AWS's storage offering S3 is pretty limited in terms of its consistency model: https://medium.com/@jim_dowling/reflections-on-s3s-architect...
In contrast, Google's filesystem (yes, not an object store, but a filesystem) is hardly known - Collosus. It provides a HA consistent metadata layer to a distributed filesystem. The closest open-source architecture is HopsFS (distributed metadata for HDFS) - disclosure I work on Hops.
[1] https://aws.amazon.com/about-aws/whats-new/2017/11/sign-up-f...
If I'm right with this, then this is basically stating: "yeah, aws also has a high scale database offering" but it's not really comparable anyhow.
Serverless Aurora can autoscale up and down based on usage, and Aurora has powerful replication functionality, but this is a far cry from Spanner's multi-master horizontal scaling.
DynamoDB can autoscale (crudely), and they just added multi-region multi-master. It's non-relational though and requires more capacity planning than Spanner.
No clue how far it is from General Availability or if it allows even remotely close to the horizontal scaling of Spanner. But it’s at least being worked on.
[1] https://aws.amazon.com/about-aws/whats-new/2017/11/sign-up-f...
Neither provide the guarantees of TrueTime.
Excerpt: "...The designers of Spanner call this ‘TrueTime’, and it provides a tight bound on clock offset between any two nodes in the system. TrueTime enables high levels of external consistency. As an open source database based on Spanner, our challenge was in providing similar guarantees of external consistency without atomic clocks."