It was Amazon S3 US Standard (us-east-1) region that went down. This caused quite a lot of pain throughout the internet because people were using S3 as storage

It was Amazon S3 US Standard (us-east-1) region that went down. This caused quite a lot of pain throughout the internet because people were using S3 as storage for their application and/or their static assets. Other AWS services were affected, but this was mostly limited to creating, updating, and deleting AWS resources (EC2 instances, RDS instances, CodeDeploy deployments, etc.). There were no issues I'm aware of for already-provisioned resources. There was also some issues with the AWS console. So Amazon.com was able to weather the storm because it did not have a hard dependency on S3 US Standard. They may not use S3 behind the scenes, they may be using another region, or they have cross-region replication to serve files from redundant regions. EDIT: Another option, they may have put S3 behind a CDN and were able to lean on the CDN's cache while S3 was down.