diff --git a/source/slang/hlsl.meta.slang b/source/slang/hlsl.meta.slang
index 7c68a641c1..bb8c20f2ee 100644
--- a/source/slang/hlsl.meta.slang
+++ b/source/slang/hlsl.meta.slang
@@ -7275,12 +7275,13 @@ vector<T, 3> cross(vector<T, 3> left, vector<T, 3> right)
 
 // Convert encoded color
 [__readNone]
-[require(cpp_cuda_glsl_hlsl_metal_spirv, sm_4_0_version)]
+[require(cpp_cuda_glsl_hlsl_metal_spirv_wgsl, sm_4_0_version)]
 int4 D3DCOLORtoUBYTE4(float4 color)
 {
     __target_switch
     {
     case hlsl: __intrinsic_asm "D3DCOLORtoUBYTE4";
+    case wgsl: __intrinsic_asm "bitcast<vec4i>(pack4x8unorm($0)).zyxw";
     default:
         let scaled = color.zyxw * 255.001999f;
         return int4(scaled);
@@ -8030,7 +8031,7 @@ vector<T,N> exp10(vector<T,N> x)
 __glsl_version(420)
 __cuda_sm_version(6.0)
 [__readNone]
-[require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
+[require(cpp_cuda_glsl_hlsl_metal_spirv_wgsl, shader5_sm_5_0)]
 float f16tof32(uint value)
 {
     __target_switch
@@ -8048,12 +8049,13 @@ float f16tof32(uint value)
             result:$$float = OpFConvert %half
         };
     }
+    case wgsl: __intrinsic_asm "unpack2x16float($0).x";
     }
 }
 
 __generic<let N : int>
 [__readNone]
-[require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
+[require(cpp_cuda_glsl_hlsl_metal_spirv_wgsl, shader5_sm_5_0)]
 vector<float, N> f16tof32(vector<uint, N> value)
 {
     __target_switch
@@ -8078,7 +8080,7 @@ vector<float, N> f16tof32(vector<uint, N> value)
 __glsl_version(420)
 __cuda_sm_version(6.0)
 [__readNone]
-[require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
+[require(cpp_cuda_glsl_hlsl_metal_spirv_wgsl, shader5_sm_5_0)]
 uint f32tof16(float value)
 {
     __target_switch
@@ -8096,12 +8098,13 @@ uint f32tof16(float value)
             result:$$uint = OpUConvert %lowBits
         };
     }
+    case wgsl: __intrinsic_asm "pack2x16float(vec2f($0,0.0))";
     }
 }
 
 __generic<let N : int>
 [__readNone]
-[require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
+[require(cpp_cuda_glsl_hlsl_metal_spirv_wgsl, shader5_sm_5_0)]
 vector<uint, N> f32tof16(vector<float, N> value)
 {
     __target_switch
@@ -8127,7 +8130,7 @@ vector<uint, N> f32tof16(vector<float, N> value)
 
 __glsl_version(420)
 [__readNone]
-[require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
+[require(cpp_cuda_glsl_hlsl_metal_spirv_wgsl, shader5_sm_5_0)]
 float f16tof32(float16_t value)
 {
     __target_switch
@@ -8143,12 +8146,13 @@ float f16tof32(float16_t value)
             result:$$float = OpFConvert $value
         };
     }
+    case wgsl: __intrinsic_asm "f32($0)";
     }
 }
 
 __generic<let N : int>
 [__readNone]
-[require(cpp_cuda_glsl_hlsl_metal_spirv, shader5_sm_5_0)]
+[require(cpp_cuda_glsl_hlsl_metal_spirv_wgsl, shader5_sm_5_0)]
 vector<float, N> f16tof32(vector<float16_t, N> value)
 {
     __target_switch
@@ -8167,7 +8171,7 @@ vector<float, N> f16tof32(vector<float16_t, N> value)
 // Convert to float16_t
 __glsl_version(420)
 [__readNone]
-[require(cuda_glsl_metal_spirv, shader5_sm_5_0)]
+[require(cuda_glsl_metal_spirv_wgsl, shader5_sm_5_0)]
 float16_t f32tof16_(float value)
 {
     __target_switch
@@ -8178,12 +8182,13 @@ float16_t f32tof16_(float value)
     case spirv: return spirv_asm {
         OpFConvert $$float16_t result $value
     };
+    case wgsl: __intrinsic_asm "f16($0)";
     }
 }
 
 __generic<let N : int>
 [__readNone]
-[require(cuda_glsl_metal_spirv, shader5_sm_5_0)]
+[require(cuda_glsl_metal_spirv_wgsl, shader5_sm_5_0)]
 vector<float16_t, N> f32tof16_(vector<float, N> value)
 {
     __target_switch
diff --git a/source/slang/slang-capabilities.capdef b/source/slang/slang-capabilities.capdef
index 3e17b3d19d..9364600411 100644
--- a/source/slang/slang-capabilities.capdef
+++ b/source/slang/slang-capabilities.capdef
@@ -330,6 +330,10 @@ alias cuda_glsl_spirv = cuda | glsl | spirv;
 /// [Compound]
 alias cuda_glsl_metal_spirv = cuda | glsl | metal | spirv;
 
+/// CUDA, GLSL, Metal, SPIRV and WGSL code-gen targets
+/// [Compound]
+alias cuda_glsl_metal_spirv_wgsl = cuda | glsl | metal | spirv | wgsl;
+
 /// CUDA, and HLSL code-gen targets
 /// [Compound]
 alias cuda_hlsl = cuda | hlsl;
diff --git a/source/slang/slang-emit-wgsl.cpp b/source/slang/slang-emit-wgsl.cpp
index 0a4cca407a..40d8ace91a 100644
--- a/source/slang/slang-emit-wgsl.cpp
+++ b/source/slang/slang-emit-wgsl.cpp
@@ -1002,4 +1002,22 @@ void WGSLSourceEmitter::emitFrontMatterImpl(TargetRequest* /* targetReq */)
     }
 }
 
+void WGSLSourceEmitter::emitIntrinsicCallExprImpl(
+    IRCall* inst,
+    UnownedStringSlice intrinsicDefinition,
+    IRInst* intrinsicInst,
+    EmitOpInfo const& inOuterPrec
+    )
+{
+    // The f16 constructor is generated for f32tof16
+    if (intrinsicDefinition.startsWith("f16"))
+    {
+        m_f16ExtensionEnabled = true;
+    }
+
+    CLikeSourceEmitter::emitIntrinsicCallExprImpl(
+        inst, intrinsicDefinition, intrinsicInst, inOuterPrec
+    );
+}
+
 } // namespace Slang
diff --git a/source/slang/slang-emit-wgsl.h b/source/slang/slang-emit-wgsl.h
index dacd11c3d3..d3cf19d91f 100644
--- a/source/slang/slang-emit-wgsl.h
+++ b/source/slang/slang-emit-wgsl.h
@@ -53,6 +53,13 @@ class WGSLSourceEmitter : public CLikeSourceEmitter
         IRInst* inst, const EmitOpInfo& outerPrec
     ) SLANG_OVERRIDE;
 
+    virtual void emitIntrinsicCallExprImpl(
+        IRCall* inst,
+        UnownedStringSlice intrinsicDefinition,
+        IRInst* intrinsicInst,
+        EmitOpInfo const& inOuterPrec
+    ) SLANG_OVERRIDE;
+
     void emit(const AddressSpace addressSpace);
 
 private: