diff --git a/graphistry/Engine.py b/graphistry/Engine.py
index 3209f5349c..19bcfdcef4 100644
--- a/graphistry/Engine.py
+++ b/graphistry/Engine.py
@@ -15,6 +15,8 @@
     # triggers ``graphistry.compute`` package init at runtime (would be circular).
     from graphistry.compute.typing import DataFrameT, PolarsFrame, PolarsSeriesT, SeriesT
     from graphistry.Plottable import Plottable
+    import cudf
+    import cupy
     import polars as pl
     # TypeGuard, NOT TypeIs (PEP 742). TypeIs additionally narrows the negative branch, and to
     # do that soundly it REQUIRES the narrowed type to be consistent with the declared input
@@ -249,6 +251,32 @@ def is_polars_df(df: object) -> "TypeGuard[PolarsFrame]":
     return df is not None and "polars" in type(df).__module__
 
 
+def is_cudf_series(series: object) -> "TypeGuard[cudf.Series]":
+    """True if ``series`` is a cuDF ``Series``.
+
+    The module check avoids importing the optional cuDF dependency for other engines;
+    the exact ``isinstance`` check prevents look-alike classes from taking the cuDF path.
+    """
+    if series is None or "cudf" not in type(series).__module__:
+        return False
+    import cudf
+
+    return isinstance(series, cudf.Series)
+
+
+def is_cupy_array(array: object) -> "TypeGuard[cupy.ndarray]":
+    """True if ``array`` is a CuPy array.
+
+    The module check avoids importing the optional GPU dependency for other
+    engines; the exact ``isinstance`` check rejects look-alikes.
+    """
+    if array is None or "cupy" not in type(array).__module__:
+        return False
+    import cupy
+
+    return isinstance(array, cupy.ndarray)
+
+
 def active_frames_are_polars(g: Any) -> bool:
     """True if ``g``'s active table (nodes, else edges) is a polars frame."""
     if g._nodes is not None:
diff --git a/graphistry/compute/gfql/index/api.py b/graphistry/compute/gfql/index/api.py
index 8fa76166bd..aa39d51e0f 100644
--- a/graphistry/compute/gfql/index/api.py
+++ b/graphistry/compute/gfql/index/api.py
@@ -8,12 +8,16 @@
 from __future__ import annotations
 
 import copy
-from typing import Dict, List, Literal, Mapping, Optional, Sequence, Set, Tuple, cast
+from typing import TYPE_CHECKING, Dict, List, Literal, Mapping, Optional, Sequence, Set, Tuple, Union, cast
 
+import numpy as np
 import pandas as pd
 
-from graphistry.Engine import EngineAbstract, Engine, EngineAbstractType, POLARS_ENGINES, is_polars_df, resolve_engine
-from graphistry.compute.typing import DataFrameT
+from graphistry.Engine import (
+    EngineAbstract, Engine, EngineAbstractType, POLARS_ENGINES,
+    is_cudf_series, is_cupy_array, is_polars_df, is_polars_series, resolve_engine,
+)
+from graphistry.compute.typing import DataFrameT, SeriesT
 from graphistry.Plottable import Plottable
 from .registry import (
     AdjacencyIndex, ColStatsFact, PartitionValue, ColStatsRole, GfqlIndexRegistry, EMPTY_REGISTRY, NodeIdIndex,
@@ -34,11 +38,31 @@
     TraceEngine, SidecarFactKind,
 )
 
+if TYPE_CHECKING:
+    import cudf
+    from graphistry.compute.typing import ArrayLike, PolarsSeriesT
+
+    CrossEngineByteSourceT = Union[ArrayLike, SeriesT, PolarsSeriesT, "cudf.Series"]
+
 # Private Plottable attachment keys. Keep access behind helpers.
 POLICY_ATTR = "_gfql_index_policy"
 REGISTRY_ATTR = "_gfql_index_registry"
 
 
+def _byte_source_nbytes(value: CrossEngineByteSourceT) -> int:
+    if is_polars_series(value):
+        return int(value.estimated_size())
+    if is_cudf_series(value):
+        return int(value.memory_usage())
+    if isinstance(value, np.ndarray):
+        return int(value.nbytes)
+    if is_cupy_array(value):
+        return int(value.nbytes)
+    if isinstance(value, pd.Series):
+        return int(value.memory_usage())
+    return 0
+
+
 # --- lightweight, thread-local index decision trace (for gfql_explain) -------
 import threading as _threading
 _TRACE = _threading.local()
@@ -591,7 +615,7 @@ def show_indexes(
         )
         usable, reason = _index_usability(kind, typed.engine, valid, query_engine)
         # other_values is shared with the ordinary adjacency index, not added memory.
-        nbytes = sum(int(getattr(typed, name).nbytes) for name in (
+        nbytes = sum(_byte_source_nbytes(getattr(typed, name)) for name in (
             "keys_sorted", "group_offsets", "row_positions",
         ))
         rows.append({
@@ -611,7 +635,7 @@ def show_indexes(
             is segment
         )
         usable, reason = _index_usability(NODE_TYPE_SEGMENT, segment.engine, valid, query_engine)
-        nbytes = sum(int(getattr(segment, name).nbytes) for name in (
+        nbytes = sum(_byte_source_nbytes(getattr(segment, name)) for name in (
             "keys_sorted", "group_offsets", "row_positions",
         ))
         rows.append({
diff --git a/graphistry/tests/compute/gfql/index/test_node_type_segments.py b/graphistry/tests/compute/gfql/index/test_node_type_segments.py
index 511bf8d044..e1fecf746f 100644
--- a/graphistry/tests/compute/gfql/index/test_node_type_segments.py
+++ b/graphistry/tests/compute/gfql/index/test_node_type_segments.py
@@ -1,8 +1,11 @@
 """Node type-segment index must preserve canonical GFQL results (edge typed-index dual)."""
 from __future__ import annotations
 
+import numpy as np
 import pandas as pd
 import pytest
+import sys
+import types
 
 import graphistry
 from graphistry.Engine import Engine
@@ -10,10 +13,28 @@
 from graphistry.compute.gfql.index import (
     get_registry, index_trace, maybe_index_hop, with_index_policy,
 )
+from graphistry.Engine import is_cupy_array
+from graphistry.compute.gfql.index.api import _byte_source_nbytes
 from graphistry.compute.gfql.index.cypher_ddl import parse_index_ddl
 from graphistry.compute.gfql.index.wire import apply_index_op
 
 
+class _CpuCudfSeries:
+    def __init__(self, values: list[int]) -> None:
+        self.values = values
+
+    def memory_usage(self) -> int:
+        return 16
+
+
+class _CpuCuPyArray:
+    nbytes = 8
+
+
+_CpuCudfSeries.__module__ = "cudf"
+_CpuCuPyArray.__module__ = "cupy"
+
+
 def _graph(engine: str):
     # Node 1 has TWO labels (membership); node 5 has a null type; node 4 is
     # label-only-False; a null type never matches any scalar segment.
@@ -37,6 +58,79 @@ def _graph(engine: str):
     return graphistry.bind(node="id", source="src", destination="dst").nodes(nodes).edges(edges)
 
 
+def test_node_segment_show_indexes_reports_cudf_bytes() -> None:
+    pytest.importorskip("cudf")
+    cupy = pytest.importorskip("cupy")
+    try:
+        device_count = cupy.cuda.runtime.getDeviceCount()
+    except Exception:
+        pytest.skip("CUDA device detection failed")
+    if device_count == 0:
+        pytest.skip("cuDF requires a CUDA device")
+    indexed = _indexed(_graph("cudf"), "cudf")
+    rows = indexed.show_indexes()
+    segment_rows = rows[rows["kind"] == "node_type_segment"]
+    assert not segment_rows.empty
+    assert segment_rows["nbytes"].sum() > 0
+
+
+def test_byte_source_nbytes_uses_pandas_measurement() -> None:
+    series = pd.Series([1], dtype="int64")
+    assert _byte_source_nbytes(series) == series.memory_usage()
+
+
+def test_byte_source_nbytes_uses_polars_measurement() -> None:
+    pl = pytest.importorskip("polars")
+    series = pl.Series([1], dtype=pl.Int64)
+    assert _byte_source_nbytes(series) == series.estimated_size()
+
+
+def test_byte_source_nbytes_uses_cudf_measurement() -> None:
+    cudf = pytest.importorskip("cudf")
+    try:
+        device_count = cudf.cuda.runtime.getDeviceCount()
+    except Exception:
+        pytest.skip("CUDA device detection failed")
+    if device_count == 0:
+        pytest.skip("cuDF requires a CUDA device")
+    series = cudf.Series([1], dtype="int64")
+    assert _byte_source_nbytes(series) == series.memory_usage()
+
+
+def test_byte_source_nbytes_uses_cudf_guard_without_cuda(monkeypatch: pytest.MonkeyPatch) -> None:
+    monkeypatch.setitem(sys.modules, "cudf", types.SimpleNamespace(Series=_CpuCudfSeries))
+    series = _CpuCudfSeries([1])
+    assert _byte_source_nbytes(series) == 16
+
+
+def test_byte_source_nbytes_defaults_to_zero_for_unknown_byte_source() -> None:
+    assert _byte_source_nbytes(object()) == 0
+
+
+def test_byte_source_nbytes_uses_numpy_measurement() -> None:
+    assert _byte_source_nbytes(np.asarray([1], dtype="int64")) == 8
+
+
+def test_byte_source_nbytes_uses_cupy_measurement() -> None:
+    cupy = pytest.importorskip("cupy")
+    try:
+        device_count = cupy.cuda.runtime.getDeviceCount()
+    except Exception:
+        pytest.skip("CUDA device detection failed")
+    if device_count == 0:
+        pytest.skip("CuPy requires a CUDA device")
+    array = cupy.asarray([1], dtype="int64")
+    assert _byte_source_nbytes(array) == 8
+    assert is_cupy_array(array)
+
+
+def test_byte_source_nbytes_uses_cupy_guard_without_cuda(monkeypatch: pytest.MonkeyPatch) -> None:
+    monkeypatch.setitem(sys.modules, "cupy", types.SimpleNamespace(ndarray=_CpuCuPyArray))
+    array = _CpuCuPyArray()
+    assert _byte_source_nbytes(array) == 8
+    assert is_cupy_array(array)
+
+
 def _pandas(frame):
     return frame.to_pandas() if hasattr(frame, "to_pandas") else frame