Amazon Web Services (AWS) began restoring operations following a widespread outage that disrupted some of the world’s most popular websites and applications, including Snapchat and Reddit.

AWS said it was seeing “significant signs of recovery”, noting that “most requests should now be succeeding” while the company worked through a backlog. However, issues persist in AWS’ US-EAST-1 region based in Northern Virginia, one of its largest global hubs. Although the company said it was “seeing early signs of recovery”, it confirmed that “multiple AWS services experienced network connectivity issues” and reported “significant API errors” in the region.

The global disruption, which began early today and lasted nearly three hours, affected more than 70 AWS services and led to widespread outages across financial, gaming and communications platforms. According to Reuters, Ookla said its outage tracking site Downdetector logged more than 4 million user reports globally. Platforms including Uber, Perplexity AI, Roblox, PayPal and Signal were affected, while UK users experienced disruptions to service from Lloyds Bank, BT, Vodafone, Ring and more.

Amazon’s own services, including its retail site, Prime Video platform and Alexa voice assistant, were also impacted.

Cloud backbone

E-Marketer senior analysts agreed that the event underscores the risks of concentrated cloud infrastructure and the need for greater resilience across digital ecosystems. Gadjo Sevilla said the outage “made evident how deeply modern commerce, communication and entertainment depend on a single cloud backbone”, warning that over-reliance on one provider “endangers brand reputation and customer trust.”

Indeed, Jacob Bourne warned that “as cloud reliance and workloads expand, these outages could hit industries harder”. Meanwhile, Grace Harmon forecasted that moving forward, businesses might “invest in contingency systems that let users quickly shift between providers like AWS and Microsoft”.

Although there is no evidence of a cyberattack behind the outage, Rafe Pilling, director of threat intelligence at Sophos, told Reuters that “with incidents like this, concern about a cyber event is natural”.

The disruption marks one of the most significant global internet incidents since last year’s Microsoft outage, which saw a defective CrowdStrike update cripple hospital, banking and airport systems worldwide.