NVIDIA vGPU Manager CVE-2026-47539

Requires local access to a Linux host running NVIDIA vGPU Virtual GPU Manager, AND the attacker must already hold high privileges (PR:H - administrative/root-equivalent on the host) to reach the vulnerable kernel-mode code path. No user interaction is needed (UI:N) and attack complexity is low (AC:L) once that privilege is held. The PR:H requirement is the dominant limiting factor: an unprivileged guest or user cannot trigger this, so realistic risk is confined to scenarios involving a malicious or compromised privileged host operator. Only Linux vGPU Manager deployments are affected; Windows and non-vGPU GPU display driver installations are out of scope.

This is a genuine but not urgent vulnerability. The CVSS 3.1 vector (AV:L/AC:L/PR:H/UI:N/S:U/C:H/I:H/A:H, 6.7) makes clear that despite the broad impact language (code execution, DoS, privesc, info-disclosure, tampering), exploitation is LOCAL and requires HIGH privileges - meaning the attacker must already hold administrative/root-level access on the vGPU hypervisor host. PR:H is the single most important limiting factor and is notable relative to most sibling CVEs in NVIDIA bulletin 5861, which are PR:L. No EPSS score, KEV listing, or public PoC is present in the provided data, so this is 'no public exploit identified at time of analysis.' There is no SSVC decision data supplied. The 'RCE' tag overstates reach: with AV:L this is local code execution, not remote. Net: a real but low-to-medium priority issue whose PR:H prerequisite substantially narrows the realistic attacker population. The CWE-681 (incorrect numeric conversion) root cause combined with C:H/I:H/A:H suggests a numeric truncation/sign-conversion error that can corrupt a size/index and lead to memory corruption in the kernel-mode layer.

An attacker who already possesses high-privilege (administrative/root) access on a Linux host running NVIDIA vGPU Virtual GPU Manager issues crafted input to the kernel-mode layer that triggers an incorrect numeric conversion, corrupting kernel memory. The resulting memory corruption is leveraged for kernel-level code execution or a host crash. No public exploit code is identified at time of analysis, and no user interaction is required.