NVIDIA Display Driver CVE-2026-47580

Requires local, already-authenticated low-privileged access (PR:L) to a Windows system running an affected NVIDIA GPU Display Driver (GeForce, RTX/Quadro/NVS, or Tesla) - there is no remote or network vector. No user interaction is needed (UI:N) and attack complexity is low (AC:L). Realistic impact is amplified in multi-user, VDI, or cloud-GPU environments where the scope-change (S:C) lets one tenant/user reach another's confidential data; on a single-user workstation the practical exposure is narrower. Limiting factors: attacker must first obtain local code execution as a standard user, and hosts patched per NVIDIA bulletin 5861 are not affected.

This is a genuine but not urgent priority - a local privilege/authorization flaw, not a remotely exploitable one. The CVSS 3.1 vector (AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N, 7.3) is internally consistent with the description: an already-authenticated low-privilege local user (PR:L) on a Windows host with the NVIDIA driver can trigger a missing-authorization condition in the kernel module. The scope change (S:C) plus C:H is what drives the score up - the flaw lets a user reach data across a trust boundary (e.g. another user's/process's data), yielding high confidentiality impact, while integrity impact is only Low (I:L) and there is no availability impact (A:N). No EPSS score, no CISA KEV listing, and no public POC were provided, so there is no signal of active or imminent exploitation; SSVC inputs (exploitation status, automatable, technical impact) were not supplied. Real-world risk is therefore moderate and gated by the local-access requirement - this is a post-compromise or malicious-insider/multi-tenant escalation issue rather than a mass-exploitation candidate. Note two data-quality conflicts: the 'Microsoft' vendor tag is incorrect (this is an NVIDIA product), and the 'Authentication Bypass' tag is loose shorthand for the actual CWE-862 missing-authorization weakness - it does not imply a network auth bypass.

A low-privileged user (or malicious process) already logged onto a Windows workstation or a multi-tenant/VDI host with an affected NVIDIA GPU driver issues crafted requests to the GPU kernel module that lacks a proper authorization check. Because the scope changes, the user reads confidential data belonging to another user or system context and tampers with some data, without needing admin rights or user interaction. No public exploit code has been identified at time of analysis.