diff --git a/External/FEXCore/Source/Interface/Core/Interpreter/VectorOps.cpp b/External/FEXCore/Source/Interface/Core/Interpreter/VectorOps.cpp
index 7a4ba3454..b8ee08d41 100644
--- a/External/FEXCore/Source/Interface/Core/Interpreter/VectorOps.cpp
+++ b/External/FEXCore/Source/Interface/Core/Interpreter/VectorOps.cpp
@@ -23,7 +23,7 @@ DEF_OP(VectorImm) {
   auto Op = IROp->C<IR::IROp_VectorImm>();
   uint8_t OpSize = IROp->Size;
 
-  uint8_t Tmp[16];
+  uint8_t Tmp[32];
 
   uint8_t Elements = OpSize / Op->Header.ElementSize;
   uint8_t Imm = Op->Immediate;
diff --git a/External/FEXCore/Source/Interface/Core/JIT/Arm64/VectorOps.cpp b/External/FEXCore/Source/Interface/Core/JIT/Arm64/VectorOps.cpp
index 27d96a5cb..7eba50d2b 100644
--- a/External/FEXCore/Source/Interface/Core/JIT/Arm64/VectorOps.cpp
+++ b/External/FEXCore/Source/Interface/Core/JIT/Arm64/VectorOps.cpp
@@ -40,15 +40,37 @@ DEF_OP(VectorImm) {
   auto Op = IROp->C<IR::IROp_VectorImm>();
 
   const uint8_t OpSize = IROp->Size;
-  const uint8_t Elements = OpSize / Op->Header.ElementSize;
+  const uint8_t ElementSize = Op->Header.ElementSize;
+  const uint8_t Elements = OpSize / ElementSize;
 
-  if (Op->Header.ElementSize == 8) {
-    // movi with 64bit element size doesn't do what we want here
-    LoadConstant(TMP1.X(), Op->Immediate);
-    dup(GetDst(Node).V2D(), TMP1.X());
-  }
-  else {
-    movi(GetDst(Node).VCast(OpSize * 8, Elements), Op->Immediate);
+  if (CanUseSVE) {
+    const auto Dst = [&] {
+      const auto Tmp = GetDst(Node).Z();
+      switch (ElementSize) {
+      case 1:
+        return Tmp.VnB();
+      case 2:
+        return Tmp.VnH();
+      case 4:
+        return Tmp.VnS();
+      case 8:
+        return Tmp.VnD();
+      default:
+        LOGMAN_MSG_A_FMT("Unhandled element size: {}", ElementSize);
+        return Tmp;
+      }
+    }();
+
+    dup(Dst, Op->Immediate);
+  } else {
+    if (ElementSize == 8) {
+      // movi with 64bit element size doesn't do what we want here
+      LoadConstant(TMP1.X(), Op->Immediate);
+      dup(GetDst(Node).V2D(), TMP1.X());
+    }
+    else {
+      movi(GetDst(Node).VCast(OpSize * 8, Elements), Op->Immediate);
+    }
   }
 }
 
diff --git a/External/FEXCore/Source/Interface/Core/JIT/x86_64/VectorOps.cpp b/External/FEXCore/Source/Interface/Core/JIT/x86_64/VectorOps.cpp
index e54f472f2..968c673d5 100644
--- a/External/FEXCore/Source/Interface/Core/JIT/x86_64/VectorOps.cpp
+++ b/External/FEXCore/Source/Interface/Core/JIT/x86_64/VectorOps.cpp
@@ -24,10 +24,11 @@ DEF_OP(VectorZero) {
 
 DEF_OP(VectorImm) {
   auto Op = IROp->C<IR::IROp_VectorImm>();
-  uint8_t OpSize = IROp->Size;
 
-  auto Dst = GetDst(Node);
-  uint64_t Imm = Op->Immediate;
+  const uint8_t OpSize = IROp->Size;
+  const uint64_t Imm = Op->Immediate;
+
+  const auto Dst = GetDst(Node);
 
   uint64_t Element{};
   switch (Op->Header.ElementSize) {
@@ -61,9 +62,13 @@ DEF_OP(VectorImm) {
 
   mov(TMP1, Element);
   vmovq(Dst, TMP1);
-  if (OpSize == 16) {
-    // Duplicate to the upper 64bits if we are 128bits
-    movddup(Dst, Dst);
+
+  if (OpSize >= 16) {
+    LOGMAN_THROW_AA_FMT(OpSize == 16 || OpSize == 32,
+                        "Can't handle a vector of size: {}", OpSize);
+
+    // Duplicate into upper elements
+    vbroadcastsd(ToYMM(Dst), Dst);
   }
 }