diff --git a/tests/integration/test_lists/test-db/l0_b200.yml b/tests/integration/test_lists/test-db/l0_b200.yml
index ba99e5284330..47cc68539fde 100644
--- a/tests/integration/test_lists/test-db/l0_b200.yml
+++ b/tests/integration/test_lists/test-db/l0_b200.yml
@@ -99,6 +99,9 @@ l0_b200:
   # investigation before enabling on this machine type (TRTLLM-15204).
   - unittest/_torch/modules/kimi_kda/test_kda_prefill_state_parity.py
   - unittest/_torch/modules/kimi_k3_attn_res/test_attn_res_op.py
+  # GPU KDA disagg transfer + peer-validation (cpu_only cases skipped by the
+  # stage's "not cpu_only" markexpr and run on the CPU-Generic stage instead).
+  - unittest/disaggregated/test_kda_mamba_transfer.py
   - unittest/_torch/modules/moe/test_kimi_k3_situ_moe.py
   - unittest/_torch/custom_ops/test_deepseek_v4_q_norm.py TIMEOUT (15)
   # ------------- modules (non-MoE) ---------------
diff --git a/tests/unittest/disaggregated/test_kda_mamba_transfer.py b/tests/unittest/disaggregated/test_kda_mamba_transfer.py
index fcef4587c05b..51e5a7ded365 100644
--- a/tests/unittest/disaggregated/test_kda_mamba_transfer.py
+++ b/tests/unittest/disaggregated/test_kda_mamba_transfer.py
@@ -342,6 +342,7 @@ def _synthetic_rank_info(tp: int, adp: bool):
     )
 
 
+@pytest.mark.cpu_only
 @pytest.mark.parametrize(
     "ctx,gen,ok",
     [
@@ -383,6 +384,7 @@ def build(cfg):
             MambaPolicy.validate_peer_compatible(ctx_ri, gen_ri, ctx_pt, gen_pt)
 
 
+@pytest.mark.cpu_only
 def test_kda_peer_validation_allows_pipeline_parallel_layer_split():
     """Peer validation must not require identical layer sets.