External risk intelligence

NVIDIA BioNeMo Untrusted Deserialization Vulnerability

CVE advisorySeverity: CRITICAL (CVSS 9.8)

CVE-2026-24164

NVIDIA BioNeMo is a framework for generative AI in drug discovery. While it involves data processing that could be exposed via APIs or research portals, it is typically deployed in internal research, lab, or development environments rather than being a public-facing edge service or internet-exposed appliance by default.

Deserialization

Nvidia Bionemo Framework

before 2026-01-21

Halo Surface Signal: 3 out of 5 — possibly public-facing.

External exposure likelihood

Horizon Alert

Summary of the vulnerability and why it matters

NVIDIA BioNeMo, a framework used for generative AI in drug discovery, has a vulnerability related to processing untrusted data, which could potentially lead to unauthorized code execution, service disruption, or data compromise. The primary concern at this time is to confirm if this specific technology is in use within our environment.

Untrusted data processing in AI framework.
Potential for code execution or data compromise.
Confirm relevance and exposure in our environment.

Attack Path

How an attacker could exploit the issue

An attacker could reach the NVIDIA BioNeMo framework over the network without needing any special access or user interaction. By sending specially crafted data, they could trigger a flaw in how the framework handles untrusted information. This could allow them to execute their own code on the system, disrupt its operation, steal sensitive data, or alter existing information.

Network access required.
Untrusted data triggers deserialization.
Leads to code execution or data compromise.

Live Threat

Current exploitation, exposure, and threat context

NVIDIA BioNeMo, when accessible over a network and without requiring user interaction or privileges, could allow an attacker to deserialize untrusted data. This might lead to unauthorized code execution, disruption of service, exposure of information, or alteration of data within the framework.

Framework data and services.
Deserializing untrusted input.
Code execution or data tampering.

Operational Fix

Recommended remediation, mitigation, and detection steps

Owners of the NVIDIA BioNeMo framework are responsible for addressing this critical deserialization vulnerability. The first practical step is to identify all instances of the framework within the environment, assess their accessibility and business criticality, and then determine the accountable owner for remediation planning.

Application or platform teams own this issue.
Verify framework reachability and criticality.
Plan and coordinate remediation activities.

Supplementary metadata

CVSS vector

Validate whether this threat affects your internet-facing exposure.

Halo Threat Intelligence helps prioritize remediation with Halo Surface Signal and H/A/L/O context. Start exposure validation with a free external attack surface trial.

Free EASM Trial See How Signal Works

Frequently asked questions

What is the NVIDIA BioNeMo framework?

NVIDIA BioNeMo is a specialized software framework designed to accelerate generative AI development within the fields of biology and drug discovery. It provides researchers and data scientists with tools to build, train, and deploy large-scale AI models for tasks like protein structure prediction and molecular design.

What does deserialization of untrusted data mean for CVE-2026-24164?

This vulnerability falls under the weakness class of deserialization of untrusted data (CWE-502). In plain English, the framework incorrectly trusts and processes incoming data structures without proper validation. An attacker can craft malicious data that the framework converts into objects, which can then trick the system into executing unauthorized commands, crashing the service, or accessing sensitive data.

How does an attacker trigger this vulnerability?

An attacker triggers this flaw by sending specially crafted data over a network to the framework. The system processes this input automatically, meaning no user interaction or elevated privileges are required to initiate the attack. Note that internal data processing that is not reachable from the network or other untrusted sources does not serve as a vector for this specific bug.

Is my instance of NVIDIA BioNeMo at risk?

According to Halo Surface Signal, while NVIDIA BioNeMo is often deployed in internal research or development labs, any instance reachable via a network is a potential concern. You should assess whether your specific implementation is exposed to external traffic or accessible by untrusted users, as these configurations are the most critical to secure.

What should I do if I run NVIDIA BioNeMo?

Your first step is to locate all instances of the framework within your infrastructure. Once identified, evaluate their network accessibility and current version to confirm if they are affected. Coordinate with the relevant platform or application owners to prioritize these systems for remediation based on their business criticality and exposure.

References

Written by Nicholas Merritt

Nicholas Merritt is VP of Security Solutions at Halo Security. With more than 25 years across application security, vulnerability prioritization, and offensive security, he helps teams translate threat data into practical exposure validation and remediation focus. He authors Halo Threat Intelligence CVE advisories using Halo Surface Signal and H/A/L/O context to prioritize internet-facing risk.