NVIDIA Triton Inference Server contains a flaw in its DALI backend that allows an attacker to cause uncontrolled resource consumption. The flaw can lead to denial of service by exhausting memory or CPU resources. This weakness is classified as CWE-400, indicating that the system does not adequately limit resource usage when processing input from potentially untrusted sources.

The affected product is NVIDIA Triton Inference Server. The specific vendors and product name are NVIDIA Triton Inference Server; no version information was supplied, so administrators should verify all running instances against NVIDIA's advisory for patched releases.

The CVSS score of 5.7 indicates moderate severity. Because the EPSS score is not available, the likelihood of exploitation cannot be quantified, and the vulnerability is not listed in CISA's KEV catalog, the risk is considered modest. The attack vector is not explicitly defined in the description, but the likely scenario is that a malicious client could send specially crafted requests to the DALI backend, inducing the server to consume excessive resources. Successful exploitation would result in a denial of service affecting all users of the inference service.