diff --git a/frigate/detectors/plugins/axelera.py b/frigate/detectors/plugins/axelera.py
index d83a8629e5..2c6d981ede 100644
--- a/frigate/detectors/plugins/axelera.py
+++ b/frigate/detectors/plugins/axelera.py
@@ -198,7 +198,8 @@ class _AxeleraInference:
         self._frame_q: queue.Queue = queue.Queue(maxsize=64)
         self._cond = threading.Condition()
         # request ids a caller is still waiting on (guarded by _cond); the pump
-        # only delivers results whose decoded token is in this set
+        # only delivers results whose decoded token is in this set. The feeder
+        # also uses it as the "real frame outstanding" signal.
         self._waiting: set = set()
         self._results: dict[int, np.ndarray] = {}
         self._next_id = 0
@@ -241,9 +242,17 @@ class _AxeleraInference:
             try:
                 frame_id, frame = self._frame_q.get(timeout=1.0)
             except queue.Empty:
-                # keep-alive: no request pending, keep the SDK stream fed so the
-                # inference timeout never kills it during quiet periods. The
-                # None id means this frame's result is dropped by the pump.
+                # A real frame may be outstanding: do NOT inject a keep-alive in
+                # that window, otherwise the zeros queue in the SDK pipeline
+                # ahead of the pending frame and delay its result by one slot
+                # per keep-alive (observed: 3 s real-frame latency under 1 s
+                # keep-alive spacing vs 14 ms for back-to-back real frames).
+                with self._cond:
+                    if self._waiting:
+                        continue
+                # truly idle: keep the SDK stream fed so its inference timeout
+                # never kills it during quiet periods. The None id means this
+                # frame's result is dropped by the pump.
                 frame_id, frame = None, np.zeros(
                     (self._height, self._width, 3), np.uint8
                 )