diff --git a/.github/workflows/deployments.yml b/.github/workflows/deployments.yml
index 5505abb6f4..d3e636298c 100644
--- a/.github/workflows/deployments.yml
+++ b/.github/workflows/deployments.yml
@@ -264,6 +264,7 @@ jobs:
     strategy:
       matrix:
         platform: ${{ fromJson(needs.metadata.outputs.multi_platform || needs.metadata.outputs.platforms).ids }}
+        cuda_version: ["11.8", "12.0"]
       fail-fast: false
     uses: ./.github/workflows/dev_environment.yml
     secrets:
@@ -272,11 +273,14 @@ jobs:
     with:
       platforms: ${{ fromJson(needs.metadata.outputs.multi_platform || needs.metadata.outputs.platforms)[format('{0}', matrix.platform)].docker_flag }}
       dockerfile: build/devdeps.ompi.Dockerfile
+      build_config_id: cu${{ matrix.cuda_version }}
+      build_args: |
+        cuda_version=${{ matrix.cuda_version }}
       registry_cache_from: ${{ needs.metadata.outputs.cache_base }}
       update_registry_cache: ${{ needs.metadata.outputs.cache_target }}
       environment: ${{ needs.metadata.outputs.environment }}
       # needed only for the cloudposse GitHub action
-      matrix_key: ${{ matrix.platform }}-ompi
+      matrix_key: ${{ matrix.platform }}-cu${{ matrix.cuda_version }}-ompi
 
   dispatch:
     name: Dispatch deployments
@@ -327,6 +331,7 @@ jobs:
     strategy:
       matrix:
         platform: ${{ fromJson(needs.metadata.outputs.multi_platform || needs.metadata.outputs.platforms).ids }}
+        cuda_version: ["11.8", "12.0"]
       fail-fast: false
     uses: ./.github/workflows/dev_environment.yml
     secrets:
@@ -335,15 +340,17 @@ jobs:
     with:
       platforms: ${{ fromJson(needs.metadata.outputs.multi_platform || needs.metadata.outputs.platforms)[format('{0}', matrix.platform)].docker_flag }}
       dockerfile: build/devdeps.ext.Dockerfile
-      build_config_id: ${{ needs.config.outputs.devdeps_toolchain }}
+      build_config_id: cu${{ matrix.cuda_version }}-${{ needs.config.outputs.devdeps_toolchain }}
       build_args: |
+        cuda_version=${{ matrix.cuda_version }}
         base_image=${{ fromJson(needs.config.outputs.json).image_hash[format('{0}-{1}', matrix.platform, needs.config.outputs.devdeps_toolchain)] }}
-        ompidev_image=${{ fromJson(needs.config.outputs.json).image_hash[format('{0}-ompi', matrix.platform)] }}
+        ompidev_image=${{ fromJson(needs.config.outputs.json).image_hash[format('{0}-cu{1}-ompi', matrix.platform, matrix.cuda_version)] }}
+        ${{ matrix.cuda_version != '11.8' && 'cuda_packages=cuda-cudart cuda-compiler libcublas-dev libcusolver libnvjitlink' || '' }}
       registry_cache_from: ${{ needs.metadata.outputs.cache_base }}
       update_registry_cache: ${{ needs.metadata.outputs.cache_target }}
       environment: ${{ needs.metadata.outputs.environment }}
       # needed only for the cloudposse GitHub action
-      matrix_key: ${{ matrix.platform }}-ext
+      matrix_key: ${{ matrix.platform }}-cu${{ matrix.cuda_version }}-ext
 
   # This job is needed only when using the cloudposse GitHub action to read
   # the output of a matrix job. This is a workaround due to current GitHub
@@ -369,6 +376,7 @@ jobs:
     strategy:
       matrix:
         platform: ${{ fromJson(needs.metadata.outputs.multi_platform || needs.metadata.outputs.platforms).ids }}
+        cuda_version: ["11.8", "12.0"]
       fail-fast: false
     uses: ./.github/workflows/docker_images.yml
     secrets:
@@ -377,9 +385,10 @@ jobs:
       DOCKERHUB_READONLY_TOKEN: ${{ secrets.DOCKERHUB_READONLY_TOKEN }}
     with:
       platforms: ${{ fromJson(needs.metadata.outputs.multi_platform || needs.metadata.outputs.platforms)[format('{0}', matrix.platform)].docker_flag }}
-      ompidev_image: ${{ fromJson(needs.config.outputs.json).image_hash[format('{0}-ompi', matrix.platform)] }}
-      devdeps_image: ${{ fromJson(needs.ext_config.outputs.json).image_hash[format('{0}-ext', matrix.platform)] }}
-      build_docs: true
+      cuda_version: ${{ matrix.cuda_version }}
+      ompidev_image: ${{ fromJson(needs.config.outputs.json).image_hash[format('{0}-cu{1}-ompi', matrix.platform, matrix.cuda_version)] }}
+      devdeps_image: ${{ fromJson(needs.ext_config.outputs.json).image_hash[format('{0}-cu{1}-ext', matrix.platform, matrix.cuda_version)] }}
+      build_docs: ${{ matrix.cuda_version == '11.8' }}
       environment: ghcr-deployment
 
   python_wheels:
diff --git a/.github/workflows/docker_images.yml b/.github/workflows/docker_images.yml
index 278178b0e5..5ed80a92cc 100644
--- a/.github/workflows/docker_images.yml
+++ b/.github/workflows/docker_images.yml
@@ -20,6 +20,10 @@ on:
       build_docs:
         required: false
         type: string
+      cuda_version:
+        required: false
+        type: string
+        description: The CUDA version with which the development dependencies were built.
       environment:
         required: false
         type: string
@@ -165,8 +169,8 @@ jobs:
           file: ./docker/build/devdeps.ext.Dockerfile
           build-args: |
             ompidev_image=${{ inputs.ompidev_image }}
-            base_image=nvidia/cuda:11.8.0-runtime-ubuntu22.04
-            cuda_root=/usr/local/cuda-11.8
+            base_image=nvidia/cuda:${{ inputs.cuda_version }}.0-runtime-ubuntu22.04
+            cuda_version=${{ inputs.cuda_version }}
             cuda_packages=
           tags: ${{ steps.metadata.outputs.tags }}
           labels: ${{ steps.metadata.outputs.labels }}
@@ -470,7 +474,8 @@ jobs:
           image_description="CUDA Quantum toolkit for heterogeneous quantum-classical workflows"
 
           platform_tag=${{ needs.metadata.outputs.platform_tag }}
-          image_tag=${platform_tag:+$platform_tag-}
+          cuda_major=`echo ${{ inputs.cuda_version }} | cut -d . -f1`
+          image_tag=${platform_tag:+$platform_tag-}${cuda_major:+cu${cuda_major}-}
           if ${{ github.event.pull_request.number != '' }} || [ -n "$(echo ${{ github.ref_name }} | grep pull-request/)" ]; then
             pr_number=`echo ${{ github.ref_name }} | grep -o [0-9]*`
             image_tag+=pr-${pr_number:-${{ github.event.pull_request.number }}}
@@ -841,7 +846,8 @@ jobs:
         run: |
           cudaqdev_hash=${{ needs.cudaqdev_image.outputs.image_hash }}
           cudaq_hash=${{ needs.cudaq_image.outputs.image_hash }}
-          artifact_name=image_publishing # changing the artifact name requires updating other workflows
+          cuda_major=`echo ${{ inputs.cuda_version }} | cut -d . -f1`
+          artifact_name=image_${cuda_major:+cu${cuda_major}_}publishing # changing the artifact name requires updating other workflows
           echo "artifact_name=$artifact_name" >> $GITHUB_OUTPUT
           info_file="$artifact_name.txt"
           echo "info_file=$info_file" >> $GITHUB_OUTPUT
@@ -852,6 +858,7 @@ jobs:
           echo "cuda-quantum-dev-image: $cudaqdev_hash"  >> "$info_file"
           echo "cuda-quantum-devdeps-image: ${{ inputs.devdeps_image }}" >> "$info_file"
           echo "platforms: ${{ inputs.platforms }}" >> "$info_file"
+          echo "cuda-version: ${{ inputs.cuda_version }}" >> "$info_file"
           cat .github/workflows/config/gitlab_commits.txt >> "$info_file"
 
       - name: Upload build info
diff --git a/.github/workflows/prebuilt_binaries.yml b/.github/workflows/prebuilt_binaries.yml
index c1414757e1..e7bb617b79 100644
--- a/.github/workflows/prebuilt_binaries.yml
+++ b/.github/workflows/prebuilt_binaries.yml
@@ -487,7 +487,8 @@ jobs:
   staging:
     name: Staging
     needs: [build_installer, build_openmpi, create_test_config, validation]
-    if: inputs.environment
+    # Allow staging even if there are not validations configured.
+    if: always() && !cancelled() && inputs.environment && !contains(needs.*.result, 'failure') && !contains(needs.*.result, 'cancelled')
     runs-on: ubuntu-latest
     permissions:
       contents: read
diff --git a/.github/workflows/python_wheels.yml b/.github/workflows/python_wheels.yml
index 0c7dce1d64..f9c92baeab 100644
--- a/.github/workflows/python_wheels.yml
+++ b/.github/workflows/python_wheels.yml
@@ -380,8 +380,9 @@ jobs:
         run: |
           wheeldeps_hash=${{ inputs.devdeps_image }}
           platform_id=`echo "${{ inputs.platform }}" | sed 's/linux\///g' | tr -d ' ' | tr ',' -`
+          cuda_major=`echo ${{ inputs.cuda_version }} | cut -d . -f1`
           build_config_id=${{ inputs.build_config_id }}
-          build_config_id=${build_config_id:-cu${{ inputs.cuda_version }}}
+          build_config_id=${build_config_id:-cu$cuda_major}
           artifact_name=python_${platform_id}_${build_config_id}_publishing # changing the artifact name requires updating other workflows
 
           echo "platform_id=$platform_id" >> $GITHUB_OUTPUT
@@ -393,6 +394,7 @@ jobs:
           echo "source-sha: ${{ github.sha }}" >> "$info_file"
           echo "cuda-quantum-wheeldeps-image: $wheeldeps_hash" >> "$info_file"
           echo "platform: ${{ inputs.platform }}" >> "$info_file"
+          echo "cuda-version: ${{ inputs.cuda_version }}" >> "$info_file"
           cat .github/workflows/config/gitlab_commits.txt >> "$info_file"
 
       - name: Upload build info
diff --git a/docker/build/assets.Dockerfile b/docker/build/assets.Dockerfile
index 4ec56f81f1..8f14d77342 100644
--- a/docker/build/assets.Dockerfile
+++ b/docker/build/assets.Dockerfile
@@ -18,7 +18,7 @@
 ARG base_image=amd64/almalinux:8
 FROM ${base_image} AS prereqs
 SHELL ["/bin/bash", "-c"]
-ARG cuda_version
+ARG cuda_version=11.8
 ENV CUDA_VERSION=${cuda_version}
 
 # When a dialogue box would be needed during install, assume default configurations.
diff --git a/docker/build/cudaq.dev.Dockerfile b/docker/build/cudaq.dev.Dockerfile
index a3b9f32dbc..c5c25340a2 100644
--- a/docker/build/cudaq.dev.Dockerfile
+++ b/docker/build/cudaq.dev.Dockerfile
@@ -17,7 +17,7 @@
 # 3) set the CC and CXX environment variable to use the same compiler toolchain
 #    as the LLVM dependencies have been built with.
 
-ARG base_image=ghcr.io/nvidia/cuda-quantum-devdeps:ext-gcc11-main
+ARG base_image=ghcr.io/nvidia/cuda-quantum-devdeps:ext-cu11-gcc11-main
 FROM $base_image
 
 ENV CUDAQ_REPO_ROOT=/workspaces/cuda-quantum
diff --git a/docker/build/devdeps.ext.Dockerfile b/docker/build/devdeps.ext.Dockerfile
index 94ee9d3420..f785391e8c 100644
--- a/docker/build/devdeps.ext.Dockerfile
+++ b/docker/build/devdeps.ext.Dockerfile
@@ -16,12 +16,24 @@
 # Must be built from the repo root with:
 #   docker build -t ghcr.io/nvidia/cuda-quantum-devdeps:ext -f docker/build/devdeps.ext.Dockerfile .
 
+ARG cuda_version=11.8
 ARG base_image=ghcr.io/nvidia/cuda-quantum-devdeps:llvm-main
-ARG ompidev_image=ghcr.io/nvidia/cuda-quantum-devdeps:ompi-main
+ARG ompidev_image=ghcr.io/nvidia/cuda-quantum-devdeps:cu11-ompi-main
 FROM $ompidev_image AS ompibuild
+ARG cuda_version
+RUN if [ -z "${cuda_version}" ]; then \
+        echo -e "\e[01;31mError: Missing argument cuda_version.\e[0m" >&2 && \
+        exit 1; \
+    fi && \        
+    if [ -n "${CUDA_VERSION}" ] && [ "${CUDA_VERSION}" != "${cuda_version}" ]; then \
+        echo -e "\e[01;31mError: CUDA version ${CUDA_VERSION} in ompidev_image does not match ${cuda_version}.\e[0m" >&2 && \
+        exit 1; \
+    fi
 
 FROM $base_image
 SHELL ["/bin/bash", "-c"]
+ARG cuda_version
+ENV CUDA_VERSION=${cuda_version}
 
 # When a dialogue box would be needed during install, assume default configurations.
 # Set here to avoid setting it for all install commands. 
@@ -125,12 +137,13 @@ ENV CUQUANTUM_PATH="$CUQUANTUM_INSTALL_PREFIX"
 ENV LD_LIBRARY_PATH="$CUQUANTUM_INSTALL_PREFIX/lib:$LD_LIBRARY_PATH"
 ENV CPATH="$CUQUANTUM_INSTALL_PREFIX/include:$CPATH"
 
-ENV CUQUANTUM_VERSION=24.03.0.4_cuda11
+ENV CUQUANTUM_VERSION=24.08.0.5
 RUN apt-get update && apt-get install -y --no-install-recommends xz-utils \
     && arch_folder=$([ "$(uname -m)" == "aarch64" ] && echo sbsa || echo x86_64) \
-    && wget -q "https://developer.download.nvidia.com/compute/cuquantum/redist/cuquantum/linux-$arch_folder/cuquantum-linux-$arch_folder-$CUQUANTUM_VERSION-archive.tar.xz" \
-    && mkdir -p "$CUQUANTUM_INSTALL_PREFIX" && tar xf cuquantum-linux-$arch_folder-$CUQUANTUM_VERSION-archive.tar.xz --strip-components 1 -C "$CUQUANTUM_INSTALL_PREFIX" \
-    && rm cuquantum-linux-$arch_folder-$CUQUANTUM_VERSION-archive.tar.xz \
+    && cuda_major=$(echo $CUDA_VERSION | cut -d . -f1) \
+    && wget -q "https://developer.download.nvidia.com/compute/cuquantum/redist/cuquantum/linux-$arch_folder/cuquantum-linux-$arch_folder-${CUQUANTUM_VERSION}_cuda${cuda_major}-archive.tar.xz" \
+    && mkdir -p "$CUQUANTUM_INSTALL_PREFIX" && tar xf cuquantum-linux-$arch_folder-${CUQUANTUM_VERSION}_cuda${cuda_major}-archive.tar.xz --strip-components 1 -C "$CUQUANTUM_INSTALL_PREFIX" \
+    && rm cuquantum-linux-$arch_folder-${CUQUANTUM_VERSION}_cuda${cuda_major}-archive.tar.xz \
     && apt-get remove -y xz-utils \
     && apt-get autoremove -y --purge && apt-get clean && rm -rf /var/lib/apt/lists/* 
 
@@ -145,19 +158,20 @@ ENV CPATH="$CUTENSOR_INSTALL_PREFIX/include:$CPATH"
 ENV CUTENSOR_VERSION=2.0.1.2
 RUN apt-get update && apt-get install -y --no-install-recommends xz-utils \
     && arch_folder=$([ "$(uname -m)" == "aarch64" ] && echo sbsa || echo x86_64) \
+    && cuda_major=$(echo $CUDA_VERSION | cut -d . -f1) \
     && wget -q "https://developer.download.nvidia.com/compute/cutensor/redist/libcutensor/linux-$arch_folder/libcutensor-linux-$arch_folder-$CUTENSOR_VERSION-archive.tar.xz" \
     && tar xf libcutensor-linux-$arch_folder-$CUTENSOR_VERSION-archive.tar.xz && cd libcutensor-linux-$arch_folder-$CUTENSOR_VERSION-archive \
-    && mkdir -p "$CUTENSOR_INSTALL_PREFIX" && mv include "$CUTENSOR_INSTALL_PREFIX" && mv lib/11 "$CUTENSOR_INSTALL_PREFIX/lib" \
+    && mkdir -p "$CUTENSOR_INSTALL_PREFIX" && mv include "$CUTENSOR_INSTALL_PREFIX" && mv lib/${cuda_major} "$CUTENSOR_INSTALL_PREFIX/lib" \
     && cd / && rm -rf libcutensor-linux-$arch_folder-$CUTENSOR_VERSION-archive* \
     && apt-get remove -y xz-utils \
     && apt-get autoremove -y --purge && apt-get clean && rm -rf /var/lib/apt/lists/* 
 
-# Install CUDA 11.8.
+# Install CUDA
 
-ARG cuda_root=/usr/local/cuda-11.8
-ARG cuda_packages="cuda-cudart-11-8 cuda-compiler-11-8 libcublas-dev-11-8 libcusolver-11-8"
+ARG cuda_packages="cuda-cudart cuda-compiler libcublas-dev libcusolver"
 RUN if [ -n "$cuda_packages" ]; then \
         arch_folder=$([ "$(uname -m)" == "aarch64" ] && echo sbsa || echo x86_64) \
+        && cuda_packages=`printf '%s\n' $cuda_packages | xargs -I {} echo {}-$(echo ${CUDA_VERSION} | tr . -)` \
         && wget -q "https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/$arch_folder/cuda-keyring_1.0-1_all.deb" \
         && dpkg -i cuda-keyring_1.0-1_all.deb \
         && apt-get update && apt-get install -y --no-install-recommends $cuda_packages \
@@ -178,7 +192,7 @@ RUN if [ -z "$CUDA_ROOT" ]; then \
         /usr/local/cuda/compat/lib; \
     fi
 
-ENV CUDA_INSTALL_PREFIX="$cuda_root"
+ENV CUDA_INSTALL_PREFIX="/usr/local/cuda-${CUDA_VERSION}"
 ENV CUDA_HOME="$CUDA_INSTALL_PREFIX"
 ENV CUDA_ROOT="$CUDA_INSTALL_PREFIX"
 ENV CUDA_PATH="$CUDA_INSTALL_PREFIX"
diff --git a/docker/build/devdeps.manylinux.Dockerfile b/docker/build/devdeps.manylinux.Dockerfile
index 9a040d37e7..e12a438fcd 100644
--- a/docker/build/devdeps.manylinux.Dockerfile
+++ b/docker/build/devdeps.manylinux.Dockerfile
@@ -101,7 +101,7 @@ RUN bash /scripts/install_prerequisites.sh
 
 # Install CUDA
 
-ARG cuda_version
+ARG cuda_version=11.8
 ENV CUDA_VERSION=${cuda_version}
 
 # Note that pip packages are available for all necessary runtime components.
diff --git a/docker/build/devdeps.ompi.Dockerfile b/docker/build/devdeps.ompi.Dockerfile
index 529b379323..1aba3d2bd2 100644
--- a/docker/build/devdeps.ompi.Dockerfile
+++ b/docker/build/devdeps.ompi.Dockerfile
@@ -15,12 +15,15 @@
 # Must be built from the repo root with:
 #   docker build -t ghcr.io/nvidia/cuda-quantum-devdeps:ompi -f docker/build/devdeps.ompi.Dockerfile .
 
-FROM nvidia/cuda:11.8.0-devel-ubuntu22.04
+ARG cuda_version=11.8
+FROM nvidia/cuda:${cuda_version}.0-devel-ubuntu22.04
 SHELL ["/bin/bash", "-c"]
 ARG DEBIAN_FRONTEND=noninteractive
+ARG cuda_version
+ENV CUDA_VERSION=${cuda_version}
 
 ARG TARGETARCH
-ENV CUDA_INSTALL_PREFIX=/usr/local/cuda-11.8
+ENV CUDA_INSTALL_PREFIX=/usr/local/cuda-${CUDA_VERSION}
 ENV COMMON_COMPILER_FLAGS="-march=x86-64-v3 -mtune=generic -O2 -pipe"
 ENV COMMON_COMPILER_FLAGS_ARM="-march=armv8-a -mtune=generic -O2 -pipe"
 
diff --git a/pyproject.toml b/pyproject.toml
index 47af11178d..194750ad58 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -18,7 +18,7 @@ requires-python = ">=3.10"
 license = { file="LICENSE" }
 dependencies = [
   'astpretty ~= 3.0',
-  'cuquantum-cu11 ~= 24.03',
+  'cuquantum-cu11 ~= 24.08',
   'graphlib-backport >= 1.0',
   'numpy >= 1.24',
   'requests >= 2.31',
diff --git a/scripts/configure_build.sh b/scripts/configure_build.sh
index deafb645df..1ee4dbadb8 100644
--- a/scripts/configure_build.sh
+++ b/scripts/configure_build.sh
@@ -63,7 +63,7 @@ if [ "$1" == "install-cuquantum" ]; then
     CUDA_ARCH_FOLDER=$([ "$(uname -m)" == "aarch64" ] && echo sbsa || echo x86_64)
 
 # [>cuQuantumInstall]
-    CUQUANTUM_VERSION=24.03.0.4
+    CUQUANTUM_VERSION=24.08.0.5
     CUQUANTUM_DOWNLOAD_URL=https://developer.download.nvidia.com/compute/cuquantum/redist/cuquantum
 
     cuquantum_archive=cuquantum-linux-${CUDA_ARCH_FOLDER}-${CUQUANTUM_VERSION}_cuda$(echo ${CUDA_VERSION} | cut -d . -f1)-archive.tar.xz