Building on our NVIDIA Inception membership, Trust3 AI aims to provide agent discovery, purposeful access control, and full-fidelity observability for generative AI applications built with NVIDIA NeMo.
san francisco, June 30, 2026 /PRNewswire/ — Trust3 AI, the enterprise agent control plane, today announced its intent to extend its unified trust layer to NVIDIA NeMo, an end-to-end platform for developing custom-generated AI models and agents. As a member of the NVIDIA Inception program, Trust3 AI is working on integrations designed to help enterprises manage agents and applications running on NeMo without standing up a separate set of controls.
When an organization moves from pilot to production, it is unlikely that a model will fall into gaps in its capabilities. It’s governance. Teams struggle to answer basic operational questions, such as which agents are running, what data and tools they have access to, what they actually did, and can they prove it to auditors. Trust3 AI’s planned NeMo integration aims to fill that gap at the point of action.
We aim to enable teams to build on NeMo once the integration is available.
- Agent discovery and inventory across NeMo-based deployments. This also includes unregistered agents (Shadow AI).
- Purpose-Based Access Control (PBAC). Auto-expiring grants tie agent access to a declared purpose rather than a static role.
- Apply policies at the moment of action, before the tool launches and before any data moves with agent ID, declared purpose, and per-request data lineage.
- Full-fidelity observability, capturing prompts, retrievals, tool calls, and data accesses allows you to replay and verify agent behavior.
- Audit evidence mapped to frameworks such as EU AI legislation and the NIST AI Risk Management Framework.
This approach is intended to complement NeMo’s own controls, including NeMo Guardrails, by adding an independent, vendor-neutral layer of trust below, rather than beside, the agent.
“As enterprises move from experimentation to production with NVIDIA NeMo, the gap they face is not functionality, but security and governance,” said Neeraj Sabharwal, co-founder of Trust3 AI. “Our goal is to place a trust layer directly below NeMo-based agents so that security and compliance teams can see every agent, secure and control every action, and attest to it without slowing down the AI shipping team.” Neeraj Sabharwal, Co-Founder, Trust3 AI
Trust3 AI was founded by the engineers behind Apache Ranger and Apache Atlas, the open source access control and metadata standards used by regulated enterprises today. The company’s One Control Plane architecture and Unified Trust Layer are built to manage any agent on any framework, across any cloud and data source.
About Trust3 AI
Trust3 AI is an agent control plane for enterprises. Founded by Balaji Ganesan, Don Bosco Durai, and Neeraj Sabharwal, the company provides AI-powered governance of data, AI, and access intelligence through a One Control Plane architecture that leverages a unified trust layer built on AI-native metadata and context. The platform enables enterprises to discover any agent, monitor every decision, and secure every action across any framework, cloud, or data source. Trust3 AI is a member of the NVIDIA Inception program and Snowflake Startup Accelerator. Learn more at trust3.ai or request a demo at trust3.ai/demo.
Forward-looking statements
This press release contains forward-looking statements regarding Trust3 AI’s product plans, including its plans to integrate with NVIDIA NeMo and the expected features, benefits, and timing of that integration.
For more information, please visit https://trust3.ai.
media contact
Trust3 AI
Ibby Rahmani, Head of Marketing
(510) 413-7300
ibby@trust3.ai
Logo – https://mma.prnewswire.com/media/2984063/Trust3_AI_Logo.jpg
