Skip to main content

CVE-2026-100968NVIDIA Megatron-LM deserialization of untrusted data

Schedule itHigh7.8EPSS not scored yet

CVE-2026-100968 is a deserialization of untrusted data in NVIDIA Megatron-LM, rated high (7.8 on CVSS 3.1).

Why this priority: High severity, with no exploitation signal yet.

Published · Updated · NVD status: Awaiting Analysis

What it is

Serialized objects from outside are turned back into live objects, which a crafted payload can abuse to run code.

How it is reached

An attacker can exploit it by someone who can already run code or sign in on the machine, with an ordinary user account, and nobody has to click anything.

What an attacker gains

A successful attack can read the data it protects, change data or behaviour and take it offline. In practice, this kind of flaw lets an attacker turn crafted data into code execution inside the receiving service.

Is it being exploited?

There is no public report of exploitation yet.

What to do

  1. No fixed version is listed yet. Find every place NVIDIA Megatron-LM runs, watch the references for a fix, and plan a replacement if it is no longer supported.
  2. Limit who can run code or sign in on the hosts that run it.

Description from NVD

NVIDIA Megatron-LM contains a vulnerability in dataset loading where an attacker could cause deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, information disclosure, and data tampering.

What this means for your architecture.

A CVE names a product. Whether it matters depends on where that product sits in your systems: what reaches it, which trust boundary it guards and what it can reach once compromised. Veriom maps your components, their relationships and your trust boundaries from your code, cloud and delivery evidence, so you can see where a component like this sits. The radar itself does not look at your systems.

Components it touches

  • APIs, queues and caches that accept serialized objects. The flaw sits between a message producer and the service that deserializes it.

Trust boundaries in play

  • Host boundary. The attacker needs a foothold on the machine first; the flaw turns a small foothold into a bigger one.
  • Between a message producer and the service that deserializes it. An attacker who crosses it can turn crafted data into code execution inside the receiving service.

Questions to answer about your own systems

  • Where does NVIDIA Megatron-LM run in your systems, and at which version?
  • Which of those hosts also run code or users you do not fully trust?
  • What can it reach once compromised: which data stores, secrets and services?

Sources

  • CVE records and CVSS scores from the NVD CVE API. This product uses the NVD API but is not endorsed or certified by the NVD.
  • Exploitation status from CISA’s Known Exploited Vulnerabilities catalog.
  • Exploit likelihood from FIRST’s Exploit Prediction Scoring System (EPSS), refreshed daily.
  • Explanations are written by Veriom from fixed templates over each record’s data, not by a person reviewing each CVE. Descriptions are quoted from the records as published.