Critical Security Vulnerabilities in NVIDIA GPU Drivers: Privilege Escalation and DoS Risks Patched

NVIDIA has disclosed several critical security vulnerabilities affecting its GPU display drivers and virtual GPU (vGPU) software. Security updates have been released to address these issues, and administrators are encouraged to apply them promptly.

Linux GPU Display Driver Vulnerability

A significant vulnerability (CVE-2025-23244) has been discovered in NVIDIA’s Linux GPU display drivers that could allow privilege escalation. According to the Common Vulnerability Scoring System (CVSS v3.1), this security flaw has received a base score of 7.8, categorizing it as “High” severity.

This vulnerability potentially allows local attackers to gain elevated permissions on affected systems, posing serious security risks for organizations using NVIDIA graphics hardware on Linux environments.

vGPU Software Vulnerabilities

Two additional vulnerabilities have been identified in NVIDIA’s virtual GPU software:

Global Resource Access Vulnerability

The first issue (CVE-2025-23245) affects the Virtual GPU Manager and could allow unauthorized access to global resources from virtual machines. This creates potential security boundaries breaches between virtualized environments.

Resource Consumption Vulnerability

The second vulnerability (CVE-2025-23246) involves uncontrolled resource consumption in the Virtual GPU Manager, which could lead to denial-of-service conditions in both Windows and Linux environments.

Both vGPU vulnerabilities have been assigned a CVSS base score of 5.5, placing them in the “Medium” severity category.

Affected Products and Available Updates

NVIDIA has released security updates for the following products:

GPU Display Drivers:

• GeForce
• NVIDIA RTX
• Quadro
• NVS
• Tesla

Virtual GPU Manager Updates for:

• Citrix Hypervisor
• VMware vSphere
• Red Hat Enterprise Linux KVM
• Ubuntu
• Microsoft Azure
• Microsoft Windows Server

Recommended Actions

IT administrators should immediately apply the appropriate security updates to all affected systems. Organizations with large NVIDIA deployments should prioritize patching based on:

1. Systems with direct internet exposure
2. Critical production environments
3. Systems processing sensitive data
4. Development and testing environments

Mitigation Strategies

While applying patches, consider implementing these additional security measures:

• Restrict physical and remote access to affected systems
• Monitor systems for unusual activity
• Ensure proper network segmentation
• Follow the principle of least privilege for all user accounts