CUDA Toolkit nvJPEG Local Auth OOB Write via Image Dimensions
CVE-2025-23275 Published on September 24, 2025
NVIDIA CUDA Toolkit for all platforms contains a vulnerability in nvJPEG where a local authenticated user may cause a GPU out-of-bounds write by providing certain image dimensions. A successful exploit of this vulnerability may lead to denial of service and information disclosure.
Vulnerability Analysis
CVE-2025-23275 can be exploited with local system access, requires user interaction and a small amount of user privileges. This vulnerability is consided to have a high level of attack complexity. The potential impact of an exploit of this vulnerability is considered to be low. considered to have a small impact on confidentiality and integrity and availability.
Weakness Type
What is a Memory Corruption Vulnerability?
The software writes data past the end, or before the beginning, of the intended buffer. Typically, this can result in corruption of data, a crash, or code execution. The software may modify an index or perform pointer arithmetic that references a memory location that is outside of the boundaries of the buffer. A subsequent write operation then produces undefined or unexpected results.
CVE-2025-23275 has been classified to as a Memory Corruption vulnerability or weakness.
Products Associated with CVE-2025-23275
Want to know whenever a new CVE is published for NVIDIA Cuda Toolkit? stack.watch will email you.
Affected Versions
NVIDIA CUDA Toolkit:- Version All versions prior to CUDA Toolkit 13.0 is affected.
- Version All versions prior to nvJPEG 13.0.0 is affected.
Exploit Probability
EPSS (Exploit Prediction Scoring System) scores estimate the probability that a vulnerability will be exploited in the wild within the next 30 days. The percentile shows you how this score compares to all other vulnerabilities.