diff --git a/utils/helm/speckle-server/templates/preview_service/deployment.yml b/utils/helm/speckle-server/templates/preview_service/deployment.yml
index d3cb013a5..2221bd996 100644
--- a/utils/helm/speckle-server/templates/preview_service/deployment.yml
+++ b/utils/helm/speckle-server/templates/preview_service/deployment.yml
@@ -48,6 +48,9 @@ spec:
           limits:
             cpu: {{ .Values.preview_service.limits.cpu }}
             memory: {{ .Values.preview_service.limits.memory }}
+            {{- if (and .Values.preview_service.nvidiaGPUEnabled (get .Values.preview_service.limits "nvidia.com/gpu"))}}
+            nvidia.com/gpu: {{ (get .Values.preview_service.limits "nvidia.com/gpu") }}
+            {{- end }}
 
         securityContext:
           allowPrivilegeEscalation: false
diff --git a/utils/helm/speckle-server/values.schema.json b/utils/helm/speckle-server/values.schema.json
index 6d2318c7f..caa81ac53 100644
--- a/utils/helm/speckle-server/values.schema.json
+++ b/utils/helm/speckle-server/values.schema.json
@@ -1915,6 +1915,11 @@
           "description": "Allows using a dedicated redis url for the preview service job queue",
           "default": false
         },
+        "nvidiaGPUEnabled": {
+          "type": "boolean",
+          "description": "If enabled, the Preview Service will be deployed with GPU support. This is only available on Kubernetes clusters with GPU support.",
+          "default": false
+        },
         "replicas": {
           "type": "number",
           "description": "The number of instances of the Preview Service pod to be deployed within the cluster.",
diff --git a/utils/helm/speckle-server/values.yaml b/utils/helm/speckle-server/values.yaml
index 1cf898736..493485228 100644
--- a/utils/helm/speckle-server/values.yaml
+++ b/utils/helm/speckle-server/values.yaml
@@ -1123,6 +1123,9 @@ preview_service:
   ## @param preview_service.dedicatedPreviewsQueue Allows using a dedicated redis url for the preview service job queue
   ##
   dedicatedPreviewsQueue: false
+  ## @param preview_service.nvidiaGPUEnabled If enabled, the Preview Service will be deployed with GPU support. This is only available on Kubernetes clusters with GPU support.
+  ##
+  nvidiaGPUEnabled: false
   ## @param preview_service.replicas The number of instances of the Preview Service pod to be deployed within the cluster.
   ##
   replicas: 1
@@ -1171,6 +1174,9 @@ preview_service:
     ## ref: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/
     ##
     memory: 4Gi
+    ## @param preview_service.limits.nvidia.com/gpu The maximum number of GPUs that will be made available to the Preview Service Pod. Ignored if preview_service.nvidiaGPUEnabled is false.
+    ## If this limit is exceeded, processes within the pod that request additional GPUs may be paused.
+    nvidia.com/gpu: 1
   autoscaling:
     ## @param preview_service.autoscaling.enabled
     ## @descriptionStart