the amount must reach the beneficiary’s provider quickly: article L133-13 of the same code states «Le montant de l’opération de paiement est crédité sur le compte du prestataire de services de paiement du bénéficiaire au plus tard à la fin du premier jour ouvrable suivant le moment de réception de l’ordre de paiement» (i.e. the amount of the payment transaction is credited to the beneficiary’s provider account no later than the end of the first business day following receipt of the payment order). A full day of paralysis therefore already exceeds the normal execution time, and every additional hour of delay counts against the provider, not the customer.

Crucially, the customer does not have to prove what went wrong inside the machines. Where a user denies having authorised a transaction or states that it was not properly executed, article L133-23 reverses the burden of proof: «il incombe à son prestataire de services de paiement de prouver que l’opération en question a été authentifiée, dûment enregistrée et comptabilisée et qu’elle n’a pas été affectée par une déficience technique ou autre» (i.e. it is for the payment service provider to prove that the transaction was authenticated, duly recorded and entered in the accounts and that it was not affected by a technical or other deficiency). The words matter: a cloud failure at a subcontractor is precisely a “déficience technique” whose absence the provider must establish, and the provider’s own logs alone are not necessarily enough. During the October outage, a customer who saw no pending incoming transfer while the app’s help bot insisted nothing was coming is exactly the person this rule protects: it is for Revolut, Lloyds, HSBC or any other provider to demonstrate that each transaction was authenticated, recorded and unaffected by the incident.

One deadline governs everything and must be treated as absolute. Article L133-24 requires the user to report an unauthorised or badly executed transaction « sans tarder » (i.e. without delay) and «au plus tard dans les treize mois suivant la date de débit sous peine de forclusion» (i.e. no later than thirteen months after the debit date, failing which the claim is time-barred). Thirteen months sounds generous, but the forclusion is automatic, and the practical advice is to complain in writing on the day itself: screenshots of the error, the exact time of the failed payment, the reference of any chat with support, and a copy of the provider’s own status page. These elements fix the date, the scope of the failure, and the link between the outage and your loss.

B. The money left but never arrived: follow the chain of liability to the right defendant

The October outage produced a second, quieter type of victim: the person waiting for money. The BBC quoted Rachael Ellen, a 25-year-old Revolut customer from Glasgow, who could neither pay for her parking nor see an expected transfer from a relative, while the app’s help bot told her no incoming payment existed. When the missing money was sent by someone else, from another bank, the legal question becomes: whom do you sue, your own provider or the sender’s? The answer follows the money along a chain that the courts apply strictly.

The Court of Appeal of Paris gave that chain its clearest recent expression on 10 July 2025, in a case about a 500 euro transfer sent from a BNP Paribas account that never appeared on the claimant’s ING account (Paris Court of Appeal, 10 July 2025, RG 24/07558, official judgment). The court held: «Il résulte de ce texte que la société BNP Paribas était responsable de la bonne transmission du virement jusqu’à sa réception par la société ING Bank laquelle était ensuite responsable de sa bonne exécution sur le compte de Mme [N].» (i.e. It follows from this text that BNP Paribas was responsible for the proper transmission of the transfer until its receipt by ING Bank, which was then responsible for its proper execution on the claimant’s account.) The payer’s provider answers first, up to receipt by the beneficiary’s provider; only then does the beneficiary’s provider take over. Suing the wrong link in the chain fails, which is exactly what happened to the claimant: «Il appartient donc au premier chef à Mme [N] qui n’a mis en cause que le prestataire de service de paiement du bénéficiaire à savoir la société ING Bank de prouver non seulement qu’elle a donné un ordre de virement à la société BNP Paribas mais encore que cet ordre a bien été correctement exécutée par cette banque qui l’a correctement transmis.» (i.e. It is therefore primarily for the claimant, who sued only the beneficiary’s provider, to prove not only that she gave a transfer order to BNP Paribas but also that the order was correctly executed and transmitted by that bank.) She could not, and the court confirmed the dismissal of all her claims and ordered her to pay costs. The lesson for October’s victims is direct: if you sent the money, complain to your own provider and demand the proof of transmission, such as the initial instruction and the Swift confirmation, which the beneficiary’s bank in that case had expressly requested; if you were waiting for money, ask the sender to obtain that same proof from their provider, because without it no court will move the liability down the chain.

A frequent objection during outages is the wrong account identifier. When panic sets in, users retype IBANs, approve duplicates, or validate a transfer twice. Article L133-21 draws a sharp line: «Si l’identifiant unique fourni par l’utilisateur du service de paiement est inexact, le prestataire de services de paiement n’est pas responsable de la mauvaise exécution ou de la non-exécution de l’opération de paiement. Toutefois, le prestataire de services de paiement du payeur s’efforce de récupérer les fonds engagés dans l’opération de paiement» (i.e. Where the unique identifier provided by the payment service user is inaccurate, the payment service provider is not liable for non-execution or defective execution. However, the payer’s provider shall make reasonable efforts to recover the funds involved.). A mistyped IBAN during the outage therefore shifts the regime: no automatic liability, but a legal duty to try to recover the funds, and a duty for the beneficiary’s provider to cooperate by sharing the information needed for recovery. Keep the exact identifier you entered, the timestamp, and any confirmation screen: it decides which of the two regimes applies to you.

Late execution has its own remedy. Where the money arrives eventually but too late, for example a transfer credited a day after the deadline while a contractual payment was due, the provider responsible for the delay must correct the value dating so the account looks as it would have looked without the delay. The Paris case above confirms the mechanism article by article, and the practical consequence is that fees charged because of the delay, such as overdraft interest debited while the expected funds were invisible, must be reversed once the provider’s responsibility is established. Again, everything turns on dated evidence: the moment the order was given, the moment the funds should have been available, and the charges posted in between.

II. Who pays for the damage: outage, force majeure and the method for claiming

A. Does a power cut at Google excuse your bank? The narrow path of force majeure

The most tempting defence for a provider is also the weakest against customers: our subcontractor failed, an event beyond our control, force majeure. The Monetary and Financial Code anticipates exactly this argument and frames the exemption narrowly. Article L133-5 provides: «La responsabilité prévue aux sections 2 à 9 du présent chapitre ne s’applique pas aux cas de force majeure, ni lorsque le prestataire de services de paiement est lié par d’autres obligations légales prévues par des législations nationales ou communautaires.» (i.e. Liability under sections 2 to 9 of this chapter does not apply in cases of force majeure, nor where the payment service provider is bound by other legal obligations under national or EU legislation.) Force majeure is therefore a possible exit, but it must be proven against the classic three-part test of French law: an event beyond control, unforeseeable, and irresistible. The general contract rule states the same idea in one sentence at article 1231-1 of the Civil Code: «Le débiteur est condamné, s’il y a lieu, au paiement de dommages et intérêts soit à raison de l’inexécution de l’obligation, soit à raison du retard dans l’exécution, s’il ne justifie pas que l’exécution a été empêchée par la force majeure.» (i.e. The debtor is ordered to pay damages for non-performance or delay unless it proves that performance was prevented by force majeure.) It is the provider, not the customer, that must justify the excuse.

Applied to October 2026, the excuse looks fragile. A power event in one zone of a London data centre, as Google described it, may well be beyond the bank’s direct control, but the dependence itself was known, documented and supervised. RTE reported that the United Kingdom had brought Google Cloud, alongside other hyperscalers, under direct regulatory oversight as critical third-party suppliers to the financial sector in July, precisely because an outage at one provider could paralyse several firms at once, and that EU regulators had designated Google Cloud among the critical technology providers on which the bloc’s finance industry depends. A risk identified by two sets of regulators, covered by a dedicated European regulation on digital operational resilience in the financial sector (Regulation (EU) 2022/2554 on digital operational resilience, the DORA Regulation), and materialising through a single point of failure is difficult to present as unforeseeable. Foreseeability does not mean the provider must prevent every incident; it means the provider must have built continuity, backup and communication plans proportionate to a risk the whole sector had mapped. The underlying European payment rule confirms the strictness: under Directive (EU) 2015/2366 on payment services (PSD2), the payer’s provider is liable for non-execution or defective execution and must refund, with only narrowly defined abnormal and unforeseeable circumstances as an escape.

Restitution and compensation must also be kept apart, because they obey different conditions. Getting your money back when a transaction failed follows the automatic mechanism of article L133-22 described above. Obtaining damages for the consequences, a missed property completion, a lost day of sales of perishable goods like those of the English produce company quoted by the BBC, penalty interest charged by a creditor, requires proving a fault, a loss and a causal link. Article 1240 of the Civil Code states the principle in a single line: «Tout fait quelconque de l’homme, qui cause à autrui un dommage, oblige celui par la faute duquel il est arrivé à le réparer.» (i.e. Any act that causes damage to another obliges the person by whose fault it occurred to repair it.) Silence can be that fault. On 1 August 2025, the Paris judicial court in interim proceedings ordered Banque Populaire Rives de Paris to close the accounts, return more than 193,000 euros, and pay a 4,000 euro provision for the proven harm, after the bank had blocked the customer’s card on a false pretext and refused for months to close the account and return the funds (Paris judicial court, interim order of 1 August 2025, RG 25/54874, official order). The judge found: «le fait qu’elle n’ait apporté aucune explication tant à M. [U] qu’à son conseil est fautif. » (i.e. the fact that it gave no explanation either to the customer or to his counsel is wrongful.) A bank that leaves customers facing a dead app, contradictory bot messages and no information about their money takes the same risk: the outage may be Google’s, but the unexplained silence toward the account holder is the provider’s own. Conversely, the Paris Court of Appeal case above shows the other side of the coin: the claimant who proves neither the transmission nor any quantified loss is dismissed and pays costs. Documented loss decides everything.

B. Claiming without losing your rights: the method that preserves every option

First, report in writing without delay, and in any event within thirteen months of the debit. The thirteen-month forclusion of article L133-24 runs «sous peine de forclusion» (i.e. failing which the claim is time-barred), which means a late complaint can extinguish even a perfectly valid claim, so the report should go through the provider’s official complaint channel on the day itself, with the screenshots, timestamps and support references already gathered. Ask expressly for three things: confirmation that the incident affected your account, the technical explanation recorded for your transactions, and the refund or re-execution of each failed or late operation with corrected value dating. Because the burden of proving proper execution lies with the provider, a precise written request forces it to produce the authentication and recording evidence the law demands, rather than a generic apology.

Second, quantify every euro of loss separately from the refund of the transaction itself. The refund restores the account; the loss is everything around it: overdraft interest posted while funds were invisible, late-payment penalties charged by a landlord, supplier or tax authority, the extra cost of an emergency alternative payment, and for professionals the margin lost on orders that could not be processed or dispatched, documented by order logs and courier records as in the produce company’s case. Courts require proven, not hypothetical, harm, and the dismissed Paris claimant is the warning: gather invoices, bank statements showing the charges, and any correspondence proving the missed deadline. Where a direct debit was taken twice or at the wrong time during the chaos, note also that authorised payee-initiated payments carry their own eight-week refund window under article L133-25 of the Monetary and Financial Code (L133-25 refund rules), a separate track worth checking on your statements.

Third, escalate in the right order. If the provider’s final answer is unsatisfactory, incomplete or simply never comes, the file goes to the competent out-of-court complaints body and then to the financial regulator, with the full dated bundle attached. The October incident already drew regulatory attention: the UK Financial Conduct Authority publicly engaged with Google and the affected firms, and the European framework requires financial entities to classify, report and follow up major ICT-related incidents under the digital operational resilience rules. A complaint that cites the exact transactions, the provider’s contradictory messages and the quantified loss is the file regulators can actually use. Where the amount justifies it and the liability is not seriously disputable, interim court proceedings can order restitution and even a provision for the obvious part of the harm, as the Paris judge did with the blocked account and the 4,000 euro provision; that route is for clear-cut files only, and it rewards the customers who built their evidence on day one.

Conclusion

The Google Cloud power event of 20 October 2026 lasted hours; its legal consequences last much longer. Services are restored and the providers state that money and data stayed safe, but that statement does not settle individual files: a payment that failed or arrived late must be refunded and the account repaired, the provider must prove its systems were not deficient, a thirteen-month forclusion punishes late complaints, and a cloud failure at a subcontractor is not an automatic excuse where regulators had mapped exactly that risk. The customers who will recover in full are the ones who reported in writing immediately, kept every timestamp and receipt, directed each claim at the right link in the payment chain, and separated the automatic refund from the proven additional loss. Outages are collective; successful claims are individual, and they are won on paper.

📄 Official circular