diff --git a/xla/service/gpu/intel_gpu_compiler.cc b/xla/service/gpu/intel_gpu_compiler.cc
index f3383ac824b1a..6c7ddc7198c19 100644
--- a/xla/service/gpu/intel_gpu_compiler.cc
+++ b/xla/service/gpu/intel_gpu_compiler.cc
@@ -54,7 +54,7 @@ IntelGpuCompiler::CompileTargetBinary(
     const HloModuleConfig& module_config, llvm::Module* llvm_module,
     const stream_executor::DeviceDescription& device_description,
     bool relocatable, const HloModule* debug_module,
-    const CompileOptions& options, std::optional<int> shard_number) {
+    std::optional<int> shard_number) {
   TF_ASSIGN_OR_RETURN(
       auto spirv_str,
       spirv::CompileToSPIRV(llvm_module,
diff --git a/xla/service/gpu/intel_gpu_compiler.h b/xla/service/gpu/intel_gpu_compiler.h
index 69d87f4f5a124..c74d9416b35fd 100644
--- a/xla/service/gpu/intel_gpu_compiler.h
+++ b/xla/service/gpu/intel_gpu_compiler.h
@@ -54,7 +54,7 @@ class IntelGpuCompiler : public GpuCompiler {
       const HloModuleConfig& module_config, llvm::Module* llvm_module,
       const stream_executor::DeviceDescription& device_description,
       bool relocatable, const HloModule* debug_module,
-      const CompileOptions& options, std::optional<int> shard_number) override;
+      std::optional<int> shard_number) override;
 
   std::vector<std::string> GetLLVMCommandLineOptions(
       const DebugOptions& debug_options) const override;