An AWS outage that began on May 7 is still causing headaches across the internet, with Amazon saying recovery is progressing more slowly than expected. The disruption has centered on the US-EAST-1 Region, and the ripple effects have already hit trading apps, analytics tools, and a long list of services that quietly depend on AWS to stay online.

In its latest update at 12:29 p.m. ET, Amazon said recovery efforts were underway but warned that ”Full recovery is still expected to take several hours.” That is the kind of sentence every engineer dreads writing and every user hates reading. Cloud outages are rarely glamorous, but they are a neat reminder that a handful of infrastructure giants can make the modern web wobble in very ordinary ways.

AWS outage reports spiked, slowed, then climbed again

DownDetector, which tracks user-submitted error reports, showed complaints jumping around 8 p.m. ET on Thursday. Reports eased into Friday before rising again around 4 p.m. ET, suggesting the problem never fully settled down. That pattern fits a broader truth about cloud failures: even when the most visible blast radius shrinks, the cleanup can drag on for hours.

The outage also appears to have reached services far outside Amazon’s own ecosystem. CNBC reported temporary disruption to FanDuel and Coinbase, while Chartbeat said its issues were tied to AWS outages. Coinbase posted that service disruptions were linked to increased temperatures in the affected AWS service and said its funds were safe.

US-EAST-1 has a history of making life awkward

This is not AWS’s first rough ride in US-EAST-1. The last major outage affecting that region happened in October 2025, when a wide range of apps appeared to be caught in the blast radius, including Amazon, Roblox, HBO Max, Venmo, Lyft, Signal, and AT&T. That’s the part cloud providers never love to advertise: redundancy is sold as insurance, but a regional failure can still leave a lot of very public services twitching at once.

Internet outages have become a recurring feature of the last 12 months, which is a polite way of saying the web is still held together by a surprising amount of concentrated plumbing. AWS powers cloud hosting and other internet services, so when one region stumbles, the effects can be immediate and annoyingly broad.

What to watch as Amazon finishes recovery

• Amazon says full recovery should still take several hours.
• The outage has centered on US-EAST-1.
• DownDetector showed complaints easing, then rising again later on Friday.

AWS has not offered much detail publicly beyond its recovery updates, and Mashable said it had reached out for more information. The real question now is whether this becomes another short-lived cloud scare or another reminder that too much of the internet still leans on a few fragile choke points.