diff --git a/tensorflow/core/framework/op_requires.h b/tensorflow/core/framework/op_requires.h
index b7cf6e859fb..d186df337d7 100644
--- a/tensorflow/core/framework/op_requires.h
+++ b/tensorflow/core/framework/op_requires.h
@@ -57,6 +57,18 @@ namespace tensorflow {
     }                                                        \
   } while (0)
 
+#define OP_REQUIRES_OK_OR_SET_PAYLOAD(CTX, PAYLOAD_KEY, PAYLOAD_VALUE, STATUS) \
+  do {                                                                         \
+    if (!TF_PREDICT_TRUE(STATUS.ok())) {                                       \
+      CheckNotInComputeAsync((CTX), "OP_REQUIRES_OK_ASYNC");                   \
+      if (!PAYLOAD_VALUE.empty()) {                                            \
+        STATUS.SetPayload(PAYLOAD_KEY, PAYLOAD_VALUE);                         \
+      }                                                                        \
+      (CTX)->CtxFailureWithWarning(__FILE__, __LINE__, STATUS);                \
+      return;                                                                  \
+    }                                                                          \
+  } while (0)
+
 #define OP_REQUIRES_ASYNC(CTX, EXP, STATUS, CALLBACK)  \
   do {                                                 \
     if (!TF_PREDICT_TRUE(EXP)) {                       \
diff --git a/tensorflow/core/protobuf/tpu/compilation_result.proto b/tensorflow/core/protobuf/tpu/compilation_result.proto
index 5dd74dead49..d88bddf9173 100644
--- a/tensorflow/core/protobuf/tpu/compilation_result.proto
+++ b/tensorflow/core/protobuf/tpu/compilation_result.proto
@@ -7,7 +7,8 @@ import "tensorflow/core/protobuf/error_codes.proto";
 
 option cc_enable_arenas = true;
 
-// Describes the result of a TPU compilation.
+// Describes the result of a TPU compilation. This is also used as TPU
+// compilation result status payload.
 message CompilationResultProto {
   // The error message, if any, returned during compilation.
   error.Code status_code = 1;
diff --git a/tensorflow/core/tpu/kernels/BUILD b/tensorflow/core/tpu/kernels/BUILD
index 6461db0cfd0..b187ca31970 100644
--- a/tensorflow/core/tpu/kernels/BUILD
+++ b/tensorflow/core/tpu/kernels/BUILD
@@ -56,6 +56,7 @@ cc_library(
         ":tpu_op_util",
         ":tpu_program_group_interface",
         ":tpu_util",
+        "//tensorflow/core/tpu:tpu_compile_interface",
         "//tensorflow/core/tpu:tpu_ops_c_api_hdrs",
         ":tpu_util_hdrs",
         "@com_google_absl//absl/strings",
diff --git a/tensorflow/core/tpu/kernels/tpu_compile_op_common.cc b/tensorflow/core/tpu/kernels/tpu_compile_op_common.cc
index fa98ab763ca..55d34622912 100644
--- a/tensorflow/core/tpu/kernels/tpu_compile_op_common.cc
+++ b/tensorflow/core/tpu/kernels/tpu_compile_op_common.cc
@@ -28,6 +28,7 @@ limitations under the License.
 #include "tensorflow/core/framework/metrics.h"
 #include "tensorflow/core/framework/resource_mgr.h"
 #include "tensorflow/core/lib/core/errors.h"
+#include "tensorflow/core/platform/status.h"
 #include "tensorflow/core/protobuf/tpu/compilation_result.pb.h"
 #include "tensorflow/core/protobuf/tpu/compile_metadata.pb.h"
 #include "tensorflow/core/protobuf/tpu/dynamic_padding.pb.h"
@@ -40,6 +41,7 @@ limitations under the License.
 #include "tensorflow/core/tpu/kernels/tpu_program_group_interface.h"
 #include "tensorflow/core/tpu/kernels/tpu_util.h"
 #include "tensorflow/core/tpu/tpu_api.h"
+#include "tensorflow/core/tpu/tpu_compile_interface.h"
 #include "tensorflow/core/tpu/tpu_configuration.h"
 #include "tensorflow/core/tpu/tpu_defs.h"
 #include "tensorflow/core/tpu/tpu_ops_c_api.h"
@@ -568,7 +570,21 @@ void TpuCompileOpKernelCommon::Compute(OpKernelContext* ctx) {
     done->store(true);
   });
 
-  OP_REQUIRES_OK(ctx, ComputeInternal(ctx));
+  Status compile_status = ComputeInternal(ctx);
+  string status_payload;
+  // Construct payload if compile_status is not ok and there's no payload for
+  // compilation yet.
+  if (!compile_status.ok() &&
+      compile_status.GetPayload(TpuCompileInterface::kTpuCompileErrorPayloadKey)
+          .empty()) {
+    tpu::CompilationResultProto proto;
+    proto.set_status_code(compile_status.code());
+    proto.set_status_error_message(compile_status.error_message());
+    status_payload = proto.SerializeAsString();
+  }
+  OP_REQUIRES_OK_OR_SET_PAYLOAD(ctx,
+                                TpuCompileInterface::kTpuCompileErrorPayloadKey,
+                                status_payload, compile_status);
 }
 
 Status TpuCompileOpKernelCommon::CompileLocallyAndFillHostCache(
@@ -788,6 +804,8 @@ Status TpuCompileOpKernelCommon::ComputeInternal(OpKernelContext* ctx) {
     }
     SerializeToTString(proto, &output.scalar<tstring>()());
     ctx->set_output(0, output);
+    status.SetPayload(TpuCompileInterface::kTpuCompileErrorPayloadKey,
+                      output.scalar<tstring>()());
   }
 
   if (status.ok()) {
@@ -841,7 +859,7 @@ Status TpuCompileOpKernelCommon::ComputeInternal(OpKernelContext* ctx) {
       }
     }
   }
-  return Status::OK();
+  return status;
 }
 }  // namespace tpu
 }  // namespace tensorflow
diff --git a/tensorflow/core/tpu/tpu_compile_interface.h b/tensorflow/core/tpu/tpu_compile_interface.h
index 7e7b1f8315a..8f73b2dcacd 100644
--- a/tensorflow/core/tpu/tpu_compile_interface.h
+++ b/tensorflow/core/tpu/tpu_compile_interface.h
@@ -28,6 +28,9 @@ class TpuCompileInterface {
   static bool RegisterImplementation(TpuCompileInterface* impl);
 
   virtual uint64_t FingerprintString(absl::string_view str) = 0;
+
+  static inline constexpr char kTpuCompileErrorPayloadKey[] =
+      "https://www.tensorflow.org/tpu-compile-error";
 };
 
 #endif  // TENSORFLOW_CORE_TPU_TPU_COMPILE_INTERFACE_H_
diff --git a/tensorflow/python/distribute/tpu_strategy_test.py b/tensorflow/python/distribute/tpu_strategy_test.py
index 9f5fdb04d1e..dcb9cce1f66 100644
--- a/tensorflow/python/distribute/tpu_strategy_test.py
+++ b/tensorflow/python/distribute/tpu_strategy_test.py
@@ -300,7 +300,7 @@ class TPUStrategyTest(test.TestCase, parameterized.TestCase):
       return strategy.experimental_local_results(
           strategy.run(step_fn, args=(next(iterator),)))
 
-    with self.assertRaisesRegex(errors.InternalError, "Compilation failure"):
+    with self.assertRaises(errors.InternalError):
       logging.info(train_fn(iterator))
 
   def test_computation_on_subset_cores(self, enable_packed_var):