What is the CVSS score of CVE-2026-24191?

CVE-2026-24191 has a CVSS 3.1 base score of 7.8 (High). CVSS vector: CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:H. EPSS exploitation probability: 0.0%.

Which versions of NVIDIA Display Driver are affected by CVE-2026-24191?

NVIDIA Display Driver for Windows (GeForce, Quadro, Tesla, vGPU) contains a local privilege escalation vulnerability (CVE-2026-24191, CVSS 7.8) enabling low-privileged users to escalate to system…

How to fix or mitigate CVE-2026-24191?

Within 24 hours: Inventory all systems running NVIDIA Display Driver; prioritize Tesla and vGPU deployments in data center environments. Within 7 days: Implement strict local user access controls to minimize low-privileged accounts; disable unnecessary GPU acceleration where operations permit; audit current user access on GPU-enabled systems. Within 30 days: Monitor NVIDIA security advisories for patch availability; document driver version baseline and prepare controlled update procedures; evaluate feasibility of permanent GPU access restrictions as compensating control.

NVIDIA Display Driver CVE-2026-24191

EUVD-2026-31931 HIGH

Time-of-check Time-of-use (TOCTOU) Race Condition (CWE-367)

2026-05-26 nvidia

GHSA-2439-7mhv-329q

Nvidia RCE Information Disclosure Microsoft Denial Of Service

7.8

CVSS 3.1 · Vendor: nvidia

Severity by source

Vendor (nvidia) PRIMARY

7.8 HIGH

AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:H

Primary rating from Vendor (nvidia) · only source for this CVE.

CVSS VectorVendor: nvidia

CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:H

Attack Vector

Local

Attack Complexity

High

Privileges Required

Low

User Interaction

None

Scope

Changed

Confidentiality

High

Integrity

High

Availability

High

Lifecycle Timeline

Analysis Generated

Jun 08, 2026 - 10:06 vuln.today

CVE Published

May 26, 2026 - 17:23 nvd

HIGH 7.8

DescriptionCVE.org

NVIDIA Display Driver for Windows contains a vulnerability where an attacker could cause a time-of-check time-of-use issue. A successful exploit of this vulnerability might lead to denial of service, escalation of privileges, information disclosure, data tampering, and code execution.

AnalysisAI

Local privilege escalation in NVIDIA Display Driver for Windows (GeForce, RTX/Quadro/NVS, Tesla, and vGPU guest/manager components) stems from a time-of-check time-of-use (TOCTOU) race condition that can be abused by a low-privileged local user. Successful exploitation may yield code execution, privilege escalation, denial of service, information disclosure, or data tampering with scope change beyond the driver's security boundary. No public exploit identified at time of analysis; EPSS is 0.01% and SSVC exploitation status is 'none', so the practical risk is contingent on local access and winning a high-complexity race.

Technical ContextAI

The flaw is a CWE-367 Time-of-Check Time-of-Use race condition within the NVIDIA kernel-mode display driver stack on Windows. TOCTOU bugs occur when a privileged component validates a resource (a handle, path, buffer, or permission state) and then operates on it later, allowing an unprivileged caller to swap or mutate the resource between the check and the use. In a kernel driver context - confirmed by the CPE coverage spanning consumer GeForce, professional RTX/Quadro/NVS, datacenter Tesla, and the NVIDIA Virtual GPU Manager and guest driver - winning the race lets a user-mode caller cause the kernel to act on attacker-controlled data, which is why the CVSS scope is Changed (S:C) and all three CIA impacts are High. The same primitive class historically appears in IOCTL handlers where user pointers or shared memory regions are re-read after validation.

RemediationAI

Apply the vendor-released patches per NVIDIA security bulletin at https://nvidia.custhelp.com/app/answers/detail/a_id/5821: upgrade GeForce, RTX/Quadro/NVS, and Tesla Windows drivers to 596.36 or later on the current branch, or to 582.53 / 539.72 on the respective maintenance branches; in virtualized deployments upgrade NVIDIA Virtual GPU Manager and guest drivers to versions beyond vGPU 16.13 (guest 539.64), vGPU 19.4 (manager/guest 582.16), and vGPU 20.0 (manager 595.94 / guest 595.97). Until patching is complete, compensating controls include restricting interactive and remote-code-execution access on systems with the vulnerable driver (since exploitation requires a local low-privileged account), tightening application allowlisting to block unauthorized binaries that could spawn race-window threads, and in vGPU/VDI environments isolating untrusted tenants onto patched hosts first - these reduce exposure but do not eliminate it for any user who can already run code locally. Disabling the GPU driver is generally not a viable workaround for systems that depend on hardware acceleration; downgrading to an older branch is not an alternative because the older branches are themselves in scope.