diff --git a/src/coreclr/jit/morph.cpp b/src/coreclr/jit/morph.cpp
index b6e1f808ea9094..57153284d5bcda 100644
--- a/src/coreclr/jit/morph.cpp
+++ b/src/coreclr/jit/morph.cpp
@@ -258,17 +258,26 @@ GenTree* Compiler::fgMorphCast(GenTree* tree)
     }
 #endif //! TARGET_64BIT
 
+#ifdef TARGET_ARMARCH
+    // AArch, unlike x86/amd64, has instructions that can cast directly from
+    // all integers (except for longs on AArch32 of course) to floats.
+    // Because there is no IL instruction conv.r4.un, uint/ulong -> float
+    // casts are always imported as CAST(float <- CAST(double <- uint/ulong)).
+    // We can eliminate the redundant intermediate cast as an optimization.
+    else if ((dstType == TYP_FLOAT) && (srcType == TYP_DOUBLE) && oper->OperIs(GT_CAST)
 #ifdef TARGET_ARM
-    else if ((dstType == TYP_FLOAT) && (srcType == TYP_DOUBLE) && (oper->gtOper == GT_CAST) &&
-             !varTypeIsLong(oper->AsCast()->CastOp()))
+             && !varTypeIsLong(oper->AsCast()->CastOp())
+#endif
+                 )
     {
-        // optimization: conv.r4(conv.r8(?)) -> conv.r4(d)
-        // except when the ultimate source is a long because there is no long-to-float helper, so it must be 2 step.
-        // This happens semi-frequently because there is no IL 'conv.r4.un'
         oper->gtType       = TYP_FLOAT;
         oper->CastToType() = TYP_FLOAT;
+
         return fgMorphTree(oper);
     }
+#endif // TARGET_ARMARCH
+
+#ifdef TARGET_ARM
     // converts long/ulong --> float/double casts into helper calls.
     else if (varTypeIsFloating(dstType) && varTypeIsLong(srcType))
     {