diff --git a/app/models/import.rb b/app/models/import.rb
index d2565a3e8..17e10dbaa 100644
--- a/app/models/import.rb
+++ b/app/models/import.rb
@@ -498,6 +498,13 @@ class Import < ApplicationRecord
       @parsed_csv = self.class.parse_csv_str(csv_content, col_sep: col_sep)
     end
 
+    # Normalizes a raw CSV numeric string into a plain, parseable decimal string
+    # based on the import's configured +number_format+ (thousands delimiter and
+    # decimal separator). Returns "" when the value is blank, the format is
+    # unknown, or the result is not a valid number.
+    #
+    # @param value [String, nil] the raw cell value from the CSV
+    # @return [String] a normalized number like "1234.56", or "" if invalid
     def sanitize_number(value)
       return "" if value.nil?
 
@@ -509,7 +516,20 @@ class Import < ApplicationRecord
 
       # Handle French/Scandinavian format specially
       if format[:delimiter] == " "
-        sanitized = sanitized.gsub(/\s+/, "") # Remove all spaces first
+        # The thousands "space" can be an ASCII space, a non-breaking space
+        # (U+00A0) or a narrow no-break space (U+202F) depending on the locale
+        # or exporter. Ruby's \s does not match those Unicode spaces, so strip
+        # every kind of whitespace via the Unicode property.
+        sanitized = sanitized.gsub(/\p{Space}/, "")
+
+        # Strip currency symbols/codes only at the leading/trailing edges (e.g.
+        # "€1 234,56" or "1 234,56 kr"). Interior characters are deliberately
+        # left in place so a misconfigured US-style value like "1,234.56" keeps
+        # its period and is rejected by the numeric guard below, rather than
+        # being silently reinterpreted as 1.23456. Digits, the separator, and a
+        # minus sign are preserved so signed values and the guard still work.
+        edge_junk = /\A[^\d#{Regexp.escape(format[:separator])}\-]+|[^\d#{Regexp.escape(format[:separator])}\-]+\z/
+        sanitized = sanitized.gsub(edge_junk, "")
       else
         sanitized = sanitized.gsub(/[^\d#{Regexp.escape(format[:delimiter])}#{Regexp.escape(format[:separator])}\-]/, "")
 
diff --git a/test/interfaces/import_interface_test.rb b/test/interfaces/import_interface_test.rb
index 2b4759277..e30819594 100644
--- a/test/interfaces/import_interface_test.rb
+++ b/test/interfaces/import_interface_test.rb
@@ -112,6 +112,94 @@ module ImportInterfaceTest
     assert_equal "1234.56", row.amount
   end
 
+  test "parses French/Scandinavian format with non-breaking-space thousands separator" do
+    import = imports(:transaction)
+    import.update!(
+      number_format: "1 234,56",
+      amount_col_label: "amount",
+      date_col_label: "date",
+      name_col_label: "name",
+      date_format: "%m/%d/%Y"
+    )
+
+    # Real-world European CSV exports use a non-breaking space (U+00A0) as the
+    # thousands separator rather than a plain ASCII space.
+    csv_data = "date,amount,name\n01/01/2024,\"1 234,56\",Test"
+    import.update!(raw_file_str: csv_data)
+    import.generate_rows_from_csv
+    import.reload
+
+    row = import.rows.first
+    assert_equal "1234.56", row.amount
+  end
+
+  test "parses French/Scandinavian format with narrow no-break-space thousands separator" do
+    import = imports(:transaction)
+    import.update!(
+      number_format: "1 234,56",
+      amount_col_label: "amount",
+      date_col_label: "date",
+      name_col_label: "name",
+      date_format: "%m/%d/%Y"
+    )
+
+    # Some locales/exporters use a narrow no-break space (U+202F) as the
+    # thousands separator, which Ruby's \s also does not match.
+    csv_data = "date,amount,name\n01/01/2024,\"1 234,56\",Test"
+    import.update!(raw_file_str: csv_data)
+    import.generate_rows_from_csv
+    import.reload
+
+    row = import.rows.first
+    assert_equal "1234.56", row.amount
+  end
+
+  test "rejects US-punctuation values under the French/Scandinavian format" do
+    import = imports(:transaction)
+    import.update!(
+      number_format: "1 234,56",
+      amount_col_label: "amount",
+      date_col_label: "date",
+      name_col_label: "name",
+      date_format: "%m/%d/%Y"
+    )
+
+    # A misconfigured/mixed row using US separators ("1,234.56") must not be
+    # silently reinterpreted as 1.23456 under a space-delimited format; the
+    # unexpected period keeps it invalid so it surfaces as blank instead.
+    csv_data = "date,amount,name\n01/01/2024,\"1,234.56\",Test"
+    import.update!(raw_file_str: csv_data)
+    import.generate_rows_from_csv
+    import.reload
+
+    row = import.rows.first
+    assert_equal "", row.amount
+  end
+
+  test "strips leading/trailing currency junk under the French/Scandinavian format" do
+    import = imports(:transaction)
+    import.update!(
+      number_format: "1 234,56",
+      amount_col_label: "amount",
+      date_col_label: "date",
+      name_col_label: "name",
+      date_format: "%m/%d/%Y"
+    )
+
+    # Currency symbols/codes at the edges (issue #2537's "1 234,56 kr" row) are
+    # stripped; the amount still parses. Interior junk is not stripped (covered
+    # by the mixed-punctuation rejection test above).
+    csv_data = "date,amount,name\n" \
+               "01/01/2024,\"1 234,56 kr\",Suffix\n" \
+               "01/02/2024,\"€1 234,56\",Prefix"
+    import.update!(raw_file_str: csv_data)
+    import.generate_rows_from_csv
+    import.reload
+
+    assert_equal "1234.56", import.rows.first.amount
+    assert_equal "1234.56", import.rows.second.amount
+  end
+
   test "parses zero-decimal currency format correctly" do
     import = imports(:transaction)
     import.update!(