diff --git a/third_party/xla/xla/service/cpu/cpu_compiler.cc b/third_party/xla/xla/service/cpu/cpu_compiler.cc
index 02ecaae5d6d8ee..bd15a94d184af4 100644
--- a/third_party/xla/xla/service/cpu/cpu_compiler.cc
+++ b/third_party/xla/xla/service/cpu/cpu_compiler.cc
@@ -261,6 +261,13 @@ static AsyncValue::Executor* GetCompilationAsyncExecutor() {
   return executor;
 }
 
+// Returns task runner that uses the global compilation thread pool.
+static cpu::JitCompiler::TaskRunner GetCompilationTaskRunner() {
+  return [](cpu::JitCompiler::Task task) {
+    GetCompilationThreadPool()->Schedule(std::move(task));
+  };
+}
+
 // For each computation in the module, determines whether that computation
 // calls a custom-call function, either directly or indirectly (e.g. because it
 // calls another computation that does).
@@ -1353,6 +1360,40 @@ CpuCompiler::CompileLegacyCpuExecutable(std::unique_ptr<HloModule> module) {
       debug_options.xla_cpu_parallel_codegen_split_count();
   VlogMaxIsa(debug_options.xla_cpu_max_isa());
 
+  const HloModuleConfig& config = module->config();
+
+  // Options for compiling LLVM IR to machine code.
+  IrCompiler::Options ir_compiler_options{
+      /*optimization_level=*/static_cast<int32_t>(CodeGenOptLevel(config)),
+      /*optimize_for_size=*/options::OptimizeForSizeRequested(config),
+      /*fast_math_flags=*/llvm_ir::GetCpuFastMathFlags(config),
+      /*disable_expensive_passes=*/
+      debug_options.xla_llvm_disable_expensive_passes(),
+      /*slp_vectorizer_disabled=*/options::SlpVectorizerDisabled(config),
+  };
+
+  // Compiler hooks to intercept compiled LLVM IR modules.
+  IrCompiler::CompilationHooks ir_compiler_hooks{
+      pre_optimization_ir_hook,
+      post_optimization_ir_hook,
+      CreateOrcJITPostCompilationHook(module.get(), &obj_files),
+  };
+
+  // Options for orchestrating the JIT compilation process.
+  JitCompiler::Options jit_compiler_options{
+      std::move(ir_compiler_options),
+      std::move(ir_compiler_hooks),
+      /*num_dylibs=*/parallel_codegen_split_count,
+      /*max_cpu_isa=*/CpuFeatureFromString(debug_options.xla_cpu_max_isa()),
+  };
+
+  TF_ASSIGN_OR_RETURN(
+      JitCompiler jit_compiler,
+      JitCompiler::Create(CompilerTargetOptions(module->config()),
+                          CodeGenOptLevel(module->config()),
+                          std::move(jit_compiler_options),
+                          GetCompilationTaskRunner()));
+
   auto jit = SimpleOrcJIT::Create(
       CompilerTargetOptions(module->config()),
       CodeGenOptLevel(module->config()),
@@ -2072,6 +2113,33 @@ CpuExecutableAotCompilationResult::LoadExecutable(
 
   const DebugOptions& debug_options = module->config().debug_options();
   VlogMaxIsa(debug_options.xla_cpu_max_isa());
+  const HloModuleConfig& config = module->config();
+
+  // Options for compiling LLVM IR to machine code.
+  IrCompiler::Options ir_compiler_options{
+      /*optimization_level=*/static_cast<int32_t>(CodeGenOptLevel(config)),
+      /*optimize_for_size=*/options::OptimizeForSizeRequested(config),
+      /*fast_math_flags=*/llvm_ir::GetCpuFastMathFlags(config),
+      /*disable_expensive_passes=*/
+      debug_options.xla_llvm_disable_expensive_passes(),
+      /*slp_vectorizer_disabled=*/options::SlpVectorizerDisabled(config),
+  };
+
+  // Options for orchestrating the JIT compilation process.
+  JitCompiler::Options jit_compiler_options{
+      std::move(ir_compiler_options),
+      IrCompiler::CompilationHooks{},
+      /*num_dylibs=*/1,
+      /*max_cpu_isa=*/CpuFeatureFromString(debug_options.xla_cpu_max_isa()),
+  };
+
+  TF_ASSIGN_OR_RETURN(
+      JitCompiler jit_compiler,
+      JitCompiler::Create(CompilerTargetOptions(module->config()),
+                          CodeGenOptLevel(module->config()),
+                          std::move(jit_compiler_options),
+                          /*task_runner=*/nullptr));
+
   auto jit = SimpleOrcJIT::Create(
       CompilerTargetOptions(module->config()),
       CodeGenOptLevel(module->config()),