Monday, October 20, 2025, began like any other day, but quickly turned into a headache for millions of people living in South America (and the rest of the world). Who didn’t try to pay for coffee with Mercado Pago’s QR code, send a Pic, or check their balance on Guala, only to find an annoying error message? Frustration grew on social media, with many users complaining about failures in payment chains, bank transfers, and online stores. If you didn’t have cash in your wallet that day, it was probably an uphill battle to run your usual errands.
However, we can’t blame the local financial app or the state of the internet in South America; the collapse was part of a problem with the internet infrastructure. That day, the invisible heart of the online world went down: Amazon Web Services (AWS). The server outage occurred in the AWS US-EAST-1 region, a crucial data center in North America, and reverberated around the world. A person trying to buy a coffee in Panama was just as affected as a photo editor trying to download a stock photo from Adobe Spain.
The Heart of the Cloud
AWS is the world’s largest cloud computing platform. It is a kind of “power plant” that stores data and runs the code for most of the websites and applications we use every day. Approximately one-third of the entire Internet depends on WS services to function. The culprit behind all this chaos was an internal error in the Domain Name System (DNS), which is the Internet’s phone book. Its job is simply to translate human-readable domain names into the numerical IP addresses that computers need to find the server.
When the DNS failed, applications couldn’t find where to go, even though their destination still existed and hadn’t been hit by a cyberattack or anything like that. An expert at the University of Notre Dame described it as “temporary Internet amnesia.” The outage affected key AWS services, such as its DynamoDB database.
The funny thing about all this is that Amazon is the US company that sponsors the most H-1B visas.
It has thousands of Indian software engineers in its ranks. The problem with it taking so long to fix the AWS issue is that on that very day they were celebrating Diwali, one of the most important Hindu celebrations of the year. It’s the Indian equivalent of a server crash on December 25… and no one being in the office to fix it.
The Mercado Pago Case
The servers are located in Virginia, United States, but there were immediate financial repercussions throughout Latin America. This is because almost all regional platforms are integrated into WS, and they were forced to suspend their services. The most visible case was that of Mercado Pago and its parent company, Mercado Libre. Users in countries such as Argentina and Brazil experienced failures at almost the same time as the chaos that ensued in the United States.
This caused widespread problems for those trying to make a payment, transfer money, or check their account balances. According to reports on Downdetector, the peak of failures in Mercado Pago was recorded around 12:43 p.m. (Argentina time). If you live as an expatriate or are a local there, you will surely realize that you could not even pay for a taxi with a QR code.
Fast bank transfers were non-existent, and even popular transfer apps such as Peaks in Brazil stopped working. Virtual wallets such as Ualá and Naranja X also joined the list of those affected. In general, all of MercadoLibre’s eCommerce services had problems even uploading products or completing purchase transactions. That day, no one was able to buy anything unless they had cash in their wallet.