The outage began around 3 a.m. Eastern Time in AWS’s US-East-1 region, according to the company’s service health dashboard. Amazon confirmed “increased error rates and latencies” across several of its cloud products, including computing and storage systems that power much of the web.
Reports of disruptions quickly spread across social media as users struggled to access a variety of services. Downdetector, a platform that tracks outages, showed spikes in problem reports for platforms like Signal, Robinhood, Coinbase, and Ring. Government and financial services were also affected, highlighting the extent of modern dependence on centralized cloud infrastructure.
By 6 a.m. ET, AWS engineers said they had begun restoring normal operations, though intermittent issues persisted through the morning. Amazon later attributed the disruption to “a subsystem responsible for DNS and network load balancer health checks,” but clarified that no security breach was involved.
This marks the company’s second major outage of 2025, following a smaller incident in March. Analysts note that as AWS continues to hold roughly one-third of the global cloud market, such interruptions have an outsized impact.
“An event like this shows how the internet’s resilience is concentrated in a few hands,” said analyst Jordan Lin of CloudMetrics Research in an interview with The Guardian. “When AWS stumbles, the entire digital ecosystem feels it.”
While service interruptions were resolved by midday, businesses reliant on Amazon’s cloud infrastructure could face lingering economic fallout. Data analytics firm NetBlocks estimated that the outage may have cost tens of millions of dollars per hour in lost productivity and transactions.
The incident also renews questions about diversification in the cloud industry. Some experts are calling for regulators to treat cloud providers as essential infrastructure, similar to utilities. Others emphasize the need for backup systems spread across multiple providers like Microsoft Azure and Google Cloud to avoid single points of failure.
Amazon has since said it is “reviewing the root cause to prevent recurrence” and will release a full incident report in the coming days.
