Redshift (Serverless) SVV_TABLE_INFO shows huge storage sizes for tiny tables

My Redshift serverless shows massive storage size usage for tiny tables that so far have had only a couple DDL statements, only inserts, and are overall tiny tables. When I run select "schema", "table", "tbl_rows", "size" FROM SVV_TABLE_INFO where "schema" = 'x' and "table = 'y' I get 36 rows with 2580 "size", which according to Redshift docs are 1mb blocks, so 2.6 GB of storage used with 36 rows. The SVV_TABLE_INFO columns empty, unsorted, vacuum_sort_benefit are all 0. The fun part is when I run select * from x.y I can copy the entire resultset to my clipboard and it comes to ~23kb total. The AWS Redshift Serverless Web GUI similarly reports a whopping 1.1 TB of storage used (the same as when running sum("size") against the SVV table btw) in the cluster. There is at most 100gb used in total. Can someone help me figure out how/where those huge storage numbers come from? ## EDIT - Full SVV Dump for one table

{
	"database": "xxx",
	"schema": "x",
	"table_id": 2234810,
	"table": "y",
	"encoded": "Y",
	"diststyle": "KEY(id)",
	"sortkey1": "received_at",
	"max_varchar": 65535,
	"sortkey1_enc": "none",
	"sortkey_num": 1,
	"size": 2580,
	"pct_used": 0.0040,
	"empty": 0,
	"unsorted": 0.00,
	"stats_off": 0.00,
	"tbl_rows": 36,
	"skew_sortkey1": 1.00,
	"skew_rows": 100.00,
	"estimated_visible_rows": 36,
	"risk_event": null,
	"vacuum_sort_benefit": 0.00,
	"create_time": "2024-08-05T07:48:07.454Z"
}