Around 2 p.m. ET on Thursday, shoppers began reporting problems with the website and mobile app of Amazon

For several hours, thousands of users were unable to complete purchases, view product prices, or access their accounts on one of the world’s largest online retail platforms. 

Reports of the outage quickly surged on Downdetector, which tracks service disruptions based on user submissions. The number of complaints jumped sharply shortly after 2 p.m. ET and eventually peaked at around 22,000 reports. Many users said they were unable to check out or place orders, while others reported that product pages were failing to load properly or were missing pricing information entirely. Some shoppers also experienced app crashes or found themselves locked out of their order histories and account pages.

Amazon Says AWS China Outage Was “User Error” Not AI Failure
An internal error, including Kiro AI opting to “delete and recreate” the environment, led to the outage, which Amazon called “user error, not AI error,” per the Financial Times.

Amazon later confirmed that the incident stemmed from an internal technical issue rather than an external attack. In a statement, a company spokesperson said the disruption was caused by a software code deployment, suggesting that a newly introduced update had triggered unexpected problems within parts of the company’s retail infrastructure.

The company said engineers worked to address the issue and that both the website and the mobile app were operating normally again once the fix was implemented later in the evening.

As the outage unfolded, some users also reported disruptions involving other Amazon services. Downdetector logged scattered issues tied to the company’s streaming platform, Prime Video. However, Amazon said its cloud services were functioning normally overall, indicating that the outage appeared to be limited to the company’s shopping systems rather than its broader cloud infrastructure.

The incident comes during a week when Amazon’s infrastructure has also faced unrelated challenges overseas. Earlier reports said drone strikes linked to the ongoing Middle East conflict damaged some Amazon data centers in Bahrain and the United Arab Emirates.

Iranian state media attributed the attack to the Islamic Revolutionary Guard Corps, claiming the facilities were targeted because of Amazon’s alleged support for U.S. military and intelligence operations. Despite the reported damage, Amazon said its cloud services continued operating normally.

By Thursday night, the shopping outage had largely subsided, bringing Amazon’s retail platform back online after several hours of disruption that briefly left thousands of customers unable to browse, shop, or complete purchases.

Why Is Claude Not Working? Everything We Know About the Anthropic Outage
Users report login issues and API failures across Claude services as Anthropic works on it