EUVD-2026-61109
GHSA-j87r-h67h-c626
Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Network-reachable inference API, no authentication required, availability-only impact; metrics align with provided vector and CWE-770 characterization.
Primary rating from Vendor (nvidia).
CVSS VectorVendor: nvidia
Lifecycle Timeline
2DescriptionCVE.org
NVIDIA Triton Inference Server for Linux contains a vulnerability where an attacker could cause an allocation of resources without limits. A successful exploit might lead to denial of service.
Articles & Coverage 1
AnalysisAI
Unbounded resource allocation in NVIDIA Triton Inference Server for Linux allows remote, unauthenticated attackers to exhaust server-side resources and cause denial of service against all tracked versions (CPE wildcard). The CVSS 7.5 High rating reflects a fully network-accessible attack path with no privileges or user interaction required, making the exposure surface broad for any Triton deployment reachable from untrusted networks. No public exploit code has been identified at time of analysis, and exploitation has not been confirmed by CISA KEV.
Technical ContextAI
NVIDIA Triton Inference Server is an open-source, production-grade ML inference serving platform for Linux, designed to host and serve AI models at scale via HTTP, gRPC, and metrics APIs. The vulnerability is rooted in CWE-770 (Allocation of Resources Without Limits or Throttling), meaning a code path in Triton's request handling, model-loading, or session management logic fails to enforce upper bounds on a finite resource - such as memory, file descriptors, thread pool slots, or request queue depth. An attacker who sends crafted inputs along this path can drive the server's resource consumption to exhaustion, starving legitimate inference clients. The CPE string cpe:2.3:a:nvidia:triton_inference_server:*:*:*:*:*:*:*:* applies a full version wildcard, indicating the flaw is not isolated to a specific release branch and potentially spans the entire product lineage on Linux.
RemediationAI
Consult the NVIDIA product security advisory at https://github.com/NVIDIA/product-security/tree/main/2026/5865 for authoritative patch guidance and fixed version information; no exact patched release version was confirmed in the available intelligence, so the advisory is the definitive source. As an immediate compensating control, restrict network access to Triton inference endpoints - default HTTP port 8000, gRPC port 8001, and metrics port 8002 - to trusted clients only using host-based firewall rules (iptables/nftables) or network segmentation, reducing the pool of potential attackers. Deploying a rate-limiting reverse proxy or API gateway (such as nginx with limit_req or Envoy with global rate limiting) in front of Triton can throttle the throughput of resource-exhausting requests; note this reduces exploitation opportunity but does not eliminate the underlying vulnerability. Monitor Triton process memory and file descriptor consumption via the metrics endpoint or system-level tooling and set alerting thresholds to detect anomalous consumption before full exhaustion occurs.
More in Triton Inference Server
View allAuthentication bypass in NVIDIA Triton Inference Server allows unauthenticated remote attackers to reach protected funct
Path traversal in NVIDIA Triton Inference Server for Linux (CWE-22) lets remote attackers manipulate file paths to reach
Authentication bypass in NVIDIA Triton Inference Server allows remote unauthenticated attackers to circumvent access con
Integer overflow in the DALI backend of NVIDIA Triton Inference Server allows authenticated remote attackers to trigger
Out-of-bounds read in the DALI backend of NVIDIA Triton Inference Server allows authenticated remote attackers to trigge
NVIDIA Triton Inference Server for Linux contains a vulnerability where a user can set the logging location to an arbitr
NVIDIA Triton Inference Server for Linux contains a vulnerability in the tracing API, where a user can corrupt system fi
NVIDIA Triton Inference Server for Linux contains a vulnerability in shared memory APIs, where a user can cause an impro
Denial of service in NVIDIA Triton Inference Server for Linux allows remote unauthenticated attackers to exhaust server
Denial of service in NVIDIA Triton Inference Server for Linux (versions through 26.03) allows a remote attacker to crash
Denial of service in NVIDIA Triton Inference Server for Linux exposes AI inference infrastructure to remote disruption t
Denial of service in NVIDIA Triton Inference Server for Linux allows remote unauthenticated attackers to trigger uncontr
Same technique Denial Of Service
View allShare
External POC / Exploit Code
Leaving vuln.today