diff --git a/GPU/Common/VertexDecoderX86.cpp b/GPU/Common/VertexDecoderX86.cpp
index 85eb948de1..5330cc3708 100644
--- a/GPU/Common/VertexDecoderX86.cpp
+++ b/GPU/Common/VertexDecoderX86.cpp
@@ -191,7 +191,7 @@ JittedVertexDecoder VertexDecoderJitCache::Compile(const VertexDecoder &dec) {
 	// Later we might want to do this when the matrices are loaded instead.
 	// This is mostly proof of concept.
 	int boneCount = 0;
-	if (dec.weighttype && g_Config.bSoftwareSkinning) {
+	if (dec.weighttype && g_Config.bSoftwareSkinning && dec.morphcount == 1) {
 		MOVAPS(XMM4, M(&threeMasks));
 		for (int i = 0; i < 8; i++) {
 			MOVUPS(XMM0, M((gstate.boneMatrix + 12 * i)));
diff --git a/GPU/Directx9/GPU_DX9.cpp b/GPU/Directx9/GPU_DX9.cpp
index 0b523aa0cb..51e364c8be 100644
--- a/GPU/Directx9/GPU_DX9.cpp
+++ b/GPU/Directx9/GPU_DX9.cpp
@@ -707,6 +707,12 @@ void DIRECTX9_GPU::Execute_VertexTypeSkinning(u32 op, u32 diff) {
 		gstate.vertType ^= diff;
 		if (diff & (GE_VTYPE_TC_MASK | GE_VTYPE_THROUGH_MASK))
 			shaderManager_->DirtyUniform(DIRTY_UVSCALEOFFSET);
+		// In this case, we may be doing weights and morphs.
+		// Update any bone matrix uniforms so it uses them correctly.
+		if ((op & GE_VTYPE_MORPHCOUNT_MASK) != 0) {
+			shaderManager_->DirtyUniform(gstate_c.deferredVertTypeDirty);
+			gstate_c.deferredVertTypeDirty = 0;
+		}
 	}
 }
 
@@ -1269,6 +1275,11 @@ void DIRECTX9_GPU::Execute_BoneMtxNum(u32 op, u32 diff) {
 				break;
 			}
 		}
+
+		const int numPlusCount = (op & 0x7F) + i;
+		for (int num = op & 0x7F; num < numPlusCount; num += 12) {
+			gstate_c.deferredVertTypeDirty |= DIRTY_BONEMATRIX0 << (num / 12);
+		}
 	}
 
 	const int count = i;
@@ -1288,6 +1299,8 @@ void DIRECTX9_GPU::Execute_BoneMtxData(u32 op, u32 diff) {
 		if (!g_Config.bSoftwareSkinning || (gstate.vertType & GE_VTYPE_MORPHCOUNT_MASK) != 0) {
 			Flush();
 			shaderManager_->DirtyUniform(DIRTY_BONEMATRIX0 << (num / 12));
+		} else {
+			gstate_c.deferredVertTypeDirty |= DIRTY_BONEMATRIX0 << (num / 12);
 		}
 		((u32 *)gstate.boneMatrix)[num] = newVal;
 	}
@@ -1776,15 +1789,18 @@ void DIRECTX9_GPU::Execute_Generic(u32 op, u32 diff) {
 }
 
 void DIRECTX9_GPU::FastLoadBoneMatrix(u32 target) {
+	const int num = gstate.boneMatrixNumber & 0x7F;
+	const int mtxNum = num / 12;
+	uint32_t uniformsToDirty = DIRTY_BONEMATRIX0 << mtxNum;
+	if ((num - 12 * mtxNum) != 0) {
+		uniformsToDirty |= DIRTY_BONEMATRIX0 << ((mtxNum + 1) & 7);
+	}
+
 	if (!g_Config.bSoftwareSkinning || (gstate.vertType & GE_VTYPE_MORPHCOUNT_MASK) != 0) {
 		Flush();
-		const int num = gstate.boneMatrixNumber & 0x7F;
-		int mtxNum = num / 12;
-		uint32_t uniformsToDirty = DIRTY_BONEMATRIX0 << mtxNum;
-		if ((num - 12 * mtxNum) != 0) {
-			uniformsToDirty |= DIRTY_BONEMATRIX0 << ((mtxNum + 1) & 7);
-		}
 		shaderManager_->DirtyUniform(uniformsToDirty);
+	} else {
+		gstate_c.deferredVertTypeDirty |= uniformsToDirty;
 	}
 	gstate.FastLoadBoneMatrix(target);
 }
diff --git a/GPU/GLES/GLES_GPU.cpp b/GPU/GLES/GLES_GPU.cpp
index e5a39ddba0..bef26c0b30 100644
--- a/GPU/GLES/GLES_GPU.cpp
+++ b/GPU/GLES/GLES_GPU.cpp
@@ -868,6 +868,12 @@ void GLES_GPU::Execute_VertexTypeSkinning(u32 op, u32 diff) {
 		gstate.vertType ^= diff;
 		if (diff & (GE_VTYPE_TC_MASK | GE_VTYPE_THROUGH_MASK))
 			shaderManager_->DirtyUniform(DIRTY_UVSCALEOFFSET);
+		// In this case, we may be doing weights and morphs.
+		// Update any bone matrix uniforms so it uses them correctly.
+		if ((op & GE_VTYPE_MORPHCOUNT_MASK) != 0) {
+			shaderManager_->DirtyUniform(gstate_c.deferredVertTypeDirty);
+			gstate_c.deferredVertTypeDirty = 0;
+		}
 	}
 }
 
@@ -1346,6 +1352,11 @@ void GLES_GPU::Execute_BoneMtxNum(u32 op, u32 diff) {
 				break;
 			}
 		}
+
+		const int numPlusCount = (op & 0x7F) + i;
+		for (int num = op & 0x7F; num < numPlusCount; num += 12) {
+			gstate_c.deferredVertTypeDirty |= DIRTY_BONEMATRIX0 << (num / 12);
+		}
 	}
 
 	const int count = i;
@@ -1365,6 +1376,8 @@ void GLES_GPU::Execute_BoneMtxData(u32 op, u32 diff) {
 		if (!g_Config.bSoftwareSkinning || (gstate.vertType & GE_VTYPE_MORPHCOUNT_MASK) != 0) {
 			Flush();
 			shaderManager_->DirtyUniform(DIRTY_BONEMATRIX0 << (num / 12));
+		} else {
+			gstate_c.deferredVertTypeDirty |= DIRTY_BONEMATRIX0 << (num / 12);
 		}
 		((u32 *)gstate.boneMatrix)[num] = newVal;
 	}
@@ -1915,15 +1928,18 @@ void GLES_GPU::Execute_Generic(u32 op, u32 diff) {
 }
 
 void GLES_GPU::FastLoadBoneMatrix(u32 target) {
+	const int num = gstate.boneMatrixNumber & 0x7F;
+	const int mtxNum = num / 12;
+	uint32_t uniformsToDirty = DIRTY_BONEMATRIX0 << mtxNum;
+	if ((num - 12 * mtxNum) != 0) {
+		uniformsToDirty |= DIRTY_BONEMATRIX0 << ((mtxNum + 1) & 7);
+	}
+
 	if (!g_Config.bSoftwareSkinning || (gstate.vertType & GE_VTYPE_MORPHCOUNT_MASK) != 0) {
 		Flush();
-		const int num = gstate.boneMatrixNumber & 0x7F;
-		int mtxNum = num / 12;
-		uint32_t uniformsToDirty = DIRTY_BONEMATRIX0 << mtxNum;
-		if ((num - 12 * mtxNum) != 0) {
-			uniformsToDirty |= DIRTY_BONEMATRIX0 << ((mtxNum + 1) & 7);
-		}
 		shaderManager_->DirtyUniform(uniformsToDirty);
+	} else {
+		gstate_c.deferredVertTypeDirty |= uniformsToDirty;
 	}
 	gstate.FastLoadBoneMatrix(target);
 }
diff --git a/GPU/GPUState.h b/GPU/GPUState.h
index 81c367c370..a23d4e615b 100644
--- a/GPU/GPUState.h
+++ b/GPU/GPUState.h
@@ -469,6 +469,7 @@ struct GPUStateCache
 	bool allowShaderBlend;
 
 	float morphWeights[8];
+	u32 deferredVertTypeDirty;
 
 	u32 curTextureWidth;
 	u32 curTextureHeight;