What is the CVSS score of CVE-2026-24158?

CVE-2026-24158 has a CVSS 3.1 base score of 7.5 (High). CVSS vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H. EPSS exploitation probability: 0.0%.

Which versions of Nvidia are affected by CVE-2026-24158?

NVIDIA Triton Inference Server deployments are vulnerable to remote denial of service attacks through malicious compressed payloads, with no patch currently available.

How to fix or mitigate CVE-2026-24158?

Within 24 hours: Inventory all Triton Inference Server instances and assess exposure (internet-facing vs. internal). Within 7 days: Implement network segmentation to restrict HTTP endpoint access and deploy WAF rules blocking large compressed payloads to Triton endpoints. Within 30 days: Monitor NVIDIA security advisories for patch availability and establish a testing/deployment plan; evaluate alternative inference server solutions if Triton remains unpatched.

Nvidia CVE-2026-24158

EUVD-2026-15013 HIGH

Memory Allocation with Excessive Size Value (CWE-789)

2026-03-24 nvidia

GHSA-272x-gpf6-6c9f

Denial Of Service Nvidia

7.5

CVSS 3.1 · NVD

Severity by source

NVD PRIMARY

7.5 HIGH

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

Primary rating from NVD · only source for this CVE.

CVSS VectorNVD

Attack Vector

Network

Attack Complexity

Low

Privileges Required

None

User Interaction

None

Scope

Unchanged

Confidentiality

None

Integrity

None

Availability

High

Lifecycle Timeline

EUVD ID Assigned

Mar 24, 2026 - 20:31 euvd

EUVD-2026-15013

Analysis Generated

Mar 24, 2026 - 20:31 vuln.today

CVE Published

Mar 24, 2026 - 20:26 nvd

HIGH 7.5

DescriptionCVE.org

NVIDIA Triton Inference Server contains a vulnerability in the HTTP endpoint where an attacker may cause a denial of service by providing a large compressed payload. A successful exploit of this vulnerability may lead to denial of service.

AnalysisAI

NVIDIA Triton Inference Server contains a denial of service vulnerability in its HTTP endpoint that can be exploited by sending large compressed payloads. The vulnerability has a CVSS score of 7.5 (High) and is exploitable remotely without authentication or user interaction. There is no evidence of active exploitation (not in CISA KEV), and no public proof-of-concept has been identified at this time.

Technical ContextAI

NVIDIA Triton Inference Server is a model serving platform for deploying AI/ML models in production environments. This vulnerability (CWE-789: Memory Allocation with Excessive Size Value) affects the HTTP endpoint's handling of compressed data, where the server fails to properly validate or limit the size of decompressed payloads. When processing compressed data, systems must guard against 'zip bomb' or compression ratio attacks where small compressed inputs expand to massive decompressed outputs, exhausting memory or CPU resources. The affected product is identified by CPE cpe:2.3:a:nvidia:triton_inference_server:*:*:*:*:*:*:*:*, indicating the vulnerability spans multiple versions of the inference server's HTTP API implementation.

RemediationAI

Consult the official NVIDIA security advisory at https://nvidia.custhelp.com/app/answers/detail/a_id/5790 for patch availability and upgrade instructions to a fixed version of Triton Inference Server. As an immediate mitigation, implement rate limiting and payload size restrictions at the reverse proxy or load balancer level in front of Triton Inference Server instances to prevent large compressed payloads from reaching the vulnerable HTTP endpoint. Additional hardening measures include restricting network access to the HTTP endpoint to trusted IP ranges only, deploying web application firewalls with decompression bomb detection capabilities, and monitoring resource utilization for anomalous spikes that may indicate exploitation attempts. If the HTTP endpoint is not required for your deployment architecture, disable it in favor of gRPC or other protocols if supported.