Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessGain low-privileged local code execution
-
DeliveryOpen NVIDIA kernel-mode driver interface
-
ExploitSend crafted IOCTL triggering out-of-bounds write
-
ExecutionCorrupt kernel memory structures
-
ImpactEscalate to SYSTEM/root or crash driver
Vulnerability Assessment
| Weakness |
CWE
ASVS 5.0
V15 - Secure Coding and Architecture Bounds checking, write limits
MITRE ATT&CK
2 techniques across 2 tactics: Execution, Privilege Escalation The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires local access to a system running the NVIDIA GPU Display Driver (Windows or Linux) with an affected GeForce, RTX/Quadro/NVS, or Tesla GPU, or the vGPU/Virtual GPU Manager component; the attacker must already hold low-privilege execution (PR:L) to reach the vulnerable kernel-mode interface. No user interaction is needed (UI:N) and complexity is low (AC:L). The key LIMITING factor is that this is not remotely exploitable (AV:L) - an attacker must first have a foothold on the host - which confines it to local privilege escalation and sandbox-escape scenarios rather than remote mass exploitation. |
| Impact | This is a genuine local privilege-escalation priority despite lacking exploitation signals. The CVSS 7.8 (HIGH) vector CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H reflects a low-privileged local attacker triggering an out-of-bounds write (CWE-787) in the kernel-mode layer of the NVIDIA GPU Display Driver, yielding full C:H/I:H/A:H impact - consistent with kernel memory corruption that escalates to SYSTEM/root. Attack vector is LOCAL (AV:L) with low complexity (AC:L) and no user interaction (UI:N), so any code already running on the host (a logged-in user, a sandboxed process, or malware) can attempt it. No EPSS score, CISA KEV listing, or public POC was provided, so there is no evidence of active exploitation; this is a vendor-reported (NVIDIA product-security) issue rather than an in-the-wild threat. No SSVC decision data was supplied. The absence of a network vector meaningfully lowers mass-exploitation risk, but because kernel driver bugs are a favored local-EoP and sandbox-escape primitive and the affected drivers are extremely widely deployed, this warrants prompt patching. No conflicting signals were detected in the provided data beyond the mislabeled 'Microsoft' vendor tag (see data_quality_flags). |
| Exploit ScenarioDerived | A low-privileged user or a process that has already achieved limited code execution on a Windows or Linux host with an NVIDIA GPU issues crafted requests (e.g., via IOCTLs/ioctls) to the kernel-mode driver, triggering an out-of-bounds write that corrupts kernel memory. By controlling the overwrite, the attacker escalates to SYSTEM/root, enabling full compromise, or alternatively crashes the driver for denial of service. No public exploit code was identified at the time of analysis, and exploitation requires local access. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Sep 30, 2026
An out-of-bounds write in the kernel-mode layer of the NVIDIA GPU Display Driver for Windows and Linux lets a local attacker who already holds low-privilege code execution corrupt kernel memory, with potential outcomes of privilege escalation, arbitrary code execution, denial of service, information disclosure and data tampering (CWE-787, CVSS 7.8). The assessed vector CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H shows low attack complexity and no user interaction, but the attack is strictly local: an attacker must first have a foothold on the host, or on a guest in a vGPU configuration, which confines it to privilege-escalation and sandbox-escape scenarios rather than remote mass exploitation. No public exploit code and no confirmed active exploitation were identified at time of analysis (no public exploit identified at time of analysis), so this is a genuine local privilege-escalation priority despite the absence of exploitation signals, and it warrants prompt patching on endpoint, workstation, and multi-tenant virtual-desktop estates.
RemediationAI
Patch status: patch available per vendor advisory. The primary action is to upgrade the NVIDIA GPU Display Driver to the corrected release specified for your driver branch and operating system in NVIDIA's security bulletin https://github.com/NVIDIA/product-security/tree/main/2026/5861 (the source data does not enumerate exact fixed version numbers, so take the version directly from the bulletin rather than assuming the newest generic download covers your branch, and validate both Windows and Linux branches plus any vGPU/Virtual GPU Manager component if it is in scope). Because exploitation requires local low-privilege access and no user interaction, treat this as a privilege-escalation and sandbox-escape hardening problem: enforce least privilege so that untrusted users and processes cannot reach the driver's kernel interfaces, apply application allowlisting to block unsigned or unexpected binaries that could act as exploit launchers (side effect: operational overhead and possible breakage of legitimate unlisted tools), and in vGPU or GPU-passthrough environments avoid exposing a shared GPU to untrusted tenants and segregate workloads by physical GPU where feasible (trade-off: reduced consolidation density and higher hardware cost). Additional compensating controls include keeping exploit-mitigation features and OS kernel protections enabled, restricting interactive logon and remote-access paths that would give an attacker the required local foothold, and monitoring driver-related crash dumps, kernel fault events and unexpected process-to-driver interaction for signs of memory-corruption attempts; these controls reduce exposure but do not eliminate the vulnerability, so driver updates remain the only complete fix.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 3 packages in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The vulnerable component is the kernel-mode portion of the NVIDIA display driver stack (the kernel driver loaded by the OS, e.g. the kernel display driver on Windows and the nvidia kernel module on Linux), which mediates privileged GPU operations such as memory allocation, mapping, MMIO access and command submission on behalf of user-mode clients. CWE-787 (out-of-bounds write) means a buffer index, length or offset derived from attacker-influenced input is not sufficiently bounds-checked, so data is written past the end of an allocated kernel buffer; because this occurs in kernel context, a successful write can corrupt kernel data structures and yield ring-0 code execution, privilege escalation from a normal user account, guest-to-host or sandbox escape in virtualized GPU setups, kernel crashes (denial of service), and disclosure or modification of data. The write is reachable through the driver's user-to-kernel entry points, which are exposed to unprivileged processes, consistent with the assessed PR:L requirement; the CPE data identifies the exposed product families as NVIDIA GeForce, NVIDIA RTX/Quadro/NVS, and NVIDIA Tesla drivers, with the tag set also naming the Virtual GPU Manager (vGPU) component, and the affected platforms as Microsoft Windows and Linux. The same memory-corruption class has historically been the dominant source of local privilege-escalation bugs in GPU drivers, which is why GPU kernel drivers are now a favored target for post-exploitation tooling that aims to disable security controls or escalate from a compromised user session.
Know when Windows 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-privilege attacker triggers a kernel-mode out-of-bounds write with no user interaction, giving full confidentiality, integrity, and availability impact via privilege escalation.
Primary rating from Vendor (nvidia).
CVSS VectorVendor: nvidia
DescriptionCVE.org
NVIDIA GPU Display Driver for Windows and Linux contains a vulnerability in the kernel mode layer where an attacker could cause an out-of-bounds write. A successful exploit of this vulnerability might lead to code execution, denial of service, escalation of privileges, information disclosure, and data tampering.
Articles & Coverage 4
Executive ViewAI
A memory-corruption flaw in the kernel-mode component of the NVIDIA GPU Display Driver (CVE-2026-47523) affects Windows and Linux hosts, and because no vendor patch is currently available, every GPU-enabled endpoint, workstation, and virtual desktop in the estate remains exposed. The risk is confined to users who already have a foothold on the machine-an unprivileged local user, or an attacker inside a vGPU guest-who could then escalate to full kernel/system control, break out of a virtualized sandbox, or crash and corrupt the host. Leadership should treat this as a contained but unmitigated privilege-escalation exposure that matters most for multi-tenant VDI, GPU-accelerated servers, and any environment where untrusted users share GPU hardware.
No vendor-released patch identified at time of analysis. Interim mitigations: restrict local low-privilege code execution and interactive access on GPU-enabled hosts to trusted users; enforce strict guest-to-host isolation and avoid mixing untrusted tenants on shared GPUs in vGPU configurations; apply least privilege and remove unnecessary local accounts; monitor for unusual kernel-mode driver activity and local privilege-escalation indicators; isolate or dedicated-host GPU workloads where operationally feasible pending a vendor fix.
Threat Intelligence Overview
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.
Memory corruption occurs when a program writes data beyond the boundaries of allocated memory regions or accesses memory in unintended ways, violating the integrity of the process's address space.
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.
Unauthenticated remote code execution in SmarterTools SmarterMail lets an attacker upload arbitrary files to any locatio
Windows Win32k fails to properly handle objects in memory, allowing local privilege escalation exploited in the wild in
A privilege escalation vulnerability (CVSS 5.5). Risk factors: actively exploited (KEV-listed), EPSS 94% exploitation pr
Windows Internet Shortcut Files (.url) contain an external control vulnerability (CVE-2025-33053, CVSS 8.8) that enables
Windows SMB contains an improper access control vulnerability (CVE-2025-33073, CVSS 8.8) enabling authenticated attacker
Twonky Server 8.5.2 on Linux and Windows allows unauthenticated access to the admin log file through a web service API b
An unrestricted file upload vulnerability exists in MiniWeb HTTP Server <= Build 300 that allows unauthenticated remote
FreeFloat FTP Server contains multiple critical design flaws that allow unauthenticated remote attackers to upload arbit
Serviio Media Server versions 1.4 through 1.8 on Windows contain an unauthenticated command injection in the /rest/actio
Twonky Server 8.5.2 uses hard-coded cryptographic keys for encrypting the administrator password. Combined with the cred
Mako Server versions 2.5 and 2.6 contain an unauthenticated OS command injection via the tutorial interface at examples/
React Native Metro Development Server binds to external interfaces by default and contains an OS command injection endpo
EUVD-2026-90102
GHSA-mj7w-whx6-f9m8
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB