diff --git a/ci-operator/config/openshift/hypershift/openshift-hypershift-main.yaml b/ci-operator/config/openshift/hypershift/openshift-hypershift-main.yaml
index b5eda614c200c..6e3f3a73c3617 100644
--- a/ci-operator/config/openshift/hypershift/openshift-hypershift-main.yaml
+++ b/ci-operator/config/openshift/hypershift/openshift-hypershift-main.yaml
@@ -240,11 +240,13 @@ tests:
     - ref: hypershift-analyze-e2e-failure
     - chain: hypershift-destroy-nested-management-cluster
     workflow: hypershift-azure-e2e-self-managed
-- always_run: true
+- always_run: false
   as: e2e-azure-v2-self-managed
   optional: true
   steps:
     cluster_profile: hypershift-azure
+    dependencies:
+      OCP_IMAGE_N1: release:n1minor
     env:
       ENABLE_HYPERSHIFT_CERT_ROTATION_SCALE: "true"
       HYPERSHIFT_AZURE_LOCATION: centralus
diff --git a/ci-operator/jobs/openshift/hypershift/openshift-hypershift-main-presubmits.yaml b/ci-operator/jobs/openshift/hypershift/openshift-hypershift-main-presubmits.yaml
index c321d66a8d6e2..efc82a36049b6 100644
--- a/ci-operator/jobs/openshift/hypershift/openshift-hypershift-main-presubmits.yaml
+++ b/ci-operator/jobs/openshift/hypershift/openshift-hypershift-main-presubmits.yaml
@@ -1576,7 +1576,7 @@ presubmits:
           secretName: result-aggregator
     trigger: (?m)^/test( | .* )e2e-azure-self-managed,?($|\s.*)
   - agent: kubernetes
-    always_run: true
+    always_run: false
     branches:
     - ^main$
     - ^main-
diff --git a/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-commands.sh b/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-commands.sh
index fe16263aaeede..974487d10fc0f 100644
--- a/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-commands.sh
+++ b/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-commands.sh
@@ -4,240 +4,6 @@ set -exuo pipefail
 
 # Use the nested management cluster kubeconfig
 export KUBECONFIG="${SHARED_DIR}/management_cluster_kubeconfig"
+export HYPERSHIFT_BINARY="${HYPERSHIFT_BINARY:-/hypershift/bin/hypershift}"
 
-# Generate unique cluster names from job ID
-PUBLIC_NAME="$(echo -n "${PROW_JOB_ID}-pub"|sha256sum|cut -c-20)"
-PRIVATE_NAME="$(echo -n "${PROW_JOB_ID}-prv"|sha256sum|cut -c-20)"
-OAUTH_LB_NAME="$(echo -n "${PROW_JOB_ID}-oau"|sha256sum|cut -c-20)"
-
-# Self-managed Azure credentials
-AZURE_CREDS="/etc/hypershift-ci-jobs-self-managed-azure/credentials.json"
-AZURE_OIDC_ISSUER_URL="https://smazure.blob.core.windows.net/smazure"
-AZURE_SA_TOKEN_ISSUER_KEY_PATH="/etc/hypershift-ci-jobs-self-managed-azure-e2e/serviceaccount-signer.private"
-AZURE_WORKLOAD_IDENTITIES_FILE="/etc/hypershift-ci-jobs-self-managed-azure-e2e/workload-identities.json"
-
-PULL_SECRET_PATH="/etc/ci-pull-credentials/.dockerconfigjson"
-
-RELEASE_IMAGE="${RELEASE_IMAGE_LATEST}"
-HC_LOCATION="${HYPERSHIFT_AZURE_LOCATION:-centralus}"
-
-# Read private NAT subnet ID from SHARED_DIR (written by setup-private-link step)
-if [[ ! -s "${SHARED_DIR}/azure_private_nat_subnet_id" ]]; then
-  echo "$(date) ERROR: azure_private_nat_subnet_id is required for the private guest cluster"
-  exit 1
-fi
-AZURE_PRIVATE_NAT_SUBNET_ID="$(cat "${SHARED_DIR}/azure_private_nat_subnet_id")"
-
-# External DNS domain flag (required for correct service publishing strategy,
-# especially for Private topology clusters where the API server must use Route)
-EXTERNAL_DNS_ARGS=""
-if [[ -n "${HYPERSHIFT_EXTERNAL_DNS_DOMAIN:-}" ]]; then
-  EXTERNAL_DNS_ARGS="--external-dns-domain=${HYPERSHIFT_EXTERNAL_DNS_DOMAIN}"
-fi
-
-# Marketplace image flags
-ETCD_STORAGE_CLASS_ARGS=""
-if [[ -n "${HYPERSHIFT_ETCD_STORAGE_CLASS:-}" ]]; then
-  ETCD_STORAGE_CLASS_ARGS="--etcd-storage-class=${HYPERSHIFT_ETCD_STORAGE_CLASS}"
-fi
-
-MARKETPLACE_ARGS=""
-if [[ -n "${HYPERSHIFT_AZURE_MARKETPLACE_IMAGE_PUBLISHER:-}" ]]; then
-  MARKETPLACE_ARGS="--marketplace-publisher=${HYPERSHIFT_AZURE_MARKETPLACE_IMAGE_PUBLISHER} --marketplace-offer=${HYPERSHIFT_AZURE_MARKETPLACE_IMAGE_OFFER}"
-  if [[ -n "${HYPERSHIFT_AZURE_MARKETPLACE_IMAGE_SKU:-}" ]]; then
-    MARKETPLACE_ARGS="${MARKETPLACE_ARGS} --marketplace-sku=${HYPERSHIFT_AZURE_MARKETPLACE_IMAGE_SKU}"
-  elif [[ -f "${SHARED_DIR}/azure-marketplace-image-sku" ]]; then
-    MARKETPLACE_ARGS="${MARKETPLACE_ARGS} --marketplace-sku=$(cat "${SHARED_DIR}/azure-marketplace-image-sku")"
-  fi
-  if [[ -n "${HYPERSHIFT_AZURE_MARKETPLACE_IMAGE_VERSION:-}" ]]; then
-    MARKETPLACE_ARGS="${MARKETPLACE_ARGS} --marketplace-version=${HYPERSHIFT_AZURE_MARKETPLACE_IMAGE_VERSION}"
-  elif [[ -f "${SHARED_DIR}/azure-marketplace-image-version" ]]; then
-    MARKETPLACE_ARGS="${MARKETPLACE_ARGS} --marketplace-version=$(cat "${SHARED_DIR}/azure-marketplace-image-version")"
-  fi
-fi
-
-# Common flags for all self-managed clusters
-COMMON_FLAGS="--node-pool-replicas=${HYPERSHIFT_NODE_COUNT} \
-  --base-domain=${HYPERSHIFT_BASE_DOMAIN} \
-  --pull-secret=${PULL_SECRET_PATH} \
-  --azure-creds=${AZURE_CREDS} \
-  --location=${HC_LOCATION} \
-  --release-image=${RELEASE_IMAGE} \
-  --oidc-issuer-url=${AZURE_OIDC_ISSUER_URL} \
-  --sa-token-issuer-private-key-path=${AZURE_SA_TOKEN_ISSUER_KEY_PATH} \
-  --workload-identities-file=${AZURE_WORKLOAD_IDENTITIES_FILE} \
-  --assign-service-principal-roles \
-  --dns-zone-rg-name=os4-common \
-  --generate-ssh \
-  ${EXTERNAL_DNS_ARGS} \
-  ${MARKETPLACE_ARGS} \
-  ${ETCD_STORAGE_CLASS_ARGS}"
-
-# Create public cluster
-echo "$(date) Creating public self-managed cluster: ${PUBLIC_NAME}"
-/usr/bin/hypershift create cluster azure \
-  --name="${PUBLIC_NAME}" \
-  ${COMMON_FLAGS} &
-PUBLIC_PID=$!
-
-# Create private cluster
-PRIVATE_EXTRA="--endpoint-access-private-nat-subnet-id=${AZURE_PRIVATE_NAT_SUBNET_ID}"
-echo "$(date) Creating private self-managed cluster: ${PRIVATE_NAME}"
-/usr/bin/hypershift create cluster azure \
-  --name="${PRIVATE_NAME}" \
-  --endpoint-access=Private \
-  ${COMMON_FLAGS} \
-  ${PRIVATE_EXTRA} &
-PRIVATE_PID=$!
-
-# Create OAuth LoadBalancer cluster
-echo "$(date) Creating OAuth LB self-managed cluster: ${OAUTH_LB_NAME}"
-/usr/bin/hypershift create cluster azure \
-  --name="${OAUTH_LB_NAME}" \
-  --oauth-publishing-strategy=LoadBalancer \
-  ${COMMON_FLAGS} &
-OAUTH_LB_PID=$!
-
-# Wait for create commands to complete
-echo "$(date) Waiting for cluster create commands to finish..."
-FAILED=0
-wait ${PUBLIC_PID} || FAILED=1
-echo "$(date) Public cluster create command completed"
-wait ${PRIVATE_PID} || FAILED=1
-echo "$(date) Private cluster create command completed"
-wait ${OAUTH_LB_PID} || FAILED=1
-echo "$(date) OAuth LB cluster create command completed"
-if [[ ${FAILED} -ne 0 ]]; then
-  echo "$(date) ERROR: One or more cluster create commands failed"
-  exit 1
-fi
-
-# Patch the public cluster to set OperatorConfiguration for Ingress Operator.
-# The v2 e2e test ValidateIngressOperatorConfiguration expects this to be set on
-# public clusters (mirroring the BeforeApply hook used in the v1 CreateCluster test).
-echo "$(date) Patching public cluster ${PUBLIC_NAME} with OperatorConfiguration..."
-oc patch hostedcluster "${PUBLIC_NAME}" -n clusters --type=merge -p '
-{
-  "spec": {
-    "operatorConfiguration": {
-      "ingressOperator": {
-        "endpointPublishingStrategy": {
-          "type": "LoadBalancerService",
-          "loadBalancer": {
-            "scope": "Internal"
-          }
-        }
-      }
-    }
-  }
-}'
-echo "$(date) Public cluster ${PUBLIC_NAME} patched with OperatorConfiguration"
-
-# Wait for clusters to become available
-echo "$(date) Waiting for public cluster to become available..."
-oc wait --timeout=30m --for=condition=Available --namespace=clusters "hostedcluster/${PUBLIC_NAME}"
-echo "$(date) Public cluster is available"
-
-echo "$(date) Waiting for private cluster to become available..."
-oc wait --timeout=30m --for=condition=Available --namespace=clusters "hostedcluster/${PRIVATE_NAME}"
-echo "$(date) Private cluster is available"
-
-echo "$(date) Waiting for OAuth LB cluster to become available..."
-oc wait --timeout=30m --for=condition=Available --namespace=clusters "hostedcluster/${OAUTH_LB_NAME}"
-echo "$(date) OAuth LB cluster is available"
-
-# Wait for version rollout to complete on all clusters in parallel (via management API, same as AWS/GCP v2)
-echo "$(date) Starting parallel version rollout checks..."
-set +e
-
-echo "$(date) Waiting for version rollout on ${PUBLIC_NAME}..."
-CLUSTER_CHECK="${PUBLIC_NAME}" timeout 45m bash -c '
-  until [[ "$(oc get -n clusters hostedcluster/${CLUSTER_CHECK} -o jsonpath='"'"'{.status.version.history[?(@.state!="")].state}'"'"')" = "Completed" ]]; do
-    sleep 15
-  done
-' &
-ROLLOUT_PID_PUB=$!
-
-echo "$(date) Waiting for version rollout on ${PRIVATE_NAME}..."
-CLUSTER_CHECK="${PRIVATE_NAME}" timeout 45m bash -c '
-  until [[ "$(oc get -n clusters hostedcluster/${CLUSTER_CHECK} -o jsonpath='"'"'{.status.version.history[?(@.state!="")].state}'"'"')" = "Completed" ]]; do
-    sleep 15
-  done
-' &
-ROLLOUT_PID_PRV=$!
-
-echo "$(date) Waiting for version rollout on ${OAUTH_LB_NAME}..."
-CLUSTER_CHECK="${OAUTH_LB_NAME}" timeout 45m bash -c '
-  until [[ "$(oc get -n clusters hostedcluster/${CLUSTER_CHECK} -o jsonpath='"'"'{.status.version.history[?(@.state!="")].state}'"'"')" = "Completed" ]]; do
-    sleep 15
-  done
-' &
-ROLLOUT_PID_OAU=$!
-
-echo "$(date) Waiting for all version rollout checks to complete..."
-FAILED_READY=0
-for CLUSTER_PID in "${PUBLIC_NAME}:${ROLLOUT_PID_PUB}" "${PRIVATE_NAME}:${ROLLOUT_PID_PRV}" "${OAUTH_LB_NAME}:${ROLLOUT_PID_OAU}"; do
-  CLUSTER="${CLUSTER_PID%%:*}"
-  PID="${CLUSTER_PID##*:}"
-  wait ${PID}
-  ROLLOUT_RC=$?
-  if [[ ${ROLLOUT_RC} -ne 0 ]]; then
-    echo "$(date) ERROR: version rollout timed out for ${CLUSTER}"
-    echo "--- Diagnostic dump for ${CLUSTER} ---"
-    echo "HostedCluster conditions:"
-    oc get -n clusters hostedcluster/${CLUSTER} -o jsonpath='{range .status.conditions[*]}{.type}{"\t"}{.status}{"\t"}{.reason}{"\t"}{.message}{"\n"}{end}' || true
-    echo ""
-    echo "NodePool status:"
-    oc get -n clusters nodepool/${CLUSTER} -o jsonpath='{range .status.conditions[*]}{.type}{"\t"}{.status}{"\t"}{.reason}{"\t"}{.message}{"\n"}{end}' || true
-    echo ""
-    echo "Machines in clusters-${CLUSTER}:"
-    oc get machines.cluster.x-k8s.io -n "clusters-${CLUSTER}" -o wide 2>/dev/null || true
-    echo ""
-    echo "AzureMachines in clusters-${CLUSTER}:"
-    oc get azuremachines.infrastructure.cluster.x-k8s.io -n "clusters-${CLUSTER}" -o wide 2>/dev/null || true
-    echo ""
-    echo "Pods not ready in clusters-${CLUSTER}:"
-    oc get pods -n "clusters-${CLUSTER}" --field-selector=status.phase!=Running,status.phase!=Succeeded 2>/dev/null || true
-    echo "--- End diagnostic dump ---"
-    cat << EOF > "${ARTIFACT_DIR}/junit_hosted_cluster_${CLUSTER}.xml"
-<?xml version="1.0" encoding="UTF-8"?>
-<testsuite name="hypershift install ${CLUSTER}" tests="1" failures="1">
-  <testcase name="hosted cluster version rollout succeeds">
-    <failure message="hosted cluster version rollout never completed">
-      <![CDATA[
-error: hosted cluster version rollout never completed for ${CLUSTER}
-Degraded: $(oc get -n clusters hostedcluster/${CLUSTER} -o jsonpath='{.status.conditions[?(@.type=="Degraded")].message}')
-ClusterVersionSucceeding: $(oc get -n clusters hostedcluster/${CLUSTER} -o jsonpath='{.status.conditions[?(@.type=="ClusterVersionSucceeding")].message}')
-      ]]>
-    </failure>
-  </testcase>
-</testsuite>
-EOF
-    FAILED_READY=1
-  else
-    echo "$(date) Version rollout completed for ${CLUSTER}"
-    cat << EOF > "${ARTIFACT_DIR}/junit_hosted_cluster_${CLUSTER}.xml"
-<?xml version="1.0" encoding="UTF-8"?>
-<testsuite name="hypershift install ${CLUSTER}" tests="1" failures="0">
-  <testcase name="hosted cluster version rollout succeeds">
-    <system-out>
-      <![CDATA[
-info: hosted cluster version rollout completed successfully for ${CLUSTER}
-      ]]>
-    </system-out>
-  </testcase>
-</testsuite>
-EOF
-  fi
-done
-set -e
-if [[ ${FAILED_READY} -ne 0 ]]; then
-  exit 1
-fi
-
-# Write cluster names to shared dir
-echo "${PUBLIC_NAME}" > "${SHARED_DIR}/cluster-name-public"
-echo "${PRIVATE_NAME}" > "${SHARED_DIR}/cluster-name-private"
-echo "${OAUTH_LB_NAME}" > "${SHARED_DIR}/cluster-name-oauth-lb"
-
-echo "$(date) All self-managed guest clusters are ready"
+/hypershift/bin/create-guests
diff --git a/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-ref.yaml b/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-ref.yaml
index c5dc785ee1951..d7e665080d3bf 100644
--- a/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-ref.yaml
+++ b/ci-operator/step-registry/hypershift/azure/create-selfmanaged-guests/hypershift-azure-create-selfmanaged-guests-ref.yaml
@@ -15,6 +15,8 @@ ref:
   dependencies:
   - name: "release:latest"
     env: RELEASE_IMAGE_LATEST
+  - name: "release:n1minor"
+    env: OCP_IMAGE_N1
   env:
   - name: HYPERSHIFT_NODE_COUNT
     default: "3"
@@ -43,7 +45,10 @@ ref:
   - name: HYPERSHIFT_ETCD_STORAGE_CLASS
     default: ""
     documentation: "The persistent volume storage class for etcd data volumes."
-  from: hypershift-operator
+  - name: HYPERSHIFT_PLATFORM
+    default: ""
+    documentation: "The cloud platform for the self-managed clusters (e.g. azure). Set in the workflow env block."
+  from: hypershift-tests
   grace_period: 5m0s
   resources:
     requests:
diff --git a/ci-operator/step-registry/hypershift/azure/destroy-selfmanaged-guests/hypershift-azure-destroy-selfmanaged-guests-chain.yaml b/ci-operator/step-registry/hypershift/azure/destroy-selfmanaged-guests/hypershift-azure-destroy-selfmanaged-guests-chain.yaml
index 041eef88b9a10..635880f8dd419 100644
--- a/ci-operator/step-registry/hypershift/azure/destroy-selfmanaged-guests/hypershift-azure-destroy-selfmanaged-guests-chain.yaml
+++ b/ci-operator/step-registry/hypershift/azure/destroy-selfmanaged-guests/hypershift-azure-destroy-selfmanaged-guests-chain.yaml
@@ -8,37 +8,15 @@ chain:
     - name: HYPERSHIFT_AZURE_LOCATION
       default: "centralus"
       documentation: "Specifies the Azure location of the clusters."
+    - name: HYPERSHIFT_PLATFORM
+      default: ""
+      documentation: "The cloud platform for the self-managed clusters (e.g. azure). Set in the workflow env block."
     commands: |-
       set -xuo pipefail
-
-      # Use the nested management cluster kubeconfig
       export KUBECONFIG="${SHARED_DIR}/management_cluster_kubeconfig"
-
-      AZURE_CREDS="/etc/hypershift-ci-jobs-self-managed-azure/credentials.json"
-      HC_LOCATION="${HYPERSHIFT_AZURE_LOCATION:-centralus}"
-
-      # Re-derive cluster names from job ID (same logic as create step)
-      PUBLIC_NAME="$(echo -n "${PROW_JOB_ID}-pub"|sha256sum|cut -c-20)"
-      PRIVATE_NAME="$(echo -n "${PROW_JOB_ID}-prv"|sha256sum|cut -c-20)"
-      OAUTH_LB_NAME="$(echo -n "${PROW_JOB_ID}-oau"|sha256sum|cut -c-20)"
-
-      # Destroy all clusters, continue per-cluster, but preserve failure signal
-      FAILED=0
-      for CLUSTER in "${PUBLIC_NAME}" "${PRIVATE_NAME}" "${OAUTH_LB_NAME}"; do
-        echo "$(date) Destroying self-managed cluster: ${CLUSTER}"
-        if ! /usr/bin/hypershift destroy cluster azure \
-          --azure-creds="${AZURE_CREDS}" \
-          --name="${CLUSTER}" \
-          --location="${HC_LOCATION}" \
-          --dns-zone-rg-name=os4-common \
-          --cluster-grace-period=40m; then
-          echo "$(date) WARNING: Failed to destroy cluster ${CLUSTER}" >&2
-          FAILED=1
-        fi
-        echo "$(date) Finished destroying cluster ${CLUSTER}"
-      done
-      exit "${FAILED}"
-    from: hypershift-operator
+      export HYPERSHIFT_BINARY="${HYPERSHIFT_BINARY:-/hypershift/bin/hypershift}"
+      /hypershift/bin/destroy-guests
+    from: hypershift-tests
     grace_period: 5m0s
     resources:
       requests:
diff --git a/ci-operator/step-registry/hypershift/azure/dump-selfmanaged-guests/hypershift-azure-dump-selfmanaged-guests-chain.yaml b/ci-operator/step-registry/hypershift/azure/dump-selfmanaged-guests/hypershift-azure-dump-selfmanaged-guests-chain.yaml
index 2c0b3ade66424..d8322e2c54cb1 100644
--- a/ci-operator/step-registry/hypershift/azure/dump-selfmanaged-guests/hypershift-azure-dump-selfmanaged-guests-chain.yaml
+++ b/ci-operator/step-registry/hypershift/azure/dump-selfmanaged-guests/hypershift-azure-dump-selfmanaged-guests-chain.yaml
@@ -3,31 +3,20 @@ chain:
   steps:
   - as: dump
     cli: latest
+    env:
+    - name: HYPERSHIFT_PLATFORM
+      default: ""
+      documentation: "The cloud platform for the self-managed clusters (e.g. azure). Set in the workflow env block."
     commands: |-
       set -xuo pipefail
-
-      # Use the nested management cluster kubeconfig
       export KUBECONFIG="${SHARED_DIR}/management_cluster_kubeconfig"
-
-      if [ -f "${SHARED_DIR}/proxy-conf.sh" ] ; then
+      if [ -f "${SHARED_DIR}/proxy-conf.sh" ]; then
+        set +x
         source "${SHARED_DIR}/proxy-conf.sh"
+        set -x
       fi
-
-      # Re-derive cluster names from job ID (same logic as create/destroy steps)
-      PUBLIC_NAME="$(echo -n "${PROW_JOB_ID}-pub"|sha256sum|cut -c-20)"
-      PRIVATE_NAME="$(echo -n "${PROW_JOB_ID}-prv"|sha256sum|cut -c-20)"
-      OAUTH_LB_NAME="$(echo -n "${PROW_JOB_ID}-oau"|sha256sum|cut -c-20)"
-
-      for CLUSTER in "${PUBLIC_NAME}" "${PRIVATE_NAME}" "${OAUTH_LB_NAME}"; do
-        echo "$(date) Dumping self-managed guest cluster: ${CLUSTER}"
-        DUMP_DIR="${ARTIFACT_DIR}/${CLUSTER}"
-        mkdir -p "${DUMP_DIR}"
-        bin/hypershift dump cluster \
-          --artifact-dir="${DUMP_DIR}" \
-          --dump-guest-cluster=true \
-          --name="${CLUSTER}" || echo "$(date) WARNING: Failed to dump cluster ${CLUSTER}"
-      done
-    from: hypershift-operator
+      /hypershift/bin/dump-guests --hypershift-binary=/hypershift/bin/hypershift
+    from: hypershift-tests
     resources:
       requests:
         cpu: 100m
diff --git a/ci-operator/step-registry/hypershift/azure/e2e/v2-self-managed/hypershift-azure-e2e-v2-self-managed-workflow.yaml b/ci-operator/step-registry/hypershift/azure/e2e/v2-self-managed/hypershift-azure-e2e-v2-self-managed-workflow.yaml
index f7c41ce18fc9a..8fc3b3d822ec3 100644
--- a/ci-operator/step-registry/hypershift/azure/e2e/v2-self-managed/hypershift-azure-e2e-v2-self-managed-workflow.yaml
+++ b/ci-operator/step-registry/hypershift/azure/e2e/v2-self-managed/hypershift-azure-e2e-v2-self-managed-workflow.yaml
@@ -33,6 +33,7 @@ workflow:
     - chain: hypershift-destroy-nested-management-cluster
     env:
       CLOUD_PROVIDER: "Azure"
+      HYPERSHIFT_PLATFORM: "azure"
       HYPERSHIFT_NODE_COUNT: "6"
       HYPERSHIFT_AZURE_LOCATION: "centralus"
       AZURE_SELF_MANAGED: "true"
diff --git a/ci-operator/step-registry/hypershift/azure/run-e2e-v2-selfmanaged/hypershift-azure-run-e2e-v2-selfmanaged-chain.yaml b/ci-operator/step-registry/hypershift/azure/run-e2e-v2-selfmanaged/hypershift-azure-run-e2e-v2-selfmanaged-chain.yaml
index 131a579ad83b2..65a380fea1401 100644
--- a/ci-operator/step-registry/hypershift/azure/run-e2e-v2-selfmanaged/hypershift-azure-run-e2e-v2-selfmanaged-chain.yaml
+++ b/ci-operator/step-registry/hypershift/azure/run-e2e-v2-selfmanaged/hypershift-azure-run-e2e-v2-selfmanaged-chain.yaml
@@ -3,66 +3,20 @@ chain:
   steps:
   - as: tests
     cli: latest
+    env:
+    - name: HYPERSHIFT_PLATFORM
+      default: ""
+      documentation: "The cloud platform for the self-managed clusters (e.g. azure). Set in the workflow env block."
     commands: |-
       set -xuo pipefail
-
-      # Use the nested management cluster kubeconfig
       export KUBECONFIG="${SHARED_DIR}/management_cluster_kubeconfig"
-      export EVENTUALLY_VERBOSE="false"
-
-      # Export private NAT subnet ID for private topology tests
-      AZURE_PRIVATE_NAT_SUBNET_ID=""
-      if [[ -f "${SHARED_DIR}/azure_private_nat_subnet_id" ]]; then
-        AZURE_PRIVATE_NAT_SUBNET_ID="$(cat "${SHARED_DIR}/azure_private_nat_subnet_id")"
-      fi
-      export AZURE_PRIVATE_NAT_SUBNET_ID
-
-      PUBLIC_NAME="$(cat "${SHARED_DIR}/cluster-name-public")"
-      PRIVATE_NAME="$(cat "${SHARED_DIR}/cluster-name-private")"
-      OAUTH_LB_NAME="$(cat "${SHARED_DIR}/cluster-name-oauth-lb")"
-
-      # Run all 3 test suites in parallel (clusters are independent)
-      echo "$(date) Running public cluster tests against ${PUBLIC_NAME}..."
-      E2E_HOSTED_CLUSTER_NAME="${PUBLIC_NAME}" \
-      E2E_HOSTED_CLUSTER_NAMESPACE=clusters \
-      bin/test-e2e-v2 \
-        --ginkgo.label-filter="self-managed-azure-public" \
-        --ginkgo.skip="KAS allowed CIDRs" \
-        --ginkgo.junit-report="${ARTIFACT_DIR}/junit_self_managed_azure_public.xml" \
-        --ginkgo.v &
-      PID_PUB=$!
-
-      echo "$(date) Running private topology tests against ${PRIVATE_NAME}..."
-      E2E_HOSTED_CLUSTER_NAME="${PRIVATE_NAME}" \
-      E2E_HOSTED_CLUSTER_NAMESPACE=clusters \
-      bin/test-e2e-v2 \
-        --ginkgo.label-filter="self-managed-azure-private" \
-        --ginkgo.junit-report="${ARTIFACT_DIR}/junit_self_managed_azure_private.xml" \
-        --ginkgo.v &
-      PID_PRV=$!
-
-      echo "$(date) Running OAuth LB tests against ${OAUTH_LB_NAME}..."
-      E2E_HOSTED_CLUSTER_NAME="${OAUTH_LB_NAME}" \
-      E2E_HOSTED_CLUSTER_NAMESPACE=clusters \
-      bin/test-e2e-v2 \
-        --ginkgo.label-filter="self-managed-azure-oauth-lb" \
-        --ginkgo.junit-report="${ARTIFACT_DIR}/junit_self_managed_azure_oauth_lb.xml" \
-        --ginkgo.v &
-      PID_OAU=$!
-
-      echo "$(date) Waiting for all test suites to complete..."
-      OVERALL_EXIT=0
-      wait ${PID_PUB} || OVERALL_EXIT=1
-      echo "$(date) Public cluster tests finished (exit=$?)"
-      wait ${PID_PRV} || OVERALL_EXIT=1
-      echo "$(date) Private topology tests finished (exit=$?)"
-      wait ${PID_OAU} || OVERALL_EXIT=1
-      echo "$(date) OAuth LB tests finished (exit=$?)"
-
-      exit ${OVERALL_EXIT}
-    timeout: 45m
+      /hypershift/bin/run-tests
+    timeout: 150m
     grace_period: 5m
     from: hypershift-tests
+    dependencies:
+    - name: "release:latest"
+      env: RELEASE_IMAGE_LATEST
     resources:
       requests:
         cpu: "1"