CVE-2026-65179: NVIDIA NeMo unsafe pickle deserialization in TabularTokenizer

CVE-2026-65179 · Severity: high · CVSS 8.8 · Published 2026-09-22

NVIDIA NeMo is a framework for building speech and language AI models. The TabularTokenizer component unsafely deserializes untrusted pickle files, allowing an attacker to execute arbitrary code, modify data, disrupt service, or steal information if a malicious .pkl file is processed.

Technologies: Nvidia Nemo. Vendors: Nvidia.

Quick answers

Is CVE-2026-65179 exploited in the wild?
No. Junglewise Threat Intelligence has no report of CVE-2026-65179 being exploited in the wild.
How severe is CVE-2026-65179?
High severity, with a CVSS score of 8.8, published 22 September 2026.
Which products and versions are affected?
Nvidia Nemo <UNKNOWN>.
Is there a fix for CVE-2026-65179?
No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.
How is CVE-2026-65179 attacked?
It is attacked over the network, in a software library.
Am I affected by CVE-2026-65179?
Check whether you run Nvidia Nemo at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

Technical details

The vulnerability exists in the TabularTokenizer class, which calls pickle.load() on user-supplied .pkl files without validation or signing. An attacker can craft a malicious pickle file that executes arbitrary Python code during deserialization. Exploitation requires the victim to process an attacker-controlled .pkl file, leading to remote code execution with the privileges of the NeMo process.

Affected products

  • NVIDIA NeMo <UNKNOWN>

Timeline

  • 2026-09-22: disclosed

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.