Snowflake is most definitely running multiple instances for anything but the smallest clusters. It's plausible that one warehouse might scale better as you increase the number of nodes but I would really be surprised if there was a big difference---the techniques for query execution are pretty well-known, so Redshift and Snowflake are probably doing the same things. I did a benchmark I while ago [1] and I'm definitely planning on adding multiple scales to evaluate how they compare as you increase data size.

[1] https://blog.fivetran.com/warehouse-benchmark-dce9f4c529c1