diff --git a/xla/backends/gpu/autotuner/triton.cc b/xla/backends/gpu/autotuner/triton.cc
index b11bb7889d721..0dfd531de48f9 100644
--- a/xla/backends/gpu/autotuner/triton.cc
+++ b/xla/backends/gpu/autotuner/triton.cc
@@ -65,6 +65,12 @@ namespace {
 std::vector<TritonGemmConfig> GetDefaultTritonConfigs(
     se::GpuComputeCapability compute_capability) {
   if (compute_capability.IsRocm()) {
+    const auto* rocm_cc = compute_capability.rocm_compute_capability();
+    if (rocm_cc->gfx9_mi300()) {
+      return GetTritonConfigsForPlatform(TritonConfigsPlatform::kMI300);
+    } else if (rocm_cc->gfx9_mi350()) {
+      return GetTritonConfigsForPlatform(TritonConfigsPlatform::kMI350);
+    }
     return GetTritonConfigsForPlatform(TritonConfigsPlatform::kDefaultRocm);
   }
 
diff --git a/xla/backends/gpu/autotuner/triton/default_configs/mi300.txtpb b/xla/backends/gpu/autotuner/triton/default_configs/mi300.txtpb
new file mode 100644
index 0000000000000..aca70ca28511e
--- /dev/null
+++ b/xla/backends/gpu/autotuner/triton/default_configs/mi300.txtpb
@@ -0,0 +1,47 @@
+# Copyright 2026 The OpenXLA Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+config { block_m: 32 block_n: 32 block_k: 256 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 64 block_n: 32 block_k: 32 split_k: 16 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 32 block_n: 64 block_k: 64 split_k: 4 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 128 block_n: 128 block_k: 64 split_k: 4 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 16 block_n: 16 block_k: 256 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 16 block_n: 128 block_k: 32 split_k: 16 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 256 block_n: 256 block_k: 32 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 256 block_k: 64 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 256 block_k: 32 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 256 block_n: 128 block_k: 64 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 128 block_k: 64 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 32 block_n: 8 block_k: 16 split_k: 1 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 64 block_n: 32 block_k: 16 split_k: 1 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 128 block_n: 32 block_k: 16 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 128 block_n: 64 block_k: 128 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 128 block_k: 32 split_k: 1 num_stages: 3 num_warps: 4 num_ctas: 1 }
+config { block_m: 128 block_n: 256 block_k: 64 split_k: 2 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 256 block_n: 128 block_k: 32 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 256 block_n: 256 block_k: 32 split_k: 1 num_stages: 1 num_warps: 8 num_ctas: 1 }
+config { block_m: 256 block_n: 256 block_k: 32 split_k: 4 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 32 block_k: 32 split_k: 8 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 64 block_n: 32 block_k: 32 split_k: 8 num_stages: 3 num_warps: 2 num_ctas: 1 }
+config { block_m: 128 block_n: 32 block_k: 32 split_k: 32 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 32 block_n: 32 block_k: 32 split_k: 8 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 64 block_n: 32 block_k: 128 split_k: 2 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 256 block_n: 8 block_k: 32 split_k: 4 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 128 block_n: 16 block_k: 128 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 32 block_n: 16 block_k: 128 split_k: 2 num_stages: 5 num_warps: 2 num_ctas: 1 }
+config { block_m: 32 block_n: 16 block_k: 128 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 64 block_n: 8 block_k: 128 split_k: 1 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 32 block_n: 16 block_k: 256 split_k: 1 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 256 block_n: 8 block_k: 16 split_k: 8 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 128 block_n: 8 block_k: 16 split_k: 32 num_stages: 1 num_warps: 2 num_ctas: 1 }
diff --git a/xla/backends/gpu/autotuner/triton/default_configs/mi350.txtpb b/xla/backends/gpu/autotuner/triton/default_configs/mi350.txtpb
new file mode 100644
index 0000000000000..849069a051ffe
--- /dev/null
+++ b/xla/backends/gpu/autotuner/triton/default_configs/mi350.txtpb
@@ -0,0 +1,54 @@
+# Copyright 2026 The OpenXLA Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+config { block_m: 32 block_n: 32 block_k: 256 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 64 block_n: 32 block_k: 32 split_k: 16 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 32 block_n: 64 block_k: 64 split_k: 4 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 128 block_n: 128 block_k: 64 split_k: 4 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 16 block_n: 16 block_k: 256 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 16 block_n: 128 block_k: 32 split_k: 16 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 256 block_n: 256 block_k: 32 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 256 block_k: 64 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 256 block_k: 32 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 256 block_n: 128 block_k: 64 split_k: 1 num_stages: 2 num_warps: 8 num_ctas: 1 }
+config { block_m: 128 block_n: 128 block_k: 64 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 32 block_n: 8 block_k: 16 split_k: 1 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 128 block_n: 32 block_k: 16 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 32 block_n: 8 block_k: 32 split_k: 1 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 128 block_n: 32 block_k: 16 split_k: 1 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 256 block_n: 256 block_k: 16 split_k: 2 num_stages: 4 num_warps: 8 num_ctas: 1 }
+config { block_m: 256 block_n: 128 block_k: 64 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 256 block_n: 128 block_k: 64 split_k: 4 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 256 block_n: 128 block_k: 16 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 16 block_n: 16 block_k: 128 split_k: 1 num_stages: 3 num_warps: 2 num_ctas: 1 }
+config { block_m: 32 block_n: 16 block_k: 128 split_k: 1 num_stages: 2 num_warps: 4 num_ctas: 1 }
+config { block_m: 64 block_n: 8 block_k: 128 split_k: 1 num_stages: 4 num_warps: 2 num_ctas: 1 }
+config { block_m: 64 block_n: 16 block_k: 64 split_k: 16 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 128 block_n: 32 block_k: 16 split_k: 16 num_stages: 3 num_warps: 4 num_ctas: 1 }
+config { block_m: 64 block_n: 32 block_k: 16 split_k: 16 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 32 block_n: 16 block_k: 32 split_k: 8 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 64 block_n: 8 block_k: 16 split_k: 8 num_stages: 3 num_warps: 2 num_ctas: 1 }
+config { block_m: 16 block_n: 8 block_k: 256 split_k: 1 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 16 block_n: 16 block_k: 128 split_k: 4 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 16 block_n: 16 block_k: 128 split_k: 8 num_stages: 1 num_warps: 2 num_ctas: 1 }
+config { block_m: 16 block_n: 16 block_k: 128 split_k: 8 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 16 block_n: 64 block_k: 128 split_k: 1 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 32 block_n: 16 block_k: 64 split_k: 8 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 64 block_n: 8 block_k: 16 split_k: 4 num_stages: 3 num_warps: 2 num_ctas: 1 }
+config { block_m: 64 block_n: 8 block_k: 64 split_k: 16 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 64 block_n: 8 block_k: 256 split_k: 1 num_stages: 2 num_warps: 2 num_ctas: 1 }
+config { block_m: 64 block_n: 16 block_k: 256 split_k: 16 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 128 block_n: 8 block_k: 32 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 128 block_n: 32 block_k: 64 split_k: 16 num_stages: 1 num_warps: 4 num_ctas: 1 }
+config { block_m: 128 block_n: 64 block_k: 16 split_k: 1 num_stages: 1 num_warps: 4 num_ctas: 1 }
diff --git a/xla/backends/gpu/autotuner/triton/triton_configs.cc b/xla/backends/gpu/autotuner/triton/triton_configs.cc
index 7d56598826e79..e4ab7a74f7b22 100644
--- a/xla/backends/gpu/autotuner/triton/triton_configs.cc
+++ b/xla/backends/gpu/autotuner/triton/triton_configs.cc
@@ -58,7 +58,10 @@ const std::vector<TritonGemmConfig>& GetTritonConfigsForPlatform(
             ParseConfig(configs::get_cuda())},
            {TritonConfigsPlatform::kDefaultRocm,
             ParseConfig(configs::get_rocm())},
-           {TritonConfigsPlatform::kHopper, ParseConfig(configs::get_h100())}});
+           {TritonConfigsPlatform::kHopper, ParseConfig(configs::get_h100())},
+           {TritonConfigsPlatform::kMI300, ParseConfig(configs::get_mi300())},
+           {TritonConfigsPlatform::kMI350,
+            ParseConfig(configs::get_mi350())}});
   return kConfigs->at(platform);
 }
 
diff --git a/xla/backends/gpu/autotuner/triton/triton_configs.h b/xla/backends/gpu/autotuner/triton/triton_configs.h
index c6b30359e9965..4d4c9e268adbd 100644
--- a/xla/backends/gpu/autotuner/triton/triton_configs.h
+++ b/xla/backends/gpu/autotuner/triton/triton_configs.h
@@ -28,6 +28,8 @@ enum class TritonConfigsPlatform {
   kDefaultCuda,
   kDefaultRocm,
   kHopper,
+  kMI300,
+  kMI350,
 };
 
 const std::vector<TritonGemmConfig>& GetTritonConfigsForPlatform(