diff --git a/pcsx2/GS/GSAlignedClass.h b/pcsx2/GS/GSAlignedClass.h
index cafed33482..4fd1873e69 100644
--- a/pcsx2/GS/GSAlignedClass.h
+++ b/pcsx2/GS/GSAlignedClass.h
@@ -33,6 +33,11 @@ public:
 		_aligned_free(p);
 	}
 
+	void* operator new(size_t size, void* ptr)
+	{
+		return ptr;
+	}
+
 	void* operator new[](size_t size)
 	{
 		return _aligned_malloc(size, i);
diff --git a/pcsx2/GS/Renderers/SW/GSRasterizer.cpp b/pcsx2/GS/Renderers/SW/GSRasterizer.cpp
index 78ba755bc0..0c9b7f3b73 100644
--- a/pcsx2/GS/Renderers/SW/GSRasterizer.cpp
+++ b/pcsx2/GS/Renderers/SW/GSRasterizer.cpp
@@ -113,7 +113,7 @@ int GSRasterizer::FindMyNextScanline(int top) const
 	return top;
 }
 
-void GSRasterizer::Queue(const std::shared_ptr<GSRasterizerData>& data)
+void GSRasterizer::Queue(const GSRingHeap::SharedPtr<GSRasterizerData>& data)
 {
 	Draw(data.get());
 }
@@ -1203,7 +1203,7 @@ GSRasterizerList::~GSRasterizerList()
 	_aligned_free(m_scanline);
 }
 
-void GSRasterizerList::Queue(const std::shared_ptr<GSRasterizerData>& data)
+void GSRasterizerList::Queue(const GSRingHeap::SharedPtr<GSRasterizerData>& data)
 {
 	GSVector4i r = data->bbox.rintersect(data->scissor);
 
diff --git a/pcsx2/GS/Renderers/SW/GSRasterizer.h b/pcsx2/GS/Renderers/SW/GSRasterizer.h
index eca6670c6d..bdfce6b447 100644
--- a/pcsx2/GS/Renderers/SW/GSRasterizer.h
+++ b/pcsx2/GS/Renderers/SW/GSRasterizer.h
@@ -20,6 +20,7 @@
 #include "GS/GSAlignedClass.h"
 #include "GS/GSPerfMon.h"
 #include "GS/GSThread_CXX11.h"
+#include "GS/GSRingHeap.h"
 
 class alignas(32) GSRasterizerData : public GSAlignedClass<32>
 {
@@ -43,7 +44,7 @@ public:
 		: scissor(GSVector4i::zero())
 		, bbox(GSVector4i::zero())
 		, primclass(GS_INVALID_CLASS)
-		, buff(NULL)
+		, buff(nullptr)
 		, vertex(NULL)
 		, vertex_count(0)
 		, index(NULL)
@@ -58,7 +59,7 @@ public:
 	virtual ~GSRasterizerData()
 	{
 		if (buff != NULL)
-			_aligned_free(buff);
+			GSRingHeap::free(buff);
 	}
 };
 
@@ -115,7 +116,7 @@ class IRasterizer : public GSAlignedClass<32>
 public:
 	virtual ~IRasterizer() {}
 
-	virtual void Queue(const std::shared_ptr<GSRasterizerData>& data) = 0;
+	virtual void Queue(const GSRingHeap::SharedPtr<GSRasterizerData>& data) = 0;
 	virtual void Sync() = 0;
 	virtual bool IsSynced() const = 0;
 	virtual int GetPixels(bool reset = true) = 0;
@@ -172,7 +173,7 @@ public:
 
 	// IRasterizer
 
-	void Queue(const std::shared_ptr<GSRasterizerData>& data);
+	void Queue(const GSRingHeap::SharedPtr<GSRasterizerData>& data);
 	void Sync() {}
 	bool IsSynced() const { return true; }
 	int GetPixels(bool reset);
@@ -182,7 +183,7 @@ public:
 class GSRasterizerList : public IRasterizer
 {
 protected:
-	using GSWorker = GSJobQueue<std::shared_ptr<GSRasterizerData>, 65536>;
+	using GSWorker = GSJobQueue<GSRingHeap::SharedPtr<GSRasterizerData>, 65536>;
 
 	GSPerfMon* m_perfmon;
 	// Worker threads depend on the rasterizers, so don't change the order.
@@ -213,7 +214,7 @@ public:
 			rl->m_r.push_back(std::unique_ptr<GSRasterizer>(new GSRasterizer(new DS(), i, threads, perfmon)));
 			auto& r = *rl->m_r[i];
 			rl->m_workers.push_back(std::unique_ptr<GSWorker>(new GSWorker(
-				[&r](std::shared_ptr<GSRasterizerData>& item) { r.Draw(item.get()); })));
+				[&r](GSRingHeap::SharedPtr<GSRasterizerData>& item) { r.Draw(item.get()); })));
 		}
 
 		return rl;
@@ -221,7 +222,7 @@ public:
 
 	// IRasterizer
 
-	void Queue(const std::shared_ptr<GSRasterizerData>& data);
+	void Queue(const GSRingHeap::SharedPtr<GSRasterizerData>& data);
 	void Sync();
 	bool IsSynced() const;
 	int GetPixels(bool reset);
diff --git a/pcsx2/GS/Renderers/SW/GSRendererSW.cpp b/pcsx2/GS/Renderers/SW/GSRendererSW.cpp
index d039b8a961..be91157975 100644
--- a/pcsx2/GS/Renderers/SW/GSRendererSW.cpp
+++ b/pcsx2/GS/Renderers/SW/GSRendererSW.cpp
@@ -335,12 +335,11 @@ void GSRendererSW::Draw()
 {
 	const GSDrawingContext* context = m_context;
 
-	SharedData* sd = new SharedData(this);
-
-	std::shared_ptr<GSRasterizerData> data(sd);
+	auto data = m_vertex_heap.make_shared<SharedData>(this).cast<GSRasterizerData>();
+	SharedData* sd = static_cast<SharedData*>(data.get());
 
 	sd->primclass = m_vt.m_primclass;
-	sd->buff = (u8*)_aligned_malloc(sizeof(GSVertexSW) * ((m_vertex.next + 1) & ~1) + sizeof(u32) * m_index.tail, 64);
+	sd->buff = (u8*)m_vertex_heap.alloc(sizeof(GSVertexSW) * ((m_vertex.next + 1) & ~1) + sizeof(u32) * m_index.tail, 64);
 	sd->vertex = (GSVertexSW*)sd->buff;
 	sd->vertex_count = m_vertex.next;
 	sd->index = (u32*)(sd->buff + sizeof(GSVertexSW) * ((m_vertex.next + 1) & ~1));
@@ -540,7 +539,7 @@ void GSRendererSW::Draw()
 	*/
 }
 
-void GSRendererSW::Queue(std::shared_ptr<GSRasterizerData>& item)
+void GSRendererSW::Queue(GSRingHeap::SharedPtr<GSRasterizerData>& item)
 {
 	SharedData* sd = (SharedData*)item.get();
 
@@ -1050,7 +1049,7 @@ bool GSRendererSW::GetScanlineGlobalData(SharedData* data)
 			{
 				gd.sel.tlu = 1;
 
-				gd.clut = (u32*)_aligned_malloc(sizeof(u32) * 256, 32); // FIXME: might address uninitialized data of the texture (0xCD) that is not in 0-15 range for 4-bpp formats
+				gd.clut = (u32*)m_vertex_heap.alloc(sizeof(u32) * 256, 32); // FIXME: might address uninitialized data of the texture (0xCD) that is not in 0-15 range for 4-bpp formats
 
 				memcpy(gd.clut, (const u32*)m_mem.m_clut, sizeof(u32) * GSLocalMemory::m_psm[context->TEX0.PSM].pal);
 			}
@@ -1327,7 +1326,7 @@ bool GSRendererSW::GetScanlineGlobalData(SharedData* data)
 		{
 			gd.sel.dthe = 1;
 
-			gd.dimx = (GSVector4i*)_aligned_malloc(sizeof(env.dimx), 32);
+			gd.dimx = (GSVector4i*)m_vertex_heap.alloc(sizeof(env.dimx), 32);
 
 			memcpy(gd.dimx, env.dimx, sizeof(env.dimx));
 		}
@@ -1444,9 +1443,9 @@ GSRendererSW::SharedData::~SharedData()
 	ReleasePages();
 
 	if (global.clut)
-		_aligned_free(global.clut);
+		GSRingHeap::free(global.clut);
 	if (global.dimx)
-		_aligned_free(global.dimx);
+		GSRingHeap::free(global.dimx);
 
 	if (LOG)
 	{
diff --git a/pcsx2/GS/Renderers/SW/GSRendererSW.h b/pcsx2/GS/Renderers/SW/GSRendererSW.h
index c25ad45f92..6295b235de 100644
--- a/pcsx2/GS/Renderers/SW/GSRendererSW.h
+++ b/pcsx2/GS/Renderers/SW/GSRendererSW.h
@@ -17,6 +17,7 @@
 
 #include "GSTextureCacheSW.h"
 #include "GSDrawScanline.h"
+#include "GS/GSRingHeap.h"
 
 class GSRendererSW : public GSRenderer
 {
@@ -68,6 +69,7 @@ class GSRendererSW : public GSRenderer
 
 protected:
 	IRasterizer* m_rl;
+	GSRingHeap m_vertex_heap;
 	GSTextureCacheSW* m_tc;
 	GSTexture* m_texture[2];
 	u8* m_output;
@@ -84,7 +86,7 @@ protected:
 	GSTexture* GetFeedbackOutput();
 
 	void Draw();
-	void Queue(std::shared_ptr<GSRasterizerData>& item);
+	void Queue(GSRingHeap::SharedPtr<GSRasterizerData>& item);
 	void Sync(int reason);
 	void InvalidateVideoMem(const GIFRegBITBLTBUF& BITBLTBUF, const GSVector4i& r);
 	void InvalidateLocalMem(const GIFRegBITBLTBUF& BITBLTBUF, const GSVector4i& r, bool clut = false);