diff --git a/src/core/game_list.cpp b/src/core/game_list.cpp
index 814cb3d4b..533c3be12 100644
--- a/src/core/game_list.cpp
+++ b/src/core/game_list.cpp
@@ -1033,7 +1033,7 @@ TinyString GameList::FormatTimespan(std::time_t timespan, bool long_format)
 }
 
 std::vector<std::pair<std::string, const GameList::Entry*>>
-GameList::GetMatchingEntriesForSerial(const gsl::span<const std::string> serials)
+GameList::GetMatchingEntriesForSerial(const std::span<const std::string> serials)
 {
   std::vector<std::pair<std::string, const GameList::Entry*>> ret;
   ret.reserve(serials.size());
diff --git a/src/core/game_list.h b/src/core/game_list.h
index 802e69fb5..1d1a7d98d 100644
--- a/src/core/game_list.h
+++ b/src/core/game_list.h
@@ -10,11 +10,10 @@
 
 #include "common/string.h"
 
-#include "gsl/span"
-
 #include <ctime>
 #include <functional>
 #include <mutex>
+#include <span>
 #include <string>
 
 class ByteStream;
@@ -107,7 +106,7 @@ std::string GetNewCoverImagePathForEntry(const Entry* entry, const char* new_fil
 /// Returns a list of (title, entry) for entries matching serials. Titles will match the gamedb title,
 /// except when two files have the same serial, in which case the filename will be used instead.
 std::vector<std::pair<std::string, const Entry*>>
-GetMatchingEntriesForSerial(const gsl::span<const std::string> serials);
+GetMatchingEntriesForSerial(const std::span<const std::string> serials);
 
 /// Downloads covers using the specified URL templates. By default, covers are saved by title, but this can be changed
 /// with the use_serial parameter. save_callback optionall takes the entry and the path the new cover is saved to.
diff --git a/src/core/gpu_hw.cpp b/src/core/gpu_hw.cpp
index 2b1fcb885..5f5a3354f 100644
--- a/src/core/gpu_hw.cpp
+++ b/src/core/gpu_hw.cpp
@@ -774,11 +774,11 @@ bool GPU_HW::CompilePipelines()
 
               plconfig.input_layout.vertex_attributes =
                 textured ?
-                  (m_using_uv_limits ? gsl::span<const GPUPipeline::VertexAttribute>(
+                  (m_using_uv_limits ? std::span<const GPUPipeline::VertexAttribute>(
                                          vertex_attributes, NUM_BATCH_TEXTURED_LIMITS_VERTEX_ATTRIBUTES) :
-                                       gsl::span<const GPUPipeline::VertexAttribute>(
+                                       std::span<const GPUPipeline::VertexAttribute>(
                                          vertex_attributes, NUM_BATCH_TEXTURED_VERTEX_ATTRIBUTES)) :
-                  gsl::span<const GPUPipeline::VertexAttribute>(vertex_attributes, NUM_BATCH_VERTEX_ATTRIBUTES);
+                  std::span<const GPUPipeline::VertexAttribute>(vertex_attributes, NUM_BATCH_VERTEX_ATTRIBUTES);
 
               plconfig.vertex_shader = batch_vertex_shaders[BoolToUInt8(textured)].get();
               plconfig.fragment_shader =
@@ -856,7 +856,7 @@ bool GPU_HW::CompilePipelines()
     GL_OBJECT_NAME(fs, "Batch Wireframe Fragment Shader");
 
     plconfig.input_layout.vertex_attributes =
-      gsl::span<const GPUPipeline::VertexAttribute>(vertex_attributes, NUM_BATCH_VERTEX_ATTRIBUTES);
+      std::span<const GPUPipeline::VertexAttribute>(vertex_attributes, NUM_BATCH_VERTEX_ATTRIBUTES);
     plconfig.blend = (m_wireframe_mode == GPUWireframeMode::OverlayWireframe) ?
                        GPUPipeline::BlendState::GetAlphaBlendingState() :
                        GPUPipeline::BlendState::GetNoBlendingState();
diff --git a/src/core/imgui_overlays.cpp b/src/core/imgui_overlays.cpp
index a13cc5df9..9484ac508 100644
--- a/src/core/imgui_overlays.cpp
+++ b/src/core/imgui_overlays.cpp
@@ -30,7 +30,6 @@
 #include "IconsFontAwesome5.h"
 #include "fmt/chrono.h"
 #include "fmt/format.h"
-#include "gsl/span"
 #include "imgui.h"
 #include "imgui_internal.h"
 
@@ -39,6 +38,7 @@
 #include <cmath>
 #include <deque>
 #include <mutex>
+#include <span>
 #include <unordered_map>
 
 #if defined(CPU_X64)
@@ -64,7 +64,7 @@ namespace SaveStateSelectorUI {
 static void Draw();
 }
 
-static std::tuple<float, float> GetMinMax(gsl::span<const float> values)
+static std::tuple<float, float> GetMinMax(std::span<const float> values)
 {
 #if defined(CPU_X64)
   __m128 vmin(_mm_loadu_ps(values.data()));
diff --git a/src/util/d3d11_device.h b/src/util/d3d11_device.h
index bf696bacc..6a522c206 100644
--- a/src/util/d3d11_device.h
+++ b/src/util/d3d11_device.h
@@ -68,7 +68,7 @@ public:
 
   std::unique_ptr<GPUFramebuffer> CreateFramebuffer(GPUTexture* rt_or_ds, GPUTexture* ds = nullptr) override;
 
-  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data) override;
+  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data) override;
   std::unique_ptr<GPUShader> CreateShaderFromSource(GPUShaderStage stage, const std::string_view& source,
                                                     const char* entry_point, DynamicHeapArray<u8>* binary) override;
   std::unique_ptr<GPUPipeline> CreatePipeline(const GPUPipeline::GraphicsConfig& config) override;
diff --git a/src/util/d3d11_pipeline.cpp b/src/util/d3d11_pipeline.cpp
index c115366c3..299f3c50c 100644
--- a/src/util/d3d11_pipeline.cpp
+++ b/src/util/d3d11_pipeline.cpp
@@ -51,7 +51,7 @@ void D3D11Shader::SetDebugName(const std::string_view& name)
   SetD3DDebugObjectName(m_shader.Get(), name);
 }
 
-std::unique_ptr<GPUShader> D3D11Device::CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data)
+std::unique_ptr<GPUShader> D3D11Device::CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data)
 {
   ComPtr<ID3D11DeviceChild> shader;
   std::vector<u8> bytecode;
diff --git a/src/util/d3d12_device.h b/src/util/d3d12_device.h
index 2b5651b3e..591fb342d 100644
--- a/src/util/d3d12_device.h
+++ b/src/util/d3d12_device.h
@@ -88,7 +88,7 @@ public:
 
   std::unique_ptr<GPUFramebuffer> CreateFramebuffer(GPUTexture* rt_or_ds, GPUTexture* ds = nullptr) override;
 
-  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data) override;
+  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data) override;
   std::unique_ptr<GPUShader> CreateShaderFromSource(GPUShaderStage stage, const std::string_view& source,
                                                     const char* entry_point, DynamicHeapArray<u8>* out_binary) override;
   std::unique_ptr<GPUPipeline> CreatePipeline(const GPUPipeline::GraphicsConfig& config) override;
diff --git a/src/util/d3d12_pipeline.cpp b/src/util/d3d12_pipeline.cpp
index 47c188a5f..f5ba720a4 100644
--- a/src/util/d3d12_pipeline.cpp
+++ b/src/util/d3d12_pipeline.cpp
@@ -25,7 +25,7 @@ void D3D12Shader::SetDebugName(const std::string_view& name)
 {
 }
 
-std::unique_ptr<GPUShader> D3D12Device::CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data)
+std::unique_ptr<GPUShader> D3D12Device::CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data)
 {
   // Can't do much at this point.
   std::vector bytecode(data.begin(), data.end());
diff --git a/src/util/gpu_device.h b/src/util/gpu_device.h
index 8bd633487..55bf8e5df 100644
--- a/src/util/gpu_device.h
+++ b/src/util/gpu_device.h
@@ -12,10 +12,9 @@
 #include "common/rectangle.h"
 #include "common/types.h"
 
-#include "gsl/span"
-
 #include <memory>
 #include <optional>
+#include <span>
 #include <string>
 #include <string_view>
 #include <tuple>
@@ -231,7 +230,7 @@ public:
 
   struct InputLayout
   {
-    gsl::span<const VertexAttribute> vertex_attributes;
+    std::span<const VertexAttribute> vertex_attributes;
     u32 vertex_stride;
 
     bool operator==(const InputLayout& rhs) const;
@@ -634,7 +633,7 @@ protected:
   virtual bool ReadPipelineCache(const std::string& filename);
   virtual bool GetPipelineCacheData(DynamicHeapArray<u8>* data);
 
-  virtual std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data) = 0;
+  virtual std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data) = 0;
   virtual std::unique_ptr<GPUShader> CreateShaderFromSource(GPUShaderStage stage, const std::string_view& source,
                                                             const char* entry_point,
                                                             DynamicHeapArray<u8>* out_binary) = 0;
diff --git a/src/util/metal_device.h b/src/util/metal_device.h
index 0f6de6815..50f6f50cb 100644
--- a/src/util/metal_device.h
+++ b/src/util/metal_device.h
@@ -224,7 +224,7 @@ public:
 
   std::unique_ptr<GPUFramebuffer> CreateFramebuffer(GPUTexture* rt_or_ds, GPUTexture* ds = nullptr) override;
 
-  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data) override;
+  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data) override;
   std::unique_ptr<GPUShader> CreateShaderFromSource(GPUShaderStage stage, const std::string_view& source,
                                                     const char* entry_point,
                                                     DynamicHeapArray<u8>* out_binary = nullptr) override;
diff --git a/src/util/metal_device.mm b/src/util/metal_device.mm
index 9ab6fbaa4..f5a044539 100644
--- a/src/util/metal_device.mm
+++ b/src/util/metal_device.mm
@@ -637,7 +637,7 @@ std::unique_ptr<GPUShader> MetalDevice::CreateShaderFromMSL(GPUShaderStage stage
   }
 }
 
-std::unique_ptr<GPUShader> MetalDevice::CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data)
+std::unique_ptr<GPUShader> MetalDevice::CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data)
 {
   const std::string_view str_data(reinterpret_cast<const char*>(data.data()), data.size());
   return CreateShaderFromMSL(stage, str_data, "main0");
diff --git a/src/util/opengl_device.h b/src/util/opengl_device.h
index de7f0bb12..960a30706 100644
--- a/src/util/opengl_device.h
+++ b/src/util/opengl_device.h
@@ -68,7 +68,7 @@ public:
 
   std::unique_ptr<GPUFramebuffer> CreateFramebuffer(GPUTexture* rt_or_ds, GPUTexture* ds = nullptr) override;
 
-  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data) override;
+  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data) override;
   std::unique_ptr<GPUShader> CreateShaderFromSource(GPUShaderStage stage, const std::string_view& source,
                                                     const char* entry_point, DynamicHeapArray<u8>* out_binary) override;
   std::unique_ptr<GPUPipeline> CreatePipeline(const GPUPipeline::GraphicsConfig& config) override;
@@ -111,7 +111,7 @@ public:
   void UnrefProgram(const OpenGLPipeline::ProgramCacheKey& key);
 
   GLuint LookupVAOCache(const OpenGLPipeline::VertexArrayCacheKey& key);
-  GLuint CreateVAO(gsl::span<const GPUPipeline::VertexAttribute> attributes, u32 stride);
+  GLuint CreateVAO(std::span<const GPUPipeline::VertexAttribute> attributes, u32 stride);
   void UnrefVAO(const OpenGLPipeline::VertexArrayCacheKey& key);
 
   void SetActiveTexture(u32 slot);
diff --git a/src/util/opengl_pipeline.cpp b/src/util/opengl_pipeline.cpp
index 995e877ff..3aaed538e 100644
--- a/src/util/opengl_pipeline.cpp
+++ b/src/util/opengl_pipeline.cpp
@@ -168,7 +168,7 @@ bool OpenGLShader::Compile()
   return true;
 }
 
-std::unique_ptr<GPUShader> OpenGLDevice::CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data)
+std::unique_ptr<GPUShader> OpenGLDevice::CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data)
 {
   // Not supported.. except spir-v maybe? but no point really...
   return {};
@@ -455,7 +455,7 @@ GLuint OpenGLDevice::LookupVAOCache(const OpenGLPipeline::VertexArrayCacheKey& k
 
   OpenGLPipeline::VertexArrayCacheItem item;
   item.vao_id =
-    CreateVAO(gsl::span<const GPUPipeline::VertexAttribute>(key.vertex_attributes, key.num_vertex_attributes),
+    CreateVAO(std::span<const GPUPipeline::VertexAttribute>(key.vertex_attributes, key.num_vertex_attributes),
               key.vertex_attribute_stride);
   if (item.vao_id == 0)
     return 0;
@@ -465,7 +465,7 @@ GLuint OpenGLDevice::LookupVAOCache(const OpenGLPipeline::VertexArrayCacheKey& k
   return item.vao_id;
 }
 
-GLuint OpenGLDevice::CreateVAO(gsl::span<const GPUPipeline::VertexAttribute> attributes, u32 stride)
+GLuint OpenGLDevice::CreateVAO(std::span<const GPUPipeline::VertexAttribute> attributes, u32 stride)
 {
   glGetError();
   GLuint vao;
diff --git a/src/util/pch.h b/src/util/pch.h
index 43c2fc050..4d1f0bb5f 100644
--- a/src/util/pch.h
+++ b/src/util/pch.h
@@ -6,7 +6,6 @@
 #include "common/types.h"
 
 #include "fmt/format.h"
-#include "gsl/span"
 
 #include <array>
 #include <atomic>
@@ -18,6 +17,7 @@
 #include <memory>
 #include <mutex>
 #include <optional>
+#include <span>
 #include <string>
 #include <string_view>
 #include <tuple>
diff --git a/src/util/spirv_compiler.cpp b/src/util/spirv_compiler.cpp
index 198685483..56e120685 100644
--- a/src/util/spirv_compiler.cpp
+++ b/src/util/spirv_compiler.cpp
@@ -173,7 +173,7 @@ std::optional<SPIRVCompiler::SPIRVCodeVector> SPIRVCompiler::CompileShader(GPUSh
 
 #ifdef __APPLE__
 
-std::optional<std::string> SPIRVCompiler::CompileSPIRVToMSL(gsl::span<const SPIRVCodeType> spv)
+std::optional<std::string> SPIRVCompiler::CompileSPIRVToMSL(std::span<const SPIRVCodeType> spv)
 {
   spirv_cross::CompilerMSL compiler(spv.data(), spv.size());
   std::string msl = compiler.compile();
diff --git a/src/util/spirv_compiler.h b/src/util/spirv_compiler.h
index 01e71f2ae..1e4885421 100644
--- a/src/util/spirv_compiler.h
+++ b/src/util/spirv_compiler.h
@@ -5,9 +5,8 @@
 
 #include "common/types.h"
 
-#include "gsl/span"
-
 #include <optional>
+#include <span>
 #include <string>
 #include <string_view>
 #include <vector>
@@ -43,7 +42,7 @@ std::optional<SPIRVCodeVector> CompileShader(GPUShaderStage stage, std::string_v
 #ifdef __APPLE__
 
 // Converts a SPIR-V shader into MSL.
-std::optional<std::string> CompileSPIRVToMSL(gsl::span<const SPIRVCodeType> spv);
+std::optional<std::string> CompileSPIRVToMSL(std::span<const SPIRVCodeType> spv);
 
 #endif
 
diff --git a/src/util/vulkan_device.h b/src/util/vulkan_device.h
index c72e0fba4..850969360 100644
--- a/src/util/vulkan_device.h
+++ b/src/util/vulkan_device.h
@@ -89,7 +89,7 @@ public:
 
   std::unique_ptr<GPUFramebuffer> CreateFramebuffer(GPUTexture* rt_or_ds, GPUTexture* ds = nullptr) override;
 
-  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data) override;
+  std::unique_ptr<GPUShader> CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data) override;
   std::unique_ptr<GPUShader> CreateShaderFromSource(GPUShaderStage stage, const std::string_view& source,
                                                     const char* entry_point, DynamicHeapArray<u8>* out_binary) override;
   std::unique_ptr<GPUPipeline> CreatePipeline(const GPUPipeline::GraphicsConfig& config) override;
diff --git a/src/util/vulkan_pipeline.cpp b/src/util/vulkan_pipeline.cpp
index 610ad9991..e5ee2996a 100644
--- a/src/util/vulkan_pipeline.cpp
+++ b/src/util/vulkan_pipeline.cpp
@@ -25,7 +25,7 @@ void VulkanShader::SetDebugName(const std::string_view& name)
   Vulkan::SetObjectName(VulkanDevice::GetInstance().GetVulkanDevice(), m_module, name);
 }
 
-std::unique_ptr<GPUShader> VulkanDevice::CreateShaderFromBinary(GPUShaderStage stage, gsl::span<const u8> data)
+std::unique_ptr<GPUShader> VulkanDevice::CreateShaderFromBinary(GPUShaderStage stage, std::span<const u8> data)
 {
   VkShaderModule mod;
 
@@ -67,7 +67,7 @@ std::unique_ptr<GPUShader> VulkanDevice::CreateShaderFromSource(GPUShaderStage s
     std::memcpy(out_binary->data(), spirv->data(), spirv_size);
   }
 
-  return CreateShaderFromBinary(stage, gsl::span<const u8>(reinterpret_cast<const u8*>(spirv->data()), spirv_size));
+  return CreateShaderFromBinary(stage, std::span<const u8>(reinterpret_cast<const u8*>(spirv->data()), spirv_size));
 }
 
 //////////////////////////////////////////////////////////////////////////