diff --git a/src/core/server/graph/tenant/loaders/Comments.ts b/src/core/server/graph/tenant/loaders/Comments.ts
index 7db234394..105374c1c 100644
--- a/src/core/server/graph/tenant/loaders/Comments.ts
+++ b/src/core/server/graph/tenant/loaders/Comments.ts
@@ -20,6 +20,9 @@ import {
   retrieveManyComments,
 } from "talk-server/models/comment";
 import { Connection } from "talk-server/models/connection";
+import { retrieveSharedModerationQueueQueuesCounts } from "talk-server/models/story/counts/shared";
+
+import { SingletonResolver } from "./util";
 
 /**
  * primeCommentsFromConnection will prime a given context with the comments
@@ -114,4 +117,11 @@ export default (ctx: Context) => ({
       // The cursor passed here is always going to be a number.
       before: before as number,
     }).then(primeCommentsFromConnection(ctx)),
+  sharedModerationQueueQueuesCounts: new SingletonResolver(() =>
+    retrieveSharedModerationQueueQueuesCounts(
+      ctx.mongo,
+      ctx.redis,
+      ctx.tenant.id
+    )
+  ),
 });
diff --git a/src/core/server/graph/tenant/loaders/Stories.ts b/src/core/server/graph/tenant/loaders/Stories.ts
index 47c946618..61857e694 100644
--- a/src/core/server/graph/tenant/loaders/Stories.ts
+++ b/src/core/server/graph/tenant/loaders/Stories.ts
@@ -11,8 +11,18 @@ import { findOrCreate } from "talk-server/services/stories";
 import { scraper } from "talk-server/services/stories/scraper";
 
 export default (ctx: TenantContext) => ({
-  findOrCreate: (input: FindOrCreateStoryInput) =>
-    findOrCreate(ctx.mongo, ctx.tenant, input, ctx.queue.scraper),
+  findOrCreate: new DataLoader(
+    (inputs: FindOrCreateStoryInput[]) =>
+      Promise.all(
+        inputs.map(input =>
+          findOrCreate(ctx.mongo, ctx.tenant, input, ctx.queue.scraper)
+        )
+      ),
+    {
+      // TODO: (wyattjoh) see if there's something we can do to improve the cache key
+      cacheKeyFn: (input: FindOrCreateStoryInput) => `${input.id}:${input.url}`,
+    }
+  ),
   story: new DataLoader<string, Story | null>(ids =>
     retrieveManyStories(ctx.mongo, ctx.tenant.id, ids)
   ),
diff --git a/src/core/server/graph/tenant/loaders/util.ts b/src/core/server/graph/tenant/loaders/util.ts
new file mode 100644
index 000000000..aaf93286a
--- /dev/null
+++ b/src/core/server/graph/tenant/loaders/util.ts
@@ -0,0 +1,26 @@
+/**
+ * SingletonResolver is a cached loader for a single result.
+ */
+export class SingletonResolver<T> {
+  private cache: Promise<T> | null = null;
+  private resolver: () => Promise<T>;
+
+  constructor(resolver: () => Promise<T>) {
+    this.resolver = resolver;
+  }
+
+  public load() {
+    if (this.cache) {
+      return this.cache;
+    }
+
+    const promise = this.resolver().then(result => {
+      return result;
+    });
+
+    // Set the promise on the cache.
+    this.cache = promise;
+
+    return promise;
+  }
+}
diff --git a/src/core/server/graph/tenant/resolvers/Query.ts b/src/core/server/graph/tenant/resolvers/Query.ts
index e6fec375c..1e6cadbd2 100644
--- a/src/core/server/graph/tenant/resolvers/Query.ts
+++ b/src/core/server/graph/tenant/resolvers/Query.ts
@@ -1,9 +1,9 @@
 import { GQLQueryTypeResolver } from "talk-server/graph/tenant/schema/__generated__/types";
-import { retrieveSharedModerationQueueQueuesCounts } from "talk-server/models/story";
+
 import { ModerationQueuesInput } from "./ModerationQueues";
 
 export const Query: GQLQueryTypeResolver<void> = {
-  story: (source, args, ctx) => ctx.loaders.Stories.findOrCreate(args),
+  story: (source, args, ctx) => ctx.loaders.Stories.findOrCreate.load(args),
   comment: (source, { id }, ctx) =>
     id ? ctx.loaders.Comments.comment.load(id) : null,
   comments: (source, args, ctx) => ctx.loaders.Comments.forFilter(args),
@@ -21,10 +21,6 @@ export const Query: GQLQueryTypeResolver<void> = {
     // We don't need to filter the connection, as this is tenant wide (tenant
     // filtering is completed at the model layer).
     connection: {},
-    counts: await retrieveSharedModerationQueueQueuesCounts(
-      ctx.mongo,
-      ctx.redis,
-      ctx.tenant.id
-    ),
+    counts: await ctx.loaders.Comments.sharedModerationQueueQueuesCounts.load(),
   }),
 };