diff --git a/backend/app.py b/backend/app.py
index 1cbac8a1..6d79a6a9 100644
--- a/backend/app.py
+++ b/backend/app.py
@@ -163,18 +163,17 @@ def _get_gpu_status() -> str:
     elif backend_type == "mlx":
         return "Metal (Apple Silicon via MLX)"
 
-    # Intel XPU (Arc / Data Center) via IPEX
-    try:
-        import intel_extension_for_pytorch  # noqa: F401
-
-        if hasattr(torch, "xpu") and torch.xpu.is_available():
-            try:
-                xpu_name = torch.xpu.get_device_name(0)
-            except Exception:
-                xpu_name = "Intel GPU"
-            return f"XPU ({xpu_name})"
-    except ImportError:
-        pass
+    # Intel XPU (Arc / Data Center) — native PyTorch 2.4+ support; IPEX optional
+    if hasattr(torch, "xpu") and torch.xpu.is_available():
+        try:
+            import intel_extension_for_pytorch  # noqa: F401 -- enhances XPU perf if available
+        except Exception:
+            pass
+        try:
+            xpu_name = torch.xpu.get_device_name(0)
+        except Exception:
+            xpu_name = "Intel GPU"
+        return f"XPU ({xpu_name})"
 
     return "None (CPU only)"
 
diff --git a/backend/backends/base.py b/backend/backends/base.py
index c566af10..d774645e 100644
--- a/backend/backends/base.py
+++ b/backend/backends/base.py
@@ -102,13 +102,12 @@ def get_torch_device(
         return "cuda"
 
     if allow_xpu:
-        try:
-            import intel_extension_for_pytorch  # noqa: F401
-
-            if hasattr(torch, "xpu") and torch.xpu.is_available():
-                return "xpu"
-        except ImportError:
-            pass
+        if hasattr(torch, "xpu") and torch.xpu.is_available():
+            try:
+                import intel_extension_for_pytorch  # noqa: F401 -- enhances XPU perf if available
+            except Exception:
+                pass
+            return "xpu"
 
     if allow_directml:
         try:
diff --git a/backend/routes/health.py b/backend/routes/health.py
index 79c513f5..0bb454f9 100644
--- a/backend/routes/health.py
+++ b/backend/routes/health.py
@@ -67,17 +67,17 @@ async def health():
 
     has_xpu = False
     xpu_name = None
-    try:
-        import intel_extension_for_pytorch as ipex  # noqa: F401 -- side-effect import enables XPU
-
-        if hasattr(torch, "xpu") and torch.xpu.is_available():
-            has_xpu = True
-            try:
-                xpu_name = torch.xpu.get_device_name(0)
-            except Exception:
-                xpu_name = "Intel GPU"
-    except ImportError:
-        pass
+    # Native XPU support in PyTorch 2.4+; IPEX optional for enhanced performance
+    if hasattr(torch, "xpu") and torch.xpu.is_available():
+        has_xpu = True
+        try:
+            import intel_extension_for_pytorch  # noqa: F401 -- enhances XPU perf if available
+        except Exception:
+            pass
+        try:
+            xpu_name = torch.xpu.get_device_name(0)
+        except Exception:
+            xpu_name = "Intel GPU"
 
     has_directml = False
     directml_name = None