diff --git a/level_zero/core/source/gen11/enable_family_full_l0_gen11.cpp b/level_zero/core/source/gen11/enable_family_full_l0_gen11.cpp
index 67b2087ad2..4b54db04f6 100644
--- a/level_zero/core/source/gen11/enable_family_full_l0_gen11.cpp
+++ b/level_zero/core/source/gen11/enable_family_full_l0_gen11.cpp
@@ -10,8 +10,6 @@
 #include "shared/source/command_stream/tbx_command_stream_receiver_hw.h"
 #include "shared/source/helpers/populate_factory.h"
 
-#include "opencl/source/mem_obj/buffer.h"
-
 #include "level_zero/core/source/helpers/l0_populate_factory.h"
 #include "level_zero/core/source/hw_helpers/l0_hw_helper.h"
 
@@ -24,7 +22,6 @@ struct EnableL0Gen11 {
         populateFactoryTable<AUBCommandStreamReceiverHw<Family>>();
         populateFactoryTable<TbxCommandStreamReceiverHw<Family>>();
         populateFactoryTable<CommandStreamReceiverHw<Family>>();
-        populateFactoryTable<BufferHw<Family>>();
         L0::populateFactoryTable<L0::L0HwHelperHw<Family>>();
     }
 };
diff --git a/level_zero/core/source/gen12lp/enable_family_full_l0_gen12lp.cpp b/level_zero/core/source/gen12lp/enable_family_full_l0_gen12lp.cpp
index d6881b4bf9..839d6ff3f5 100644
--- a/level_zero/core/source/gen12lp/enable_family_full_l0_gen12lp.cpp
+++ b/level_zero/core/source/gen12lp/enable_family_full_l0_gen12lp.cpp
@@ -10,8 +10,6 @@
 #include "shared/source/command_stream/tbx_command_stream_receiver_hw.h"
 #include "shared/source/helpers/populate_factory.h"
 
-#include "opencl/source/mem_obj/buffer.h"
-
 #include "level_zero/core/source/helpers/l0_populate_factory.h"
 #include "level_zero/core/source/hw_helpers/l0_hw_helper.h"
 namespace NEO {
@@ -23,7 +21,6 @@ struct EnableL0Gen12LP {
         populateFactoryTable<AUBCommandStreamReceiverHw<Family>>();
         populateFactoryTable<TbxCommandStreamReceiverHw<Family>>();
         populateFactoryTable<CommandStreamReceiverHw<Family>>();
-        populateFactoryTable<BufferHw<Family>>();
         L0::populateFactoryTable<L0::L0HwHelperHw<Family>>();
     }
 };
diff --git a/level_zero/core/source/gen8/enable_family_full_l0_gen8.cpp b/level_zero/core/source/gen8/enable_family_full_l0_gen8.cpp
index 13a7f26380..4f5bef0094 100644
--- a/level_zero/core/source/gen8/enable_family_full_l0_gen8.cpp
+++ b/level_zero/core/source/gen8/enable_family_full_l0_gen8.cpp
@@ -10,8 +10,6 @@
 #include "shared/source/command_stream/tbx_command_stream_receiver_hw.h"
 #include "shared/source/helpers/populate_factory.h"
 
-#include "opencl/source/mem_obj/buffer.h"
-
 namespace NEO {
 
 typedef BDWFamily Family;
@@ -21,7 +19,6 @@ struct EnableL0Gen8 {
         populateFactoryTable<AUBCommandStreamReceiverHw<Family>>();
         populateFactoryTable<TbxCommandStreamReceiverHw<Family>>();
         populateFactoryTable<CommandStreamReceiverHw<Family>>();
-        populateFactoryTable<BufferHw<Family>>();
     }
 };
 
diff --git a/level_zero/core/source/gen9/enable_family_full_l0_gen9.cpp b/level_zero/core/source/gen9/enable_family_full_l0_gen9.cpp
index dbf4a0b2c0..8a6ac121d1 100644
--- a/level_zero/core/source/gen9/enable_family_full_l0_gen9.cpp
+++ b/level_zero/core/source/gen9/enable_family_full_l0_gen9.cpp
@@ -10,8 +10,6 @@
 #include "shared/source/command_stream/tbx_command_stream_receiver_hw.h"
 #include "shared/source/helpers/populate_factory.h"
 
-#include "opencl/source/mem_obj/buffer.h"
-
 #include "level_zero/core/source/helpers/l0_populate_factory.h"
 #include "level_zero/core/source/hw_helpers/l0_hw_helper.h"
 
@@ -24,7 +22,6 @@ struct EnableL0Gen9 {
         populateFactoryTable<AUBCommandStreamReceiverHw<Family>>();
         populateFactoryTable<TbxCommandStreamReceiverHw<Family>>();
         populateFactoryTable<CommandStreamReceiverHw<Family>>();
-        populateFactoryTable<BufferHw<Family>>();
         L0::populateFactoryTable<L0::L0HwHelperHw<Family>>();
     }
 };
diff --git a/level_zero/core/source/kernel/kernel_imp.cpp b/level_zero/core/source/kernel/kernel_imp.cpp
index 857a2649a1..4ecfa23367 100644
--- a/level_zero/core/source/kernel/kernel_imp.cpp
+++ b/level_zero/core/source/kernel/kernel_imp.cpp
@@ -25,8 +25,6 @@
 #include "shared/source/program/kernel_info.h"
 #include "shared/source/utilities/arrayref.h"
 
-#include "opencl/source/mem_obj/buffer.h"
-
 #include "level_zero/core/source/debugger/debugger_l0.h"
 #include "level_zero/core/source/device/device.h"
 #include "level_zero/core/source/device/device_imp.h"
@@ -85,17 +83,33 @@ KernelImmutableData::~KernelImmutableData() {
 
 inline void patchWithImplicitSurface(ArrayRef<uint8_t> crossThreadData, ArrayRef<uint8_t> surfaceStateHeap,
                                      uintptr_t ptrToPatchInCrossThreadData, NEO::GraphicsAllocation &allocation,
-                                     const NEO::ArgDescPointer &ptr, const NEO::Device &device, bool useGlobalAtomics) {
+                                     const NEO::ArgDescPointer &ptr, Device &device, bool useGlobalAtomics) {
     if (false == crossThreadData.empty()) {
         NEO::patchPointer(crossThreadData, ptr, ptrToPatchInCrossThreadData);
     }
 
     if ((false == surfaceStateHeap.empty()) && (NEO::isValidOffset(ptr.bindful))) {
         auto surfaceState = surfaceStateHeap.begin() + ptr.bindful;
-        void *addressToPatch = reinterpret_cast<void *>(allocation.getUnderlyingBuffer());
+        auto addressToPatch = allocation.getGpuAddress();
         size_t sizeToPatch = allocation.getUnderlyingBufferSize();
-        NEO::Buffer::setSurfaceState(&device, surfaceState, false, false, sizeToPatch, addressToPatch, 0,
-                                     &allocation, 0, 0, useGlobalAtomics, device.getNumGenericSubDevices() > 1);
+
+        auto neoDevice = device.getNEODevice();
+
+        auto &hwInfo = neoDevice->getHardwareInfo();
+        auto &hwHelper = NEO::HwHelper::get(hwInfo.platform.eRenderCoreFamily);
+
+        NEO::EncodeSurfaceStateArgs args;
+        args.outMemory = surfaceState;
+        args.size = sizeToPatch;
+        args.graphicsAddress = addressToPatch;
+        args.gmmHelper = neoDevice->getGmmHelper();
+        args.allocation = &allocation;
+        args.useGlobalAtomics = useGlobalAtomics;
+        args.numAvailableDevices = neoDevice->getNumGenericSubDevices();
+        args.areMultipleSubDevicesInContext = args.numAvailableDevices > 1;
+        args.mocs = device.getMOCS(true, false);
+
+        hwHelper.encodeBufferSurfaceState(args);
     }
 }
 
@@ -187,7 +201,7 @@ void KernelImmutableData::initialize(NEO::KernelInfo *kernelInfo, Device *device
         patchWithImplicitSurface(crossThredDataArrayRef, surfaceStateHeapArrayRef,
                                  static_cast<uintptr_t>(globalConstBuffer->getGpuAddressToPatch()),
                                  *globalConstBuffer, kernelDescriptor->payloadMappings.implicitArgs.globalConstantsSurfaceAddress,
-                                 *neoDevice, kernelDescriptor->kernelAttributes.flags.useGlobalAtomics);
+                                 *device, kernelDescriptor->kernelAttributes.flags.useGlobalAtomics);
         this->residencyContainer.push_back(globalConstBuffer);
     } else if (nullptr != globalConstBuffer) {
         this->residencyContainer.push_back(globalConstBuffer);
@@ -199,7 +213,7 @@ void KernelImmutableData::initialize(NEO::KernelInfo *kernelInfo, Device *device
         patchWithImplicitSurface(crossThredDataArrayRef, surfaceStateHeapArrayRef,
                                  static_cast<uintptr_t>(globalVarBuffer->getGpuAddressToPatch()),
                                  *globalVarBuffer, kernelDescriptor->payloadMappings.implicitArgs.globalVariablesSurfaceAddress,
-                                 *neoDevice, kernelDescriptor->kernelAttributes.flags.useGlobalAtomics);
+                                 *device, kernelDescriptor->kernelAttributes.flags.useGlobalAtomics);
         this->residencyContainer.push_back(globalVarBuffer);
     } else if (nullptr != globalVarBuffer) {
         this->residencyContainer.push_back(globalVarBuffer);
@@ -757,7 +771,7 @@ NEO::GraphicsAllocation *KernelImp::allocatePrivateMemoryGraphicsAllocation() {
 
 void KernelImp::patchCrossthreadDataWithPrivateAllocation(NEO::GraphicsAllocation *privateAllocation) {
     auto &kernelAttributes = kernelImmData->getDescriptor().kernelAttributes;
-    auto neoDevice = module->getDevice()->getNEODevice();
+    auto device = module->getDevice();
 
     ArrayRef<uint8_t> crossThredDataArrayRef = ArrayRef<uint8_t>(this->crossThreadData.get(), this->crossThreadDataSize);
     ArrayRef<uint8_t> surfaceStateHeapArrayRef = ArrayRef<uint8_t>(this->surfaceStateHeapData.get(), this->surfaceStateHeapDataSize);
@@ -765,7 +779,7 @@ void KernelImp::patchCrossthreadDataWithPrivateAllocation(NEO::GraphicsAllocatio
     patchWithImplicitSurface(crossThredDataArrayRef, surfaceStateHeapArrayRef,
                              static_cast<uintptr_t>(privateAllocation->getGpuAddressToPatch()),
                              *privateAllocation, kernelImmData->getDescriptor().payloadMappings.implicitArgs.privateMemoryAddress,
-                             *neoDevice, kernelAttributes.flags.useGlobalAtomics);
+                             *device, kernelAttributes.flags.useGlobalAtomics);
 }
 
 ze_result_t KernelImp::initialize(const ze_kernel_desc_t *desc) {
@@ -933,7 +947,7 @@ void KernelImp::setDebugSurface() {
         patchWithImplicitSurface(ArrayRef<uint8_t>(), surfaceStateHeapRef,
                                  0,
                                  *device->getDebugSurface(), this->getImmutableData()->getDescriptor().payloadMappings.implicitArgs.systemThreadSurfaceAddress,
-                                 *device->getNEODevice(), getKernelDescriptor().kernelAttributes.flags.useGlobalAtomics);
+                                 *device, getKernelDescriptor().kernelAttributes.flags.useGlobalAtomics);
     }
 }
 void *KernelImp::patchBindlessSurfaceState(NEO::GraphicsAllocation *alloc, uint32_t bindless) {
diff --git a/level_zero/core/source/xe_hp_core/enable_family_full_l0_xe_hp_core.cpp b/level_zero/core/source/xe_hp_core/enable_family_full_l0_xe_hp_core.cpp
index a66dc6c7da..7a94be2e42 100644
--- a/level_zero/core/source/xe_hp_core/enable_family_full_l0_xe_hp_core.cpp
+++ b/level_zero/core/source/xe_hp_core/enable_family_full_l0_xe_hp_core.cpp
@@ -10,8 +10,6 @@
 #include "shared/source/command_stream/tbx_command_stream_receiver_hw.h"
 #include "shared/source/helpers/populate_factory.h"
 
-#include "opencl/source/mem_obj/buffer.h"
-
 #include "level_zero/core/source/helpers/l0_populate_factory.h"
 #include "level_zero/core/source/hw_helpers/l0_hw_helper.h"
 
@@ -24,7 +22,6 @@ struct EnableL0XeHpCore {
         populateFactoryTable<AUBCommandStreamReceiverHw<Family>>();
         populateFactoryTable<TbxCommandStreamReceiverHw<Family>>();
         populateFactoryTable<CommandStreamReceiverHw<Family>>();
-        populateFactoryTable<BufferHw<Family>>();
         L0::populateFactoryTable<L0::L0HwHelperHw<Family>>();
     }
 };
diff --git a/shared/source/helpers/hw_helper.h b/shared/source/helpers/hw_helper.h
index 2f6fdf6083..954d62a37b 100644
--- a/shared/source/helpers/hw_helper.h
+++ b/shared/source/helpers/hw_helper.h
@@ -31,6 +31,7 @@ class TagAllocatorBase;
 class Gmm;
 struct AllocationData;
 struct AllocationProperties;
+struct EncodeSurfaceStateArgs;
 struct EngineControl;
 struct RootDeviceEnvironment;
 struct PipeControlArgs;
@@ -147,6 +148,7 @@ class HwHelper {
     virtual size_t getMax3dImageWidthOrHeight() const = 0;
     virtual uint64_t getMaxMemAllocSize() const = 0;
     virtual bool isStatelesToStatefullWithOffsetSupported() const = 0;
+    virtual void encodeBufferSurfaceState(EncodeSurfaceStateArgs &args) = 0;
 
   protected:
     HwHelper() = default;
@@ -369,6 +371,7 @@ class HwHelperHw : public HwHelper {
     size_t getMax3dImageWidthOrHeight() const override;
     uint64_t getMaxMemAllocSize() const override;
     bool isStatelesToStatefullWithOffsetSupported() const override;
+    void encodeBufferSurfaceState(EncodeSurfaceStateArgs &args) override;
 
   protected:
     static const AuxTranslationMode defaultAuxTranslationMode;
diff --git a/shared/source/helpers/hw_helper_base.inl b/shared/source/helpers/hw_helper_base.inl
index 93ec8876d4..39f105ac96 100644
--- a/shared/source/helpers/hw_helper_base.inl
+++ b/shared/source/helpers/hw_helper_base.inl
@@ -5,6 +5,7 @@
  *
  */
 
+#include "shared/source/command_container/command_encoder.h"
 #include "shared/source/execution_environment/root_device_environment.h"
 #include "shared/source/gmm_helper/gmm.h"
 #include "shared/source/gmm_helper/gmm_helper.h"
@@ -627,4 +628,10 @@ template <typename GfxFamily>
 uint64_t HwHelperHw<GfxFamily>::getRenderSurfaceStateBaseAddress(void *renderSurfaceState) const {
     return reinterpret_cast<typename GfxFamily::RENDER_SURFACE_STATE *>(renderSurfaceState)->getSurfaceBaseAddress();
 }
+
+template <typename GfxFamily>
+void HwHelperHw<GfxFamily>::encodeBufferSurfaceState(EncodeSurfaceStateArgs &args) {
+    EncodeSurfaceState<GfxFamily>::encodeBuffer(args);
+}
+
 } // namespace NEO