Amazon Web Services said Monday evening that its cloud infrastructure had fully returned to normal after a widespread outage disrupted thousands of companies and several major websites for much of the day. The failure, which originated in AWS’s main U.S. East-1 region in northern Virginia, caused hours of downtime across critical internet platforms including Snapchat, Roblox, and Duolingo.
The disruption began shortly after 3 a.m. Eastern time, when AWS reported technical problems tied to its Domain Name System, or DNS, which translates website names into IP addresses. The issue affected DynamoDB, a core database system used by numerous online services. At its peak, the outage impacted more than 70 AWS services and left an estimated 1,000 companies struggling to operate.
By late afternoon, Amazon said it had resolved the problem and that “all AWS services returned to normal operations” around 6 p.m. Eastern time. The company added that some systems were still clearing message backlogs but that performance was largely restored. Amazon plans to release a detailed post-event summary outlining the cause of the failure.
The outage rippled across industries and continents, taking down government sites in the United Kingdom, online banking services, and popular consumer apps. Downdetector, which tracks user reports of service disruptions, showed spikes of outages for Hulu, Lyft, Disney+, Reddit, and The New York Times. In the United States, Amazon.com itself experienced temporary issues, along with internal systems used by its warehouse and delivery staff. Some Amazon workers reported being instructed to stand by as internal logistics and payroll systems went offline.
Financial institutions were also affected. Lloyds Banking Group in the U.K. confirmed disruptions to online services, though most were restored within an hour. Airlines including United and Delta experienced brief system problems that prevented customers from checking in or managing bookings online.
Technology analysts say the outage underscores the fragility of global dependence on a small number of cloud providers. AWS controls roughly one-third of the world’s cloud infrastructure market, according to Synergy Research Group, ahead of Microsoft and Google. Millions of companies rely on AWS to run web servers, store data, and deliver digital services.
Experts said the problem appeared to be technical rather than the result of a cyberattack. Rob Jardin, chief digital officer at NymVPN, described it as likely stemming from a network overload or failure in a key data center component. “Because so many systems are connected to AWS, a single malfunction can cascade quickly across the internet,” he said.
AWS has faced major outages in the past, including one in 2021 that halted deliveries and crippled streaming platforms, and another in 2023 that disrupted websites for several hours. Monday’s incident followed a series of recent failures across the tech industry, including a global Microsoft crash in July 2024 that shut down Windows systems and grounded flights.
Industry experts said the event highlights the concentration of digital risk. Mike Chapple, an information technology professor at the University of Notre Dame, noted that services like DynamoDB are “the record-keepers of the modern internet.” He added that while the data itself appears secure, “the problem arose in the systems that tell other applications where to find that data.”
