AWS Records Third Outage in Three Months in US-West-2, Disrupting Apple Pay, Reddit, and PlayStation for 64 Minutes
On the morning of July 24, AWS experienced a 64-minute outage in the Oregon region, the third failure in three months, affecting Apple Pay, DoorDash, Reddit, PlayStation Network and other services.
What Went Down at 3:55 AM on the West Coast of the U.S.
At 3:55 AM PDT on July 24, Amazon Web Services recorded network connectivity issues in the US-West-2 region, in Oregon. Within twenty minutes, reports from users about Apple Pay, DoorDash, Reddit, Hulu, and PlayStation Network surged globally. AWS listed seven affected services, including CloudFront, Global Accelerator, and Elastic Container Service, indicating a failure across multiple layers of the Oregon network, rather than an isolated product issue.
By 4:59 AM PDT, the company declared all network routes restored. Total duration: 64 minutes. AWS confirmed that no customer data was lost and did not disclose the specific technical cause of the event.
The Pattern: October 2025, May 2026, July 2026
In October 2025, the US-East-1 region experienced instability for hours due to a race condition within the DynamoDB DNS system: two automated management processes running in different availability zones conflicted. AWS published a detailed post-incident report a few weeks later. In May 2026, a hardware failure in cooling systems at a Virginia data center took down EC2 instances, affecting operations for 7 to 14 hours. A new report was issued.
The July event in Oregon was described only as a "connectivity issue." AWS had not released a post-incident analysis as of the publication of this article. The historical pattern suggests a window of two to four weeks for such a document to appear.
According to Axis Intelligence's analysis of AWS outage history from 2025 to 2026, the throughline among major events is automation behaving in ways not anticipated by engineers, concentrated in regions that carry a disproportionate amount of internet traffic. US-West-2 is one of the densest AWS regions for enterprise workloads and North American consumer platforms.
What the July Event Reveals for Enterprises
For CIOs operating US-West-2 as their primary region without failover configured, 64 minutes represent more than just operational downtime. Apple Pay, DoorDash, and PlayStation Network were exposed to the same event due to their centralized dependence on the region, each with demanding internal SLAs.
The impact is especially relevant in agent-based workloads. A language model chaining API calls in a loop requires robust retry mechanisms and configured circuit breakers; without them, a 64-minute interruption generates a chain of abandoned tasks with no automatic recovery. The real cost of downtime is not the hour of inactivity, but the process that was not resumed.
U.S. and India: The Same Lesson, Different Contexts
In the United States, the third outage in three months feeds the argument that cloud integrators have been presenting to CIOs for months: multi-cloud and multi-region are not paranoia architectural options, they are requirements for contracts with SLAs above 99.99% availability. Each post-incident analysis from AWS published in the last three episodes revealed the same pattern: automated subsystems that passed testing failed in production under real load conditions.
In India, TCS, Infosys, and Wipro manage cloud services for global clients running workloads in US-West-2. Each documented interruption enters into availability reports that support infrastructure outsourcing contracts. The third occurrence in three months reinforces clients’ positions renegotiating managed infrastructure contracts: the demand for multi-cloud architecture shifts from being a compliance clause to a condition for signing.
Without a published RCA, the question remains open: was the Oregon problem isolated or is it symptomatic of an automation scale that has exceeded AWS's safe validation capacity?