EUVD-2026-90064
GHSA-jhr5-vhjh-hw8h
Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessObtain low-priv local/guest access
-
DeliveryCraft malformed request to vGPU kernel layer
-
ExploitTrigger out-of-bounds read in kernel mode
-
ExecutionLeak kernel memory or crash service
-
ImpactEscalate privileges or disclose data
Vulnerability Assessment
| Weakness |
CWE
ASVS 5.0
V15 - Secure Coding and Architecture Bounds checking, read limits
MITRE ATT&CK
1 technique across 1 tactic: Collection The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Requires the target to run the NVIDIA vGPU Virtual GPU Manager on Linux with the kernel-mode driver loaded, and the attacker must already hold low-privileged local access to the host or an authorized guest/tenant with vGPU (guest driver) access (PR:L, AV:L). No user interaction is needed (UI:N). Exploitation is limited by the need for pre-existing local authenticated access and applies to the Linux vGPU Manager / guest driver components only - there is no remote or unauthenticated attack path. Risk is elevated specifically in multi-tenant vGPU/VDI/cloud-GPU deployments where the flaw could weaken tenant/host isolation. |
| Impact | The CVSS 3.1 vector (AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, 7.8 HIGH) describes a locally-exploitable flaw requiring low privileges and no user interaction - an already-authenticated local actor on the hypervisor host or a guest with vGPU access, not a remote attacker. No EPSS score, CISA KEV listing, or public POC was provided, so exploitation status is 'no public exploit identified at time of analysis.' The 'RCE' tag overstates reach: with AV:L this is local code execution, not network RCE, and 'Buffer Overflow' is loose since the root cause is an out-of-bounds READ (CWE-125), which most directly yields information disclosure and denial of service - the claimed integrity/code-execution impacts (I:H) are the vendor's worst-case chaining, not a guaranteed primitive from a read primitive alone. Net assessment: a genuine but medium-priority local privilege/isolation issue, most consequential in multi-tenant vGPU/VDI/cloud-GPU deployments where cross-tenant boundaries matter; not a drop-everything internet-facing threat. No conflicting network-vs-local signals - description and CVSS agree on local access. |
| Exploit ScenarioDerived | An authenticated low-privileged user on a Linux host running the NVIDIA vGPU Virtual GPU Manager (or a tenant with guest-driver access in a VDI/cloud-GPU environment) issues a crafted request to the kernel-mode layer that triggers an out-of-bounds read. The leaked kernel memory can disclose sensitive data or crash the service (DoS), and in a favorable configuration may be chained toward privilege escalation or code execution. No public exploit code is currently identified, and exploitation requires prior local access. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Sep 30, 2026
Out-of-bounds read in NVIDIA vGPU Virtual GPU Manager for Linux allows a local authenticated attacker with low privileges to potentially achieve code execution, denial of service, or information disclosure. The vulnerability resides in the kernel-mode driver and affects both the vGPU Manager and guest driver components. No public exploit code has been identified, and the flaw is not known to be actively exploited. Integrity impact is assessed as low, but confidentiality and availability impacts are high, making this a significant risk in multi-tenant vGPU deployments.
RemediationAI
NVIDIA has published a security advisory at https://github.com/NVIDIA/product-security/tree/main/2026/5861; consult it for patch availability and specific fixed versions. As a workaround, restrict local access to the hypervisor host and vGPU-enabled guests, limit vGPU access to trusted tenants, and monitor for anomalous GPU or kernel activity. Note that restricting access may impact multi-tenant functionality, and no single workaround fully mitigates the risk without patching.
Affected VersionsENISA
Affects your stack
Find out whether the stacks and servers you run are on an affected version.
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The vulnerability is a classic out-of-bounds read (CWE-125) in the kernel mode layer of NVIDIA's vGPU virtualization stack for Linux. The vGPU Virtual GPU Manager and guest driver facilitate GPU sharing across virtual machines; a flaw in memory bounds checking could allow a malicious or compromised guest or a local user on the host to read beyond allocated buffers. Because the code runs in kernel mode, successful exploitation could leak sensitive information from kernel memory, crash the system (denial of service), or potentially lead to privilege escalation. The CVSS vector indicates the attack is local (AV:L), requires low privileges (PR:L), and no user interaction (UI:N), with high confidentiality and availability impacts but low integrity impact per independent assessment.
Know when Virtual Gpu Manager gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
Local low-priv access with no UI (AV:L/AC:L/PR:L/UI:N); OOB read strongly supports info-disclosure (C:H) and DoS (A:H) but only weak integrity impact (I:L), keeping S:U.
Primary rating from Vendor (nvidia).
CVSS VectorVendor: nvidia
DescriptionCVE.org
NVIDIA vGPU Virtual GPU Manager for Linux contains a vulnerability in the kernel mode layer where an attacker could cause an out-of-bounds read. A successful exploit of this vulnerability might lead to code execution, denial of service, escalation of privileges, information disclosure, and data tampering.
Articles & Coverage 2
Executive ViewAI
NVIDIA vGPU Virtual GPU Manager for Linux contains a high-severity kernel-mode flaw that could allow a local low-privileged user to execute code, disrupt service, or access sensitive data on vGPU-enabled systems. Because the affected components are used in multi-tenant virtualization, successful exploitation could undermine shared GPU infrastructure, expose tenant data, or cause availability outages. Leadership should prioritize inventory and interim mitigation for all Linux vGPU deployments until NVIDIA releases a fixed version; no public exploit code has been identified at time of analysis.
No vendor-released patch identified at time of analysis; interim mitigations include restricting local access to vGPU hosts to trusted users, applying least privilege, isolating or segmenting multi-tenant vGPU workloads, disabling vGPU where not required, enabling kernel and driver logging, monitoring for abnormal GPU driver crashes or data access, and tracking the NVIDIA advisory for CVE-2026-47520.
Threat Intelligence Overview
NVD
CVE.org
EPSS
Vulnrichment
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
A buffer overflow occurs when a program writes more data to a memory buffer than it was allocated to hold, causing the excess data to spill into adjacent memory regions.
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.
Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.
Denial of Service attacks render applications or systems unavailable by overwhelming resources or triggering failure conditions.
More in Virtual Gpu Manager
View allLocal privilege escalation and code execution in the NVIDIA Display Driver for Linux (GeForce, RTX/Quadro/NVS, Tesla, an
An out-of-bounds read in the kernel-mode layer of NVIDIA GPU Display Driver on Windows and Linux allows a local, authent
Local privilege escalation and potential code execution in NVIDIA GPU Display Driver for Windows and Linux arise from a
Local privilege escalation in the kernel-mode layer of the NVIDIA GPU Display Driver for Linux allows an unprivileged lo
Privilege escalation in NVIDIA's GPU Display Driver for Windows lets a local, low-privileged authenticated user bypass a
Local privilege escalation in NVIDIA Display Driver for Linux (GeForce, RTX/Quadro/NVS, Tesla, and Virtual GPU Manager b
Local privilege escalation in NVIDIA Display Driver for Windows (GeForce, RTX/Quadro/NVS, Tesla, and vGPU guest/manager
Incorrect resource transfer between spheres in the NVIDIA vGPU Virtual GPU Manager for Linux allows a locally positioned
An out-of-bounds write in the kernel-mode layer of NVIDIA vGPU Virtual GPU Manager for Linux lets a local attacker with
Local privilege escalation, code execution, denial of service, and information disclosure affect Linux hosts running NVI
Local privilege escalation in NVIDIA vGPU Virtual GPU Manager for Linux stems from an out-of-bounds read in firmware (CW
Local attackers with low privileges on a Linux host running NVIDIA vGPU Virtual GPU Manager can trigger an out-of-bounds
Exploit-DB