diff --git a/source/cloud/azure/azureml.md b/source/cloud/azure/azureml.md
index 7915544e..62b56eda 100644
--- a/source/cloud/azure/azureml.md
+++ b/source/cloud/azure/azureml.md
@@ -45,13 +45,6 @@ from the script, the VM will boot and inspect it by running each line of the scr
 to see where it fails.
 ```
 
-```{warning}
-AzureML is pinned to **CUDA 12** as its GPU instances ship NVIDIA driver 535.x,
-while CUDA 13 images require driver 580.65.06 or newer. Driver 535.x has itself
-reached end-of-life and no longer receives updates, so we cannot guarantee that
-the instructions below will work on AzureML.
-```
-
 ```bash
 #!/bin/bash
 set -e
@@ -62,7 +55,7 @@ conda create -y -n rapids \
     --override-channels \
     {{ rapids_conda_channels }} \
     -c microsoft \
-   rapids={{ rapids_version }} python=3.13 'cuda-version>=12.0,<=12.9' \
+    {{ rapids_conda_packages }} \
     'azure-identity>=1.19' \
     ipykernel
 
@@ -181,7 +174,7 @@ touch ./training-code/__init__.py
 
 # create a Dockerfile defining the image the code will run in
 cat > ./training-code/Dockerfile <<EOF
-FROM {{ rapids_container_cuda12 }}
+FROM {{ rapids_container }}
 
 RUN conda install --yes -c conda-forge 'dask-ml>=2024.4.4' \
  && pip install azureml-mlflow
diff --git a/source/conf.py b/source/conf.py
index 3a159f9d..734e8678 100644
--- a/source/conf.py
+++ b/source/conf.py
@@ -47,8 +47,6 @@
         "rapids_pip_version": stable_version,
         "rapids_cuda_major": cuda_major,
         "rapids_cuda_version_range": stable_cuda_range,
-        # AzureML is pinned to CUDA 12 currently (driver 535.x supports only CUDA 12.x).
-        "rapids_container_cuda12": f"rapidsai/base:{stable_version}-cuda12-py{python_version}",
     },
     "nightly": {
         "rapids_version": f"{nightly_version}",
@@ -62,8 +60,6 @@
         "rapids_pip_version": f"{nightly_version}.*,>=0.0.0a0",
         "rapids_cuda_major": cuda_major,
         "rapids_cuda_version_range": nightly_cuda_range,
-        # AzureML is pinned to CUDA 12 currently (driver 535.x supports only CUDA 12.x).
-        "rapids_container_cuda12": f"rapidsai/base:{nightly_version + 'a'}-cuda12-py{python_version}",
     },
 }
 rapids_version = (