High severity7.5NVD Advisory· Published Sep 21, 2026· Updated Sep 29, 2026
CVE-2026-94627
CVE-2026-94627
Description
vLLM Mooncake connector through 0.29.0 fails to properly manage GPU KV cache block ownership when concurrent child requests share a single transfer ID in prefill/decode disaggregated deployments. Attackers can trigger GPU memory exhaustion by submitting completion requests with multiple prompts, causing orphaned KV cache blocks to accumulate until process restart and eventually preventing legitimate requests from executing.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
Affected products
3Patches
Vulnerability mechanics
References
3- github.com/vllm-project/vllm/pull/49796nvdIssue TrackingPatch
- www.vulncheck.com/advisories/vllm-through-0.29.0-gpu-kv-cache-leak-via-mooncake-transfer-id-collisionnvdThird Party Advisory
- github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.pynvdProduct
News mentions
1- VLLM: Batch of 17 Vulnerabilities Disclosed, Affecting LLM Inference Engine StabilityVypr Intelligence · Sep 21, 2026