My Dedicated Endpoint is stuck at 1% / "Initializing" for a long time.
Last updated: July 31, 2026
Long initialization is often expected: the endpoint may be waking from sleep (min_replicas = 0) or loading with Host KV Cache enabled, which adds init time, and large models simply take longer.
However, if it looks genuinely hung after 30+ minutes, contact us with the model name, your team ID, endpoint ID, and the exact status text. Occasional hangs can also come from capacity pressure on our side — we'll check and help.