diff --git a/.dev-tools/.gitignore b/.dev-tools/.gitignore
index b93b0f69e..12262726f 100644
--- a/.dev-tools/.gitignore
+++ b/.dev-tools/.gitignore
@@ -8,6 +8,7 @@ platforms/gke/base/core/workloads/inference_gateway/manifests/*
 platforms/gke/base/core/workloads/jobset/manifests/*
 platforms/gke/base/core/workloads/kueue/manifests/*
 platforms/gke/base/core/workloads/lws/manifests/*
+platforms/gke/base/core/workloads/nri_device_injector/manifests/*
 platforms/gke/base/core/workloads/nvidia_nim/*
 platforms/gke/base/core/workloads/priority_class/manifests/*
 platforms/gke/base/kubernetes/*
diff --git a/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-2.yaml b/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-2.yaml
new file mode 100644
index 000000000..a5d3c9f0e
--- /dev/null
+++ b/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-2.yaml
@@ -0,0 +1,96 @@
+# Copyright 2026 Google LLC
+#
+# 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.
+
+apiVersion: cloud.google.com/v1
+kind: ComputeClass
+metadata:
+  name: gpu-rtx-pro-6000-96gb-x1-2
+spec:
+  activeMigration:
+    optimizeRulePriority: true
+  nodePoolConfig:
+    imageStreaming:
+      enabled: true
+  nodePoolAutoCreation:
+    enabled: true
+  priorities:
+    # Use a specific reservation
+    # - gpu:
+    #     count: 1
+    #     driverVersion: latest
+    #     type: nvidia-rtx-pro-6000
+    #   machineType: g4-standard-24
+    #   maxPodsPerNode: 32
+    #   reservations:
+    #     affinity: Specific
+    #     specific:
+    #       - name: nvidia-rtx-pro-6000-specific
+    #         reservationBlock:
+    #           name: <RESERVATION_NAME>
+    #   spot: false
+
+    # Use any reservation
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-24
+      maxPodsPerNode: 32
+      reservations:
+        affinity: AnyBestEffort
+      spot: false
+
+    # Use on-demand
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-24
+      maxPodsPerNode: 32
+      spot: false
+
+    # Use DWS FlexStart with 7 day limit
+    - flexStart:
+        enabled: true
+        nodeRecycling:
+          leadTimeSeconds: 3600
+      gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-24
+      maxPodsPerNode: 32
+      maxRunDurationSeconds: 604800
+
+    # Use DWS FlexStart with 1 day limit
+    - flexStart:
+        enabled: true
+        nodeRecycling:
+          leadTimeSeconds: 3600
+      gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-24
+      maxPodsPerNode: 32
+      maxRunDurationSeconds: 86400
+
+    # Use spot
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-24
+      maxPodsPerNode: 32
+      spot: true
diff --git a/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-4.yaml b/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-4.yaml
new file mode 100644
index 000000000..2fd214151
--- /dev/null
+++ b/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-4.yaml
@@ -0,0 +1,96 @@
+# Copyright 2026 Google LLC
+#
+# 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.
+
+apiVersion: cloud.google.com/v1
+kind: ComputeClass
+metadata:
+  name: gpu-rtx-pro-6000-96gb-x1-4
+spec:
+  activeMigration:
+    optimizeRulePriority: true
+  nodePoolConfig:
+    imageStreaming:
+      enabled: true
+  nodePoolAutoCreation:
+    enabled: true
+  priorities:
+    # Use a specific reservation
+    # - gpu:
+    #     count: 1
+    #     driverVersion: latest
+    #     type: nvidia-rtx-pro-6000
+    #   machineType: g4-standard-12
+    #   maxPodsPerNode: 32
+    #   reservations:
+    #     affinity: Specific
+    #     specific:
+    #       - name: nvidia-rtx-pro-6000-specific
+    #         reservationBlock:
+    #           name: <RESERVATION_NAME>
+    #   spot: false
+
+    # Use any reservation
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-12
+      maxPodsPerNode: 32
+      reservations:
+        affinity: AnyBestEffort
+      spot: false
+
+    # Use on-demand
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-12
+      maxPodsPerNode: 32
+      spot: false
+
+    # Use DWS FlexStart with 7 day limit
+    - flexStart:
+        enabled: true
+        nodeRecycling:
+          leadTimeSeconds: 3600
+      gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-12
+      maxPodsPerNode: 32
+      maxRunDurationSeconds: 604800
+
+    # Use DWS FlexStart with 1 day limit
+    - flexStart:
+        enabled: true
+        nodeRecycling:
+          leadTimeSeconds: 3600
+      gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-12
+      maxPodsPerNode: 32
+      maxRunDurationSeconds: 86400
+
+    # Use spot
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-12
+      maxPodsPerNode: 32
+      spot: true
diff --git a/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-8.yaml b/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-8.yaml
new file mode 100644
index 000000000..df6c0baa1
--- /dev/null
+++ b/platforms/gke/base/core/custom_compute_class/templates/manifests/gpu/rtx-pro-6000-96gb/custom-compute-gpu-rtx-pro-6000-96gb-x1-8.yaml
@@ -0,0 +1,96 @@
+# Copyright 2026 Google LLC
+#
+# 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.
+
+apiVersion: cloud.google.com/v1
+kind: ComputeClass
+metadata:
+  name: gpu-rtx-pro-6000-96gb-x1-8
+spec:
+  activeMigration:
+    optimizeRulePriority: true
+  nodePoolConfig:
+    imageStreaming:
+      enabled: true
+  nodePoolAutoCreation:
+    enabled: true
+  priorities:
+    # Use a specific reservation
+    # - gpu:
+    #     count: 1
+    #     driverVersion: latest
+    #     type: nvidia-rtx-pro-6000
+    #   machineType: g4-standard-6
+    #   maxPodsPerNode: 32
+    #   reservations:
+    #     affinity: Specific
+    #     specific:
+    #       - name: nvidia-rtx-pro-6000-specific
+    #         reservationBlock:
+    #           name: <RESERVATION_NAME>
+    #   spot: false
+
+    # Use any reservation
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-6
+      maxPodsPerNode: 32
+      reservations:
+        affinity: AnyBestEffort
+      spot: false
+
+    # Use on-demand
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-6
+      maxPodsPerNode: 32
+      spot: false
+
+    # Use DWS FlexStart with 7 day limit
+    - flexStart:
+        enabled: true
+        nodeRecycling:
+          leadTimeSeconds: 3600
+      gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-6
+      maxPodsPerNode: 32
+      maxRunDurationSeconds: 604800
+
+    # Use DWS FlexStart with 1 day limit
+    - flexStart:
+        enabled: true
+        nodeRecycling:
+          leadTimeSeconds: 3600
+      gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-6
+      maxPodsPerNode: 32
+      maxRunDurationSeconds: 86400
+
+    # Use spot
+    - gpu:
+        count: 1
+        driverVersion: latest
+        type: nvidia-rtx-pro-6000
+      machineType: g4-standard-6
+      maxPodsPerNode: 32
+      spot: true
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/.terraform.lock.hcl b/platforms/gke/base/core/workloads/nri_device_injector/.terraform.lock.hcl
new file mode 100644
index 000000000..27d625960
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/.terraform.lock.hcl
@@ -0,0 +1,42 @@
+# This file is maintained automatically by "terraform init".
+# Manual edits may be lost in future updates.
+
+provider "registry.terraform.io/hashicorp/google" {
+  version     = "7.6.0"
+  constraints = "7.6.0"
+  hashes = [
+    "h1:JYsO3fV5OtaNuRTdjGZC1Z3Ku1ZIrRJGwXwsBjtWudk=",
+    "zh:0c70c768b0a34d7a61de70d0e85cf0057820556647bbce2384972a45d7092e4e",
+    "zh:0cb7aab89cd435c5c8dab9231ea176d64fdf1df1125db15a6b9ead978a93c0b2",
+    "zh:32f25c42214bb356bb67cef6057c9904f2878cd053a7760e5ee3737619f28638",
+    "zh:38b05b1171ab086c88b95d379120fb6c28c9e895ae924557c11c35e138319119",
+    "zh:39d8206d453a614fa0be3aeac8ea3921fb3ab7ed122205cbbcc2a41ca6176cb5",
+    "zh:58d9059aa6b4aab5ede4fc173dcdc7b4d042d0b1a1ab55407dd345931d7f4815",
+    "zh:a4bc001c8ac7700d0107155296250c3b8969511e1a488f3b318f3db62362eef2",
+    "zh:cc75e25db4bb672ebc200a89d6cff9ff0b9911e14e188d1b4429bb3511d2b35f",
+    "zh:d7f7639930735f17b2b4f73814204a9a050186ea7e1c2671a52e0fa7ddf7a001",
+    "zh:f569b65999264a9416862bca5cd2a6177d94ccb0424f3a4ef424428912b9cb3c",
+    "zh:ff1190ae618dae9243de59caf4149abb4a9b775cb6439f119cd32a30f1a21820",
+    "zh:ff15b7b86787f6fd186211e7c37a72f2cc70374b284aaf063e1f989717441161",
+  ]
+}
+
+provider "registry.terraform.io/hashicorp/local" {
+  version     = "2.5.3"
+  constraints = "2.5.3"
+  hashes = [
+    "h1:1Nkh16jQJMp0EuDmvP/96f5Unnir0z12WyDuoR6HjMo=",
+    "zh:284d4b5b572eacd456e605e94372f740f6de27b71b4e1fd49b63745d8ecd4927",
+    "zh:40d9dfc9c549e406b5aab73c023aa485633c1b6b730c933d7bcc2fa67fd1ae6e",
+    "zh:6243509bb208656eb9dc17d3c525c89acdd27f08def427a0dce22d5db90a4c8b",
+    "zh:78d5eefdd9e494defcb3c68d282b8f96630502cac21d1ea161f53cfe9bb483b3",
+    "zh:885d85869f927853b6fe330e235cd03c337ac3b933b0d9ae827ec32fa1fdcdbf",
+    "zh:bab66af51039bdfcccf85b25fe562cbba2f54f6b3812202f4873ade834ec201d",
+    "zh:c505ff1bf9442a889ac7dca3ac05a8ee6f852e0118dd9a61796a2f6ff4837f09",
+    "zh:d36c0b5770841ddb6eaf0499ba3de48e5d4fc99f4829b6ab66b0fab59b1aaf4f",
+    "zh:ddb6a407c7f3ec63efb4dad5f948b54f7f4434ee1a2607a49680d494b1776fe1",
+    "zh:e0dafdd4500bec23d3ff221e3a9b60621c5273e5df867bc59ef6b7e41f5c91f6",
+    "zh:ece8742fd2882a8fc9d6efd20e2590010d43db386b920b2a9c220cfecc18de47",
+    "zh:f4c6b3eb8f39105004cf720e202f04f57e3578441cfb76ca27611139bc116a82",
+  ]
+}
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/_cluster.auto.tfvars b/platforms/gke/base/core/workloads/nri_device_injector/_cluster.auto.tfvars
new file mode 120000
index 000000000..4d9954e5a
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/_cluster.auto.tfvars
@@ -0,0 +1 @@
+../../../_shared_config/cluster.auto.tfvars
\ No newline at end of file
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/_cluster_variables.tf b/platforms/gke/base/core/workloads/nri_device_injector/_cluster_variables.tf
new file mode 120000
index 000000000..3f2c29e19
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/_cluster_variables.tf
@@ -0,0 +1 @@
+../../../_shared_config/cluster_variables.tf
\ No newline at end of file
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/_platform.auto.tfvars b/platforms/gke/base/core/workloads/nri_device_injector/_platform.auto.tfvars
new file mode 120000
index 000000000..c3133e727
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/_platform.auto.tfvars
@@ -0,0 +1 @@
+../../../_shared_config/platform.auto.tfvars
\ No newline at end of file
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/_platform_variables.tf b/platforms/gke/base/core/workloads/nri_device_injector/_platform_variables.tf
new file mode 120000
index 000000000..c68738baa
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/_platform_variables.tf
@@ -0,0 +1 @@
+../../../_shared_config/platform_variables.tf
\ No newline at end of file
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/_workloads.auto.tfvars b/platforms/gke/base/core/workloads/nri_device_injector/_workloads.auto.tfvars
new file mode 120000
index 000000000..b65551f53
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/_workloads.auto.tfvars
@@ -0,0 +1 @@
+../../../_shared_config/workloads.auto.tfvars
\ No newline at end of file
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/_workloads_variables.tf b/platforms/gke/base/core/workloads/nri_device_injector/_workloads_variables.tf
new file mode 120000
index 000000000..fec5c48ce
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/_workloads_variables.tf
@@ -0,0 +1 @@
+../../../_shared_config/workloads_variables.tf
\ No newline at end of file
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/main.tf b/platforms/gke/base/core/workloads/nri_device_injector/main.tf
new file mode 100644
index 000000000..ea8b5384c
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/main.tf
@@ -0,0 +1,62 @@
+# Copyright 2025 Google LLC
+#
+# 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.
+
+locals {
+  kubeconfig_directory = "${path.module}/../../../kubernetes/kubeconfig"
+  kubeconfig_file      = "${local.kubeconfig_directory}/${local.kubeconfig_file_name}"
+
+  manifests_directory         = "${local.namespace_directory}/kube-system"
+  namespace_directory         = "${local.manifests_directory_root}/namespace"
+  version_manifests_directory = "${path.module}/manifests"
+}
+
+data "local_file" "kubeconfig" {
+  filename = local.kubeconfig_file
+}
+
+resource "terraform_data" "manifests" {
+  input = {
+    manifests_dir         = local.manifests_directory
+    version_manifests_dir = local.version_manifests_directory
+  }
+
+  provisioner "local-exec" {
+    command     = <<EOT
+mkdir -p "${self.input.version_manifests_dir}" && \
+mkdir -p "${self.input.manifests_dir}" && \
+curl --silent --show-error --location https://raw.githubusercontent.com/GoogleCloudPlatform/container-engine-accelerators/refs/heads/master/nri_device_injector/nri-device-injector.yaml --output "${self.input.version_manifests_dir}/manifests.yaml" && \
+cp -r "${self.input.version_manifests_dir}"/* "${self.input.manifests_dir}/"
+EOT
+    interpreter = ["bash", "-c"]
+    working_dir = path.module
+  }
+
+  triggers_replace = {
+    manifests_dir         = local.manifests_directory
+    version_manifests_dir = local.version_manifests_directory
+  }
+}
+
+module "kubectl_apply_manifests" {
+  depends_on = [
+    terraform_data.manifests,
+  ]
+
+  source = "../../../modules/kubectl_apply"
+
+  apply_server_side           = true
+  kubeconfig_file             = data.local_file.kubeconfig.filename
+  manifest                    = local.version_manifests_directory
+  manifest_includes_namespace = true
+}
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/project.tf b/platforms/gke/base/core/workloads/nri_device_injector/project.tf
new file mode 100644
index 000000000..a50809807
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/project.tf
@@ -0,0 +1,17 @@
+# Copyright 2025 Google LLC
+#
+# 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.
+
+data "google_project" "cluster" {
+  project_id = local.cluster_project_id
+}
diff --git a/platforms/gke/base/core/workloads/nri_device_injector/versions.tf b/platforms/gke/base/core/workloads/nri_device_injector/versions.tf
new file mode 100644
index 000000000..a61ab18ab
--- /dev/null
+++ b/platforms/gke/base/core/workloads/nri_device_injector/versions.tf
@@ -0,0 +1,32 @@
+# Copyright 2025 Google LLC
+#
+# 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.
+
+terraform {
+  required_version = ">= 1.5.7"
+
+  required_providers {
+    google = {
+      source  = "hashicorp/google"
+      version = "7.6.0"
+    }
+    local = {
+      source  = "hashicorp/local"
+      version = "2.5.3"
+    }
+  }
+
+  provider_meta "google" {
+    module_name = "cloud-solutions/acp_gke_base_core_workloads_nri_deploy-v1"
+  }
+}
diff --git a/platforms/gke/base/use-cases/inference-ref-arch/terraform/README.md b/platforms/gke/base/use-cases/inference-ref-arch/terraform/README.md
index f526f9588..a6427dc63 100644
--- a/platforms/gke/base/use-cases/inference-ref-arch/terraform/README.md
+++ b/platforms/gke/base/use-cases/inference-ref-arch/terraform/README.md
@@ -206,6 +206,10 @@ For more information about providing values for Terraform input variables, see
           <li>gpu-l4-24gb-x2</li>
           <li>gpu-l4-24gb-x4</li>
           <li>gpu-l4-24gb-x8</li>
+          <li>gpu-rtx-pro-6000-96gb-x1</li>
+          <li>gpu-rtx-pro-6000-96gb-x1-2</li>
+          <li>gpu-rtx-pro-6000-96gb-x1-4</li>
+          <li>gpu-rtx-pro-6000-96gb-x1-8</li>
         <ul>
       </details>
     - <details>
diff --git a/platforms/gke/base/use-cases/inference-ref-arch/terraform/deploy-standard.sh b/platforms/gke/base/use-cases/inference-ref-arch/terraform/deploy-standard.sh
index c83ae354b..d2f6a6a66 100755
--- a/platforms/gke/base/use-cases/inference-ref-arch/terraform/deploy-standard.sh
+++ b/platforms/gke/base/use-cases/inference-ref-arch/terraform/deploy-standard.sh
@@ -50,6 +50,7 @@ declare -a CORE_TERRASERVICES_APPLY=(
   "workloads/jobset"
   "workloads/lws"
   "workloads/priority_class"
+  "workloads/nri_device_injector"
   "workloads/kueue"
 )
 CORE_TERRASERVICES_APPLY="${CORE_TERRASERVICES_APPLY[*]}" "${ACP_PLATFORM_CORE_DIR}/deploy.sh"
diff --git a/platforms/gke/base/use-cases/inference-ref-arch/terraform/teardown-standard.sh b/platforms/gke/base/use-cases/inference-ref-arch/terraform/teardown-standard.sh
index 1d0e1ba9a..e59cf3f97 100755
--- a/platforms/gke/base/use-cases/inference-ref-arch/terraform/teardown-standard.sh
+++ b/platforms/gke/base/use-cases/inference-ref-arch/terraform/teardown-standard.sh
@@ -69,6 +69,7 @@ done
 if [ "${ACP_TEARDOWN_CORE_PLATFORM}" = "true" ]; then
   declare -a CORE_TERRASERVICES_DESTROY=(
     "workloads/kueue"
+    "workloads/nri_device_injector"
     "workloads/priority_class"
     "workloads/lws"
     "workloads/jobset"