diff --git a/migrator/scripts/dump_affiliations_db.py b/migrator/scripts/dump_affiliations_db.py
new file mode 100644
index 000000000..d7205c54f
--- /dev/null
+++ b/migrator/scripts/dump_affiliations_db.py
@@ -0,0 +1,84 @@
+"""Parse affiliations from ROR dumps into CSV format, importable via COPY.
+
+To use call ``load_file(DATA_PATH, "affiliations.csv")``.
+"""
+
+import csv
+import orjson as json
+import uuid
+
+from idutils import normalize_ror
+from invenio_rdm_migrator.utils import ts
+
+DATA_PATH = "v1.32-2023-09-14-ror-data.zip"  # https://zenodo.org/record/8346986
+
+
+VOCABULARIES_AFFILIATION_SCHEMES = {
+    "grid",
+    "gnd",
+    "isni",
+    "ror",
+}
+
+
+def transform_affiliation(data):
+    """Applies the transformation to the stream entry."""
+    affiliation = {
+        "$schema": "local://affiliations/affiliation-v1.0.0.json",
+    }
+
+    affiliation["id"] = normalize_ror(data.get("id"))
+    if not affiliation["id"]:
+        return
+
+    affiliation["name"] = data.get("name")
+    if not affiliation["name"]:
+        return
+
+    acronyms = data.get("acronyms") or []
+    if acronyms:
+        affiliation["acronym"] = acronyms[0]
+
+    affiliation["title"] = {"en": affiliation["name"]}
+    for label in data.get("labels", []):
+        affiliation["title"][label["iso639"]] = label["label"]
+
+    # The ROR is always listed in identifiers, expected by serialization
+    affiliation["identifiers"] = [{"identifier": affiliation["id"], "scheme": "ror"}]
+    for scheme, identifier in data.get("external_ids", {}).items():
+        scheme = scheme.lower()
+        if scheme in VOCABULARIES_AFFILIATION_SCHEMES:
+            value = identifier.get("preferred") or identifier.get("all")[0]
+            affiliation["identifiers"].append({"identifier": value, "scheme": scheme})
+
+    return affiliation
+
+
+def load_file(datafile, outpath):
+    """Load the data file and dump as CSV."""
+    with open(outpath, "w") as fout, open(datafile, "rb") as fp:
+        print(f"[{ts()}] loading {datafile}")
+        writer = csv.writer(fout)
+        entries = json.loads(fp.read())
+        for idx, data in enumerate(entries):
+            if idx % 1000 == 0:
+                print(f"[{ts()}] {idx}")
+            try:
+                affiliation = transform_affiliation(data)
+                if not affiliation:
+                    print(f"[{ts()}] Failed to transform #{idx}:\n{data}\n")
+                    continue
+                affiliation_id = affiliation.pop("id")
+                creation_ts = ts()
+                writer.writerow(
+                    (
+                        str(uuid.uuid4()),  # id
+                        affiliation_id,  # pid
+                        json.dumps(affiliation),  # json
+                        creation_ts,  # created
+                        creation_ts,  # updated (same as created)
+                        1,  # version_id
+                    )
+                )
+            except Exception as ex:
+                print(f"[{ts()}] Exception for line {idx}:\n{data}\n\n{ex}\n")