University of Toronto researchers presented GPUBreach at Black Hat 2026 — a Rowhammer attack that escalates from an unprivileged CUDA kernel to a host root shell, bypassing IOMMU, threatening shared AI cloud GPU infrastructure.
Oligo Security presented at Black Hat USA 2026 this week: attackers have weaponised Ray, the open-source AI orchestration platform, into a self-propagating botnet with over 200,000 exposed servers. Compromised clusters autonomously scan for and infect new Ray deployments worldwide.
A 15-year-old Linux kernel use-after-free bug with a public 97%-reliable exploit gives any local attacker root in under five seconds. GPU training clusters, Jupyter servers, and LLM inference nodes are directly in scope.
CVE-2026-21858 gives unauthenticated attackers full code execution on n8n workflow servers. Langflow's new IDOR vulnerability has been added to CISA KEV. GitHub Copilot's MCP integration carries a prompt-injection-to-RCE chain. AI automation infrastructure is becoming a primary attack surface.
CVE-2026-33017, a CVSS 9.8 unauthenticated RCE in Langflow, has been added to the CISA KEV catalog after active exploitation deploying a self-spreading Monero cryptominer across exposed AI workflow environments.
Zenity Labs has documented attackers exploiting unpatched LiteLLM vulnerabilities and unauthenticated Ollama endpoints to run autonomous pentesting campaigns, steal compute, and exfiltrate data from victim organisations.