Amazon Web Services has an AWS billing glitch on its hands: one that conjured phantom invoices ranging from a few million dollars to, in at least one reported case, $1.5 trillion for cloud services customers had never actually used.
The incident began at 1:33:36 on 17 July 2026, according to the AWS Health Dashboard, when the Billing Console began displaying incorrect estimated charges. Amazon said it started seeing inaccurate billing data from late Thursday, and by Friday morning had confirmed that an attempted rollback ‘did not resolve the issue.’
The bug sits in what Amazon calls its billing computation subsystem. It doesn’t just inflate raw usage figures: according to Mezha.net, the glitch also incorrectly displayed charges for used services and credits, meaning the phantom numbers weren’t a single simple multiplication error but something messier underneath.
From Bangalore to Bolsover: The AWS Billing Glitch Goes Global
The reach of the problem was wider than a single region or account type. The Guardian reports the disruption hit customers globally, from Bangalore to Bolsover, with at least one account shown a bill of $1.5 trillion. The snippet from TechCrunch cited a high of close to $2.5 billion based on Reddit screenshots; The Guardian’s higher figure may reflect a different customer’s alert or a later stage of the incident. Both outlets agree the numbers were entirely fabricated.
Among those caught in it: Dan Harvey, head of marketing at Learning Through Landscapes, a Hampshire-based charity that runs a school grounds audit app. His AWS account normally costs less than £1 per month. Harvey told The Guardian he ‘almost have a heart attack’ before the tech support team looked into it.
That’s the part the marketing decks never cover. AWS’s customer base isn’t only enterprise architects running petabyte-scale data lakes. It includes small charities, students, and solo developers whose monthly bills round to the price of a flat white. For those users, a phantom invoice at any scale, let alone one with twelve zeroes, is a genuinely alarming way to start a Friday.
What Amazon Has Said, and What It Hasn’t Fixed Yet
Amazon’s position, stated clearly on the status page: the estimates ‘do not reflect actual usage and charges.’ Nobody is actually being billed a trillion dollars. The relief is real, even if the reassurance arrives via a status page update rather than a direct customer communication.
What Amazon has been less clear about is the fix timeline. The original rollback failed. As of the status page update, engineers were evaluating whether to resume estimated billing computations after internal monitoring indicated the billing computation subsystem had been restored. The issue was expected to persist for several more hours.
Mezha.net notes that Amazon has stated it is taking steps to prevent similar incidents in the future and to ensure billing accuracy. That’s the standard post-incident language, and it’s fine as far as it goes. The more pointed question is how a change to the billing computation subsystem passed internal checks and reached production in a state that could generate trillion-dollar phantom invoices.
AWS is a core profit engine for Amazon and one of the largest cloud platforms on the planet. Its billing infrastructure is not an afterthought. That a single subsystem change could corrupt estimated charges globally, for accounts of every size, and resist an initial rollback, suggests the post-mortem will be an interesting read when it eventually surfaces.
For now, customers who woke up to a multi-billion (or multi-trillion) dollar bill can breathe out. Watch for Amazon’s full incident report, and specifically whether it explains why the first fix didn’t work.
