GCD AI RADAR
incidents/AIR-2026-0138

CVE-2026-69147: vLLM video backend GPU memory allocation flaw

Attack on AI2 Disclosed vulnerabilityInfrastructure vulnerability

A vulnerability in vLLM prior to version 0.28.0 allows chat completion requests to force unexpected video backend decoders. This bypasses static GPU memory budgeting and can lead to unmanaged CUDA allocations.

disclosed CVENo confirmed exploitation, so this record is tracked as a disclosed vulnerability and not counted as an incident. It moves into the counts if CISA or the vendor confirms exploitation. Why

Details

type
Attack on AI
An AI model, agent or AI-powered product was the thing exploited, abused or leaked.
categories
  • Infrastructure vulnerability
    CVEs and advisories in AI frameworks, inference servers, MCP servers, AI developer tools and hosting stacks.
owasp llm top 10
mitre atlas
no technique for these categories
Both rows are derived from the type and categories (OWASP 2025 edition, ATLAS release 2026.09), not classified per record. How
severity
2 · Disclosed vulnerability
CVE or vendor advisory, patched or patchable, no confirmed exploitation.
targets
Framework / tooling
products
vLLM
incident date
16 Sep 2026
first reported
16 Sep 2026
origin
NVD
Vulnerability record from NVD / the GitHub Advisory Database. This product uses the NVD API but is not endorsed or certified by the NVD.
classifier confidence
0.95 · published at or above 0.80; see the methodology
tags
CVE-2026-69147 · vllm · denial-of-service
record
created 2026-10-18 06:33Z · updated 2026-10-18 07:00Z · json

Sources · 2

  1. 2026-09-16
    CVE-2026-69147: vLLM is an inference and serving engine for large language models.
    NVD · nvd.nist.gov/vuln/detail/CVE-2026-69147
  2. 2026-09-17
    CVE-2026-69147: vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation
    GitHub Advisory Database · github.com/advisories/GHSA-8pw2-6jv3-mj5j

Titles and links belong to their publishers. The summary above is written by GCD AI RADAR in its own words; data is CC BY-SA 4.0.

Related incidents