A significant failure of Amazon Web Services (AWS) on Monday, October 20, 2025, caused a ripple effect throughout the entire digital universe affecting the work of dozens of high-profile organizations, including financial trading platforms and popular smart home appliances. The incident, which was based on technical malfunction in an important region of the US-EAST-1 of AWS, acted as a powerful wake-up call to the world regarding the heavy dependence of the modern internet on big cloud providers.
The outage started soon after 3.11 AM ET (7.11 UTC) and the outages rapidly jumped on outage monitoring tools such as Downdetector. AWS established that the problems were in its huge US-EAST-1 zone, which is in Northern Virginia.
The Technical Trigger: A DNS Failure in DynamoDB
The engineers of AWS were fast to respond and in several hours the company located the possible root cause, which is a particular problem with the Domain Name System (DNS) resolution of the DynamoDB API endpoint in US-EAST-1.
DynamoDB, the highly managed NoSQL database of AWS is one of the building blocks of numerous applications. The failure of the DNS, which is more or less the phone book of the internet, to convert the name of the DynamoDB API endpoint to its appropriate network address disastrously caused a huge choke point.
This fundamental failure rapidly made a digital domino effect. Due to the fact that many of the global AWS services are supported by DynamoDB, and the US-EAST-1 region, in general, this outage spread much further than the region itself. It also impacted global services and features based on US-EAST-1 endpoint, including IAM (Identity and Access Management) updates and DynamoDB Global Tables. Since IAM controls authentication and authorization access to practically all AWS services, its failure resulted in mass paralysis of applications that rely on it.
Blast Radius: Trading to Smart Homes
The list of platforms affected looked like a list of day-to-day digital life, which describes the tremendous concentration risk involved in the cloud ecosystem.
 |
| (https://downdetector.com) |
Social and Communication
There were serious problems with service in Snapchat, Signal, Zoom, or Slack.
Finance and Trading
Necessary financial applications were brought to their knees, which include crypto exchange Coinbase, trading platform Robinhood, and peer-to-peer payment service Venmo. The big banks in UK such as the Lloyds Bank, Halifax and the Bank of Scotland were also affected as well as the government services such as the HM Revenue and Customs (HMRC) site.
Gaming and Entertainment
Gaming platforms such as Fortnite and Roblox removed users, and streaming companies, such as Prime Video and Hulu, also claimed problems.
The Amazon Ecosystem
The services of Amazon itself were not protected. The primary Amazon.com retail site and Prime Video were slowed down and people said Alexa smart speakers and ring doorbells systems had ceased operation; both systems are based on fundamental cloud connections. Exasperated users of smart homes were also reported and jokingly said that because of the loss of service of AWS, they have to turn on or turn off their lights by hand.
Market Calm Amidst Chaos
Although the operational mess was immense, the financial effect on the stock of Amazon (AMZN) was not substantial. The stock, which started its pre-market at a low point, ended at a loss of only 0.68 percent of $213.04. Analysts interpreted this quietness in the market as a signal of investor trust in the overall strength and dominance of AWS in the market despite its hiccups in operations.
More importantly, speculation was soon debunked by cybersecurity experts that it was a state-sponsored cyberattack, and this news had gone around social media (partly due to the geopolitical tensions). Experts unanimously indicated the DNS/DynamoDB problem as an internal IT infrastructure failed.
In further development of the morning, AWS made "initial mitigations" and announced indicators of major success, suggesting that customers should re-attempt any unsuccessful requests as the systems cleared a backlog. The event highlights a sobering truth to companies: depending on a single cloud vendor, even the largest in the world, requires solid multi-region or even multi-cloud resiliency plans to protect the core services of the company against single points of failure.