diff --git a/python/sglang/srt/models/deepseek_v2.py b/python/sglang/srt/models/deepseek_v2.py
index 021173194..5411052e9 100644
--- a/python/sglang/srt/models/deepseek_v2.py
+++ b/python/sglang/srt/models/deepseek_v2.py
@@ -3593,8 +3593,10 @@ class DeepseekV2ForCausalLM(nn.Module):
             self.config.q_lora_rank is not None
         )
         cached_a_proj = {} if fuse_qkv_a_proj else None
-        fuse_wk_and_weights_proj = is_nsa_indexer_wk_and_weights_proj_fused(
-            self.config, self.quant_config
+        # Fuse wk and weights_proj when NSA Indexer is enabled and quant_config is FP4. For nextn, fp4 is disabled so we cannot fuse.
+        fuse_wk_and_weights_proj = (
+            is_nsa_indexer_wk_and_weights_proj_fused(self.config, self.quant_config)
+            and not is_nextn
         )
         cached_wk_and_weights_proj = {} if fuse_wk_and_weights_proj else None