diff --git a/src/chunk_manager/backend.ts b/src/chunk_manager/backend.ts
index df56bb02bb..004fbf635d 100644
--- a/src/chunk_manager/backend.ts
+++ b/src/chunk_manager/backend.ts
@@ -30,6 +30,7 @@ import {
   ChunkState,
   getChunkDownloadStatisticIndex,
   getChunkStateStatisticIndex,
+  MemoryLimitFlags,
   numChunkMemoryStatistics,
   numChunkStatistics,
   REQUEST_CHUNK_STATISTICS_RPC_ID,
@@ -727,6 +728,13 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
 
   gpuMemoryChanged = new NullarySignal();
 
+  /**
+   * Bit mask of `MemoryLimitFlags` indicating which memory limits are currently
+   * preventing visible chunks from being loaded.  Written at the end of each
+   * `process()` pass and observed by the frontend to show a status message.
+   */
+  memoryLimitReached: SharedWatchableValue<number>;
+
   private numQueued = 0;
   private numFailed = 0;
   private gpuMemoryGeneration = 0;
@@ -751,6 +759,7 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
       getCapacity(options.downloadCapacity),
     ];
     this.computeCapacity = getCapacity(options.computeCapacity);
+    this.memoryLimitReached = rpc.get(options.memoryLimitReached);
   }
 
   scheduleUpdate() {
@@ -912,7 +921,11 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
     this.addChunkToQueues_(chunk);
   }
 
-  private processGPUPromotions_() {
+  /**
+   * @returns true if a `VISIBLE`-tier chunk could not be promoted to GPU memory
+   *     because the GPU memory limit is full of equal-or-higher-priority chunks.
+   */
+  private processGPUPromotions_(): boolean {
     const queueManager = this;
     function evictFromGPUMemory(chunk: Chunk) {
       queueManager.freeChunkGPUMemory(chunk);
@@ -941,11 +954,14 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
           evictFromGPUMemory,
         )
       ) {
-        break;
+        // If the blocked chunk is currently visible, the GPU memory limit is
+        // preventing data the user is looking at from being displayed.
+        return priorityTier === ChunkPriorityTier.VISIBLE;
       }
       this.copyChunkToGPU(promotionCandidate);
       this.updateChunkState(promotionCandidate, ChunkState.GPU_MEMORY);
     }
+    return false;
   }
 
   freeChunkGPUMemory(chunk: Chunk) {
@@ -1003,7 +1019,14 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
     rpc.invoke("Chunk.update", msg, transfers);
   }
 
-  private processQueuePromotions_() {
+  /**
+   * @returns true if a `VISIBLE`-tier chunk could not be downloaded because the
+   *     system memory limit is full of equal-or-higher-priority chunks.  The
+   *     per-source download/compute concurrency limits are request-rate limits
+   *     rather than memory limits and are intentionally not reported here.
+   */
+  private processQueuePromotions_(): boolean {
+    let systemMemoryBlocked = false;
     const evict = (chunk: Chunk) => {
       switch (chunk.state) {
         case ChunkState.DOWNLOADING:
@@ -1061,6 +1084,9 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
             evict,
           )
         ) {
+          if (priorityTier === ChunkPriorityTier.VISIBLE) {
+            systemMemoryBlocked = true;
+          }
           return;
         }
         this.updateChunkState(promotionCandidate, ChunkState.DOWNLOADING);
@@ -1084,6 +1110,7 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
       this.computeEvictionQueue.candidates(),
       this.computeCapacity,
     );
+    return systemMemoryBlocked;
   }
 
   process() {
@@ -1092,9 +1119,17 @@ export class ChunkQueueManager extends SharedObjectCounterpart {
     }
     this.updatePending = null;
     const gpuMemoryGeneration = this.gpuMemoryGeneration;
-    this.processGPUPromotions_();
-    this.processQueuePromotions_();
+    const gpuMemoryBlocked = this.processGPUPromotions_();
+    const systemMemoryBlocked = this.processQueuePromotions_();
     this.logStatistics();
+    let memoryLimitFlags = MemoryLimitFlags.NONE;
+    if (gpuMemoryBlocked) {
+      memoryLimitFlags |= MemoryLimitFlags.GPU;
+    }
+    if (systemMemoryBlocked) {
+      memoryLimitFlags |= MemoryLimitFlags.SYSTEM;
+    }
+    this.memoryLimitReached.value = memoryLimitFlags;
     if (this.gpuMemoryGeneration !== gpuMemoryGeneration) {
       this.gpuMemoryChanged.dispatch();
     }
diff --git a/src/chunk_manager/base.ts b/src/chunk_manager/base.ts
index d727085373..fd80b0801a 100644
--- a/src/chunk_manager/base.ts
+++ b/src/chunk_manager/base.ts
@@ -73,6 +73,17 @@ export enum ChunkMemoryStatistics {
 
 export const numChunkMemoryStatistics = 3;
 
+/**
+ * Bit flags indicating which memory limits are currently blocking visible
+ * (`ChunkPriorityTier.VISIBLE`) chunks from being loaded.  Shared from the
+ * backend to the frontend so the UI can surface a status message.
+ */
+export enum MemoryLimitFlags {
+  NONE = 0,
+  GPU = 1,
+  SYSTEM = 2,
+}
+
 export const numChunkDownloadStatistics = 2;
 
 export const numChunkStatistics =
diff --git a/src/chunk_manager/frontend.ts b/src/chunk_manager/frontend.ts
index 3df76ce075..b37105b4d8 100644
--- a/src/chunk_manager/frontend.ts
+++ b/src/chunk_manager/frontend.ts
@@ -24,6 +24,7 @@ import {
   CHUNK_QUEUE_MANAGER_RPC_ID,
   CHUNK_SOURCE_INVALIDATE_RPC_ID,
   ChunkState,
+  MemoryLimitFlags,
   REQUEST_CHUNK_STATISTICS_RPC_ID,
 } from "#src/chunk_manager/base.js";
 import { SharedWatchableValue } from "#src/shared_watchable_value.js";
@@ -112,6 +113,12 @@ export class ChunkQueueManager extends SharedObject {
 
   enablePrefetch = new TrackableBoolean(true, true);
 
+  /**
+   * Bit mask of `MemoryLimitFlags` set by the backend indicating which memory
+   * limits are currently preventing visible chunks from being loaded.
+   */
+  memoryLimitReached: SharedWatchableValue<number>;
+
   constructor(
     rpc: RPC,
     public gl: GL,
@@ -125,6 +132,10 @@ export class ChunkQueueManager extends SharedObject {
   ) {
     super();
 
+    this.memoryLimitReached = this.registerDisposer(
+      SharedWatchableValue.make(rpc, MemoryLimitFlags.NONE),
+    );
+
     const makeCapacityCounterparts = (capacity: CapacitySpecification) => {
       return {
         itemLimit: this.registerDisposer(
@@ -144,6 +155,7 @@ export class ChunkQueueManager extends SharedObject {
       enablePrefetch: this.registerDisposer(
         SharedWatchableValue.makeFromExisting(rpc, this.enablePrefetch),
       ).rpcId,
+      memoryLimitReached: this.memoryLimitReached.rpcId,
     });
   }
 
diff --git a/src/ui/memory_limit_status.ts b/src/ui/memory_limit_status.ts
new file mode 100644
index 0000000000..a40f68e0b7
--- /dev/null
+++ b/src/ui/memory_limit_status.ts
@@ -0,0 +1,128 @@
+/**
+ * @license
+ * Copyright 2026 Google Inc.
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *      http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/**
+ * @file Shows a dismissable status message when Neuroglancer cannot display
+ * visible data because it has reached its GPU or system memory limit.
+ */
+
+import { MemoryLimitFlags } from "#src/chunk_manager/base.js";
+import type { ChunkQueueManager } from "#src/chunk_manager/frontend.js";
+import { StatusMessage } from "#src/status.js";
+import type { RefCounted } from "#src/util/disposable.js";
+
+/**
+ * How long the memory limit must stay clear before the status message is
+ * hidden and the dismissed state is reset.  Debouncing the falling edge avoids
+ * flapping the message during ordinary chunk churn (e.g. while panning) and
+ * ensures the message only reappears after memory pressure has genuinely
+ * subsided and is hit again.
+ */
+const CLEAR_DEBOUNCE_MS = 1500;
+
+function describeMemoryLimit(flags: number): string {
+  const limits: string[] = [];
+  if (flags & MemoryLimitFlags.GPU) {
+    limits.push("GPU");
+  }
+  if (flags & MemoryLimitFlags.SYSTEM) {
+    limits.push("system");
+  }
+  const which = limits.join(" and ");
+  return (
+    `Some data is not being displayed because Neuroglancer has reached its ` +
+    `${which} memory limit. You can increase the memory limits in the ` +
+    `settings panel (gear icon at the top right).`
+  );
+}
+
+/**
+ * Registers a handler that surfaces a status message whenever the memory limit
+ * blocks visible chunks from loading, and hides it once memory pressure clears.
+ *
+ * The message can be dismissed by the user; once dismissed it will not reappear
+ * until the memory limit has been continuously clear for `CLEAR_DEBOUNCE_MS`
+ * and is subsequently reached again.
+ */
+export function registerMemoryLimitStatusMessage(
+  context: RefCounted,
+  chunkQueueManager: ChunkQueueManager,
+) {
+  const watchable = chunkQueueManager.memoryLimitReached;
+  let statusMessage: StatusMessage | undefined;
+  let shownFlags = MemoryLimitFlags.NONE;
+  let dismissed = false;
+  let clearTimer: number | undefined;
+
+  const hideMessage = () => {
+    if (statusMessage !== undefined) {
+      statusMessage.dispose();
+      statusMessage = undefined;
+    }
+    shownFlags = MemoryLimitFlags.NONE;
+  };
+
+  const showMessage = (flags: number) => {
+    if (statusMessage !== undefined && shownFlags === flags) {
+      return;
+    }
+    hideMessage();
+    shownFlags = flags;
+    const message = new StatusMessage(/*delay=*/ false);
+    message.setPreventFocusChangeOnMouseDown(true);
+    message.element.textContent = describeMemoryLimit(flags) + " ";
+    const dismissButton = document.createElement("button");
+    dismissButton.textContent = "Dismiss";
+    dismissButton.addEventListener("click", () => {
+      dismissed = true;
+      hideMessage();
+    });
+    message.element.appendChild(dismissButton);
+    statusMessage = message;
+  };
+
+  const cancelClearTimer = () => {
+    if (clearTimer !== undefined) {
+      window.clearTimeout(clearTimer);
+      clearTimer = undefined;
+    }
+  };
+
+  const update = () => {
+    const flags = watchable.value;
+    if (flags !== MemoryLimitFlags.NONE) {
+      cancelClearTimer();
+      if (!dismissed) {
+        showMessage(flags);
+      }
+    } else if (clearTimer === undefined) {
+      // Debounce the falling edge: only clear once memory pressure has stayed
+      // resolved for a while.
+      clearTimer = window.setTimeout(() => {
+        clearTimer = undefined;
+        dismissed = false;
+        hideMessage();
+      }, CLEAR_DEBOUNCE_MS);
+    }
+  };
+
+  context.registerDisposer(watchable.changed.add(update));
+  context.registerDisposer(() => {
+    cancelClearTimer();
+    hideMessage();
+  });
+  update();
+}
diff --git a/src/viewer.ts b/src/viewer.ts
index a985d0c2c3..b1aec47f2a 100644
--- a/src/viewer.ts
+++ b/src/viewer.ts
@@ -89,6 +89,7 @@ import {
   LayerListPanelState,
 } from "#src/ui/layer_list_panel.js";
 import { LayerSidePanelManager } from "#src/ui/layer_side_panel.js";
+import { registerMemoryLimitStatusMessage } from "#src/ui/memory_limit_status.js";
 import { setupPositionDropHandlers } from "#src/ui/position_drag_and_drop.js";
 import { ScreenshotDialog } from "#src/ui/screenshot_menu.js";
 import { SelectionDetailsPanel } from "#src/ui/selection_details.js";
@@ -675,6 +676,8 @@ export class Viewer extends RefCounted implements ViewerState {
       }),
     );
 
+    registerMemoryLimitStatusMessage(this, this.dataContext.chunkQueueManager);
+
     this.makeUI();
     this.updateShowBorders();