Phoenix 24
Incorrect estimates triggered panic without affecting actual invoices.
Seattle | July 2026
Amazon Web Services has corrected a global billing-system failure that displayed wildly inaccurate cost estimates to customers, with some accounts appearing to owe billions or even trillions of dollars despite having generated only minimal cloud-service expenses.
Infobae reported that the incident affected the AWS Billing and Cost Management Console, Cost Explorer and related usage reports. Customers began receiving alerts after the platform’s estimated billing system assigned incorrect prices to their activity, transforming ordinary monthly spending into impossible financial obligations.
The malfunction began during the night of July 16, when the AWS console started presenting inaccurate estimated billing data. The company later identified the origin as a unit-pricing problem inside the subsystem responsible for calculating projected costs.
AWS emphasized that the displayed amounts did not represent actual usage or finalized charges. Customers were not required to make payments, modify their accounts or suspend their cloud services while engineers investigated the failure.
One of the most widely reported cases involved Bill Radjewski, who operates a college football data platform. His AWS expenses had reportedly remained close to one cent per month, yet he received an alert suggesting that his accumulated usage had exceeded $1.5 billion and that his next invoice could surpass $3 billion.
Other customers reported estimated costs of $22 billion, $75 billion and $110 billion. One account displayed approximately $7.1 trillion in usage charges, a figure greater than the market value of many of the world’s largest corporations combined.
A software developer with a monthly spending limit of $36 saw an estimate of approximately $34 billion. The contrast demonstrated that customer-defined budgets and normal usage patterns had not caused the numbers. The calculation system itself was applying incorrect pricing information.
The erroneous estimates created immediate anxiety because AWS budget alerts are designed to warn customers when cloud spending exceeds established limits. Businesses depend on those notifications to detect misconfigured servers, compromised credentials or unexpected increases in computing demand.
An abnormal cloud bill can sometimes indicate that unauthorized users have accessed an account and activated expensive infrastructure. Customers receiving the alerts therefore had legitimate reasons to investigate urgently, even when the amounts appeared mathematically impossible.
AWS initially paused estimated billing calculations and attempted to reverse a recent change to the computation subsystem. When the first rollback did not fully resolve the discrepancy, engineers pursued additional mitigation measures and began restoring the last accurate billing information.
The company subsequently recomputed affected cost and usage data to replace the inflated figures. Throughout the process, AWS maintained that the incident concerned estimates shown within its management tools rather than the underlying records used to determine final invoices.
That distinction protected customers from the most serious potential consequence. The system displayed extraordinary liabilities, but it did not actually withdraw those sums or finalize them as legitimate charges.
The event nevertheless exposed how strongly organizations depend on accurate cloud-cost information. AWS provides infrastructure for companies, public institutions, developers, media platforms, financial services and nonprofit organizations around the world.
Many clients operate under a pay-as-you-go model in which computing, storage, database activity and data transfer are measured continuously. Billing dashboards therefore function as operational controls rather than simple monthly statements.
A small error in unit pricing can become enormous when multiplied across millions of transactions, processing hours or data records. In this case, the calculations created numbers disconnected from actual consumption but still visible through official account interfaces.
The incident did not cause a general interruption of websites or applications hosted on AWS. Computing services continued operating, distinguishing the event from a conventional cloud outage involving inaccessible servers, databases or network resources.
Its principal impact was financial uncertainty and psychological distress. Customers responsible for company budgets were forced to determine whether the notifications reflected fraud, runaway infrastructure or a platform-wide malfunction.
The episode also generated humor across social media as users compared their fictional debts. Some joked that they had briefly become responsible for sums larger than national economies, while others criticized AWS for allowing automated alerts to distribute such alarming information.
Behind those reactions was a serious issue of trust. Cloud customers delegate not only technological infrastructure to service providers but also the measurement of the resources for which they must pay.
Unlike a physical product with a visible price, cloud spending can fluctuate according to processing time, storage volume, network traffic, software requests and regional pricing. Clients must therefore rely on the provider’s internal systems to calculate costs accurately.
AWS documentation explains that monthly charges remain estimates until the billing period closes. The platform recalculates usage throughout the month so customers can monitor trends, but the final invoice is issued only after the period ends.
This structure prevented inaccurate dashboard information from automatically becoming a completed financial obligation. It also gave AWS time to restore the correct estimates before billing was finalized.
The failure illustrates the importance of separating estimated costs from actual invoices. Businesses receiving an unexpected alert should verify account activity, examine service-level usage and consult the provider’s operational-status information before making major changes.
Security precautions remain necessary because genuine cloud-cost anomalies can result from exposed credentials, cyberattacks or incorrectly configured automated systems. The extraordinary scale and simultaneous global appearance of these estimates, however, pointed toward a centralized calculation error.
AWS’s response included suspending the affected computation process, attempting a rollback, restoring known accurate data and recalculating the estimates. Customers were repeatedly informed that no action was required.
The incident also demonstrated why automated financial systems require safeguards capable of identifying implausible outputs before notifications reach users. An account that normally spends cents or a few dollars should trigger internal validation before receiving a multibillion-dollar projection.
Such controls could compare estimated charges with historical patterns, account limits and the maximum realistic cost of the resources in use. Exceptional increases would still be investigated, but customers could receive a warning that the data was under verification rather than a definitive-looking total.
AWS remains the world’s largest cloud-infrastructure provider, and the scale of its operations means that even a temporary billing discrepancy can affect customers across industries and countries.
The malfunction was resolved without turning the fictional figures into real debts. Yet the experience showed that in digital infrastructure, an incorrect number can produce genuine fear long before money changes hands.
Phoenix24 | In the cloud, trust depends on every calculation. En la nube, la confianza depende de cada cálculo.