It's been a while since I've used Redshift, but isn't it still dependent on data coming in via a COPY from S3? Any sort of Redshift materialized view offering would depend on batches of data landing in an underlying table or tables. The closest service offering from AWS is probably using Kinesis analytics (or Flink on KA) using their flavor of streaming SQL to join Kinesis streams forming new ones.
With the introduction of Spectrum you can back a Redshift table with data in S3 directly. I'm not sure how that interacts with the materialized views though. Probably not supported yet but I would expect it to be eventually.
