diff --git a/rpcs3/Emu/RSX/VK/VKCompute.cpp b/rpcs3/Emu/RSX/VK/VKCompute.cpp
index 5a88537ce1..69f85e7c38 100644
--- a/rpcs3/Emu/RSX/VK/VKCompute.cpp
+++ b/rpcs3/Emu/RSX/VK/VKCompute.cpp
@@ -81,6 +81,7 @@ namespace vk
 			case vk::driver_vendor::DOZEN:
 				// Actual optimal size depends on the D3D device. Use 32 since it should work well on both AMD and NVIDIA
 			case vk::driver_vendor::NVIDIA:
+			case vk::driver_vendor::NVK:
 				// Warps are multiples of 32. Increasing kernel depth seems to hurt performance (Nier, Big Duck sample)
 				unroll_loops = true;
 				optimal_kernel_size = 1;
diff --git a/rpcs3/Emu/RSX/VK/VKFragmentProgram.cpp b/rpcs3/Emu/RSX/VK/VKFragmentProgram.cpp
index da77ce54aa..aab1406ac5 100644
--- a/rpcs3/Emu/RSX/VK/VKFragmentProgram.cpp
+++ b/rpcs3/Emu/RSX/VK/VKFragmentProgram.cpp
@@ -247,7 +247,7 @@ void VKFragmentDecompilerThread::insertGlobalFunctions(std::stringstream &OS)
 	m_shader_props.emulate_coverage_tests = g_cfg.video.antialiasing_level == msaa_level::none;
 	m_shader_props.emulate_shadow_compare = device_props.emulate_depth_compare;
 	m_shader_props.low_precision_tests = device_props.has_low_precision_rounding && !(m_prog.ctrl & RSX_SHADER_CONTROL_ATTRIBUTE_INTERPOLATION);
-	m_shader_props.disable_early_discard = vk::get_driver_vendor() != vk::driver_vendor::NVIDIA;
+	m_shader_props.disable_early_discard = vk::is_NVIDIA(vk::get_driver_vendor());
 	m_shader_props.supports_native_fp16 = device_props.has_native_half_support;
 	m_shader_props.ROP_output_rounding = g_cfg.video.shader_precision != gpu_preset_level::low;
 	m_shader_props.require_tex1D_ops = properties.has_tex1D;
@@ -402,7 +402,7 @@ void VKFragmentProgram::Decompile(const RSXFragmentProgram& prog)
 	}
 
 	decompiler.device_props.emulate_depth_compare = !pdev->get_formats_support().d24_unorm_s8;
-	decompiler.device_props.has_low_precision_rounding = vk::get_driver_vendor() == vk::driver_vendor::NVIDIA;
+	decompiler.device_props.has_low_precision_rounding = vk::is_NVIDIA(vk::get_driver_vendor());
 	decompiler.Task();
 
 	shader.create(::glsl::program_domain::glsl_fragment_program, source);
diff --git a/rpcs3/Emu/RSX/VK/VKGSRender.cpp b/rpcs3/Emu/RSX/VK/VKGSRender.cpp
index 4968b832e8..8c765abd5c 100644
--- a/rpcs3/Emu/RSX/VK/VKGSRender.cpp
+++ b/rpcs3/Emu/RSX/VK/VKGSRender.cpp
@@ -743,7 +743,7 @@ VKGSRender::VKGSRender(utils::serial* ar) noexcept : GSRender(ar)
 
 	// NVIDIA has broken attribute interpolation
 	backend_config.supports_normalized_barycentrics = (
-		vk::get_driver_vendor() != vk::driver_vendor::NVIDIA ||
+		!vk::is_NVIDIA(vk::get_driver_vendor()) ||
 		!m_device->get_barycoords_support() ||
 		g_cfg.video.shader_precision == gpu_preset_level::low);
 
@@ -758,7 +758,7 @@ VKGSRender::VKGSRender(utils::serial* ar) noexcept : GSRender(ar)
 
 	// NOTE: On NVIDIA cards going back decades (including the PS3) there is a slight normalization inaccuracy in compressed formats.
 	// Confirmed in BLES01916 (The Evil Within) which uses RGB565 for some virtual texturing data.
-	backend_config.supports_hw_renormalization = (vk::get_driver_vendor() == vk::driver_vendor::NVIDIA);
+	backend_config.supports_hw_renormalization = vk::is_NVIDIA(vk::get_driver_vendor());
 
 	// Conditional rendering support
 	// Do not use on MVK due to a speedhack we rely on (streaming results without stopping the current renderpass)
@@ -800,6 +800,10 @@ VKGSRender::VKGSRender(utils::serial* ar) noexcept : GSRender(ar)
 			g_cfg.video.vk.asynchronous_scheduler.set(vk_gpu_scheduler_mode::safe);
 		}
 		break;
+	case vk::driver_vendor::NVK:
+		// TODO: Verify if this driver crashes or not
+		rsx_log.warning("NVK behavior with passthrough DMA is unknown. Proceed with caution.");
+		break;
 #if !defined(_WIN32)
 		// Anything running on AMDGPU kernel driver will not work due to the check for fd-backed memory allocations
 	case vk::driver_vendor::RADV:
@@ -2407,7 +2411,8 @@ void VKGSRender::patch_transform_constants(rsx::context* ctx, u32 index, u32 cou
 		vk::driver_vendor::AMD,
 		vk::driver_vendor::RADV,
 		vk::driver_vendor::LAVAPIPE,
-		vk::driver_vendor::NVIDIA
+		vk::driver_vendor::NVIDIA,
+		vk::driver_vendor::NVK
 	};
 
 	const auto driver_vendor = vk::get_driver_vendor();
diff --git a/rpcs3/Emu/RSX/VK/VKVertexProgram.cpp b/rpcs3/Emu/RSX/VK/VKVertexProgram.cpp
index 2faba9d014..390ca0bb39 100644
--- a/rpcs3/Emu/RSX/VK/VKVertexProgram.cpp
+++ b/rpcs3/Emu/RSX/VK/VKVertexProgram.cpp
@@ -207,8 +207,8 @@ void VKVertexDecompilerThread::insertMainStart(std::stringstream & OS)
 	properties2.require_lit_emulation = properties.has_lit_op;
 	properties2.emulate_zclip_transform = true;
 	properties2.emulate_depth_clip_only = vk::g_render_device->get_shader_types_support().allow_float64;
-	properties2.low_precision_tests = vk::get_driver_vendor() == vk::driver_vendor::NVIDIA;
-	properties2.require_explicit_invariance = (vk::get_driver_vendor() == vk::driver_vendor::NVIDIA && g_cfg.video.shader_precision != gpu_preset_level::low);
+	properties2.low_precision_tests = vk::is_NVIDIA(vk::get_driver_vendor());
+	properties2.require_explicit_invariance = (vk::is_NVIDIA(vk::get_driver_vendor()) && g_cfg.video.shader_precision != gpu_preset_level::low);
 
 	glsl::insert_glsl_legacy_function(OS, properties2);
 	glsl::insert_vertex_input_fetch(OS, glsl::glsl_rules_vulkan);
diff --git a/rpcs3/Emu/RSX/VK/vkutils/chip_class.h b/rpcs3/Emu/RSX/VK/vkutils/chip_class.h
index 68c78e4eb8..7c33fb911b 100644
--- a/rpcs3/Emu/RSX/VK/vkutils/chip_class.h
+++ b/rpcs3/Emu/RSX/VK/vkutils/chip_class.h
@@ -9,12 +9,17 @@ namespace vk
 	enum class chip_class
 	{
 		unknown,
+
+		// AMD
 		AMD_gcn_generic,
 		AMD_polaris,
 		AMD_vega,
 		AMD_navi1x,
 		AMD_navi2x,
 		AMD_navi3x,
+		_AMD_ENUM_MAX_, // Do not insert AMD enums beyond this point
+
+		// NVIDIA
 		NV_generic,
 		NV_kepler,
 		NV_maxwell,
@@ -23,9 +28,15 @@ namespace vk
 		NV_turing,
 		NV_ampere,
 		NV_lovelace,
+		_NV_ENUM_MAX_, // Do not insert NV enums beyond this point
+
+		// APPLE
 		MVK_apple,
+
+		// INTEL
 		INTEL_generic,
-		INTEL_alchemist
+		INTEL_alchemist,
+		_INTEL_ENUM_MAX, // Do not insert INTEL enums beyond this point
 	};
 
 	enum class driver_vendor
@@ -38,7 +49,8 @@ namespace vk
 		ANV,
 		MVK,
 		DOZEN,
-		LAVAPIPE
+		LAVAPIPE,
+		NVK
 	};
 
 	driver_vendor get_driver_vendor();
@@ -57,5 +69,11 @@ namespace vk
 	chip_class get_chip_family();
 	chip_class get_chip_family(u32 vendor_id, u32 device_id);
 
-	static inline bool is_NVIDIA(chip_class chip) { return chip >= chip_class::NV_generic && chip <= chip_class::NV_ampere; }
+	static inline bool is_NVIDIA(chip_class chip) { return chip >= chip_class::NV_generic && chip < chip_class::_NV_ENUM_MAX_; }
+	static inline bool is_AMD(chip_class chip) { return chip >= chip_class::AMD_gcn_generic && chip < chip_class::_AMD_ENUM_MAX_; }
+	static inline bool is_INTEL(chip_class chip) { return chip >= chip_class::INTEL_generic && chip < chip_class::_INTEL_ENUM_MAX; }
+
+	static inline bool is_NVIDIA(driver_vendor vendor) { return vendor == driver_vendor::NVIDIA || vendor == driver_vendor::NVK; }
+	static inline bool is_AMD(driver_vendor vendor) { return vendor == driver_vendor::AMD || vendor == driver_vendor::RADV; }
+	static inline bool is_INTEL(driver_vendor vendor) { return vendor == driver_vendor::INTEL || vendor == driver_vendor::ANV; }
 }
diff --git a/rpcs3/Emu/RSX/VK/vkutils/device.cpp b/rpcs3/Emu/RSX/VK/vkutils/device.cpp
index 2cc5f25006..339b21015d 100644
--- a/rpcs3/Emu/RSX/VK/vkutils/device.cpp
+++ b/rpcs3/Emu/RSX/VK/vkutils/device.cpp
@@ -243,6 +243,11 @@ namespace vk
 
 			if (gpu_name.find("NVIDIA") != umax || gpu_name.find("GeForce") != umax || gpu_name.find("Quadro") != umax)
 			{
+				if (gpu_name.find("NVK") != umax)
+				{
+					return driver_vendor::NVK;
+				}
+
 				return driver_vendor::NVIDIA;
 			}
 
@@ -281,6 +286,8 @@ namespace vk
 				return driver_vendor::DOZEN;
 			case VK_DRIVER_ID_MESA_LLVMPIPE:
 				return driver_vendor::LAVAPIPE;
+			case VK_DRIVER_ID_MESA_NVK:
+				return driver_vendor::NVK;
 			default:
 				// Mobile?
 				return driver_vendor::unknown;