Stained Glass Transform for NVIDIA Nemotron 3
The fastest way to put your most sensitive data to work with an open model is to start with SGT early access.
In an agent workflow, sensitive data is everywhere
In agentic and long-running workflows, sensitive data leaves the data owner's trust zone on every LLM call. The hard reality of enterprise AI is that sensitive data often leaks into application layer surfaces and persists across the serving stack: inference logs, request caches, GPU memory, observability tooling. Stained Glass Transform™ keeps agents working on protected representations, so raw data is never there to expose.
Bring your sensitive data
SGT for Nemotron 3 is rolling out to a limited cohort, and admission is by application. Tell us what you're building and where your sensitive agent's LLM runs today.
What to expect in early access
A guided path from your first agentic workflow's definition to protected inference on the capacity you choose.
Enterprises are looking to scale agentic AI across their environments, including workflows that depend on proprietary data. With NVIDIA Nemotron and Protopia's Stained Glass Transform, they can extend open-model deployments to more of those workloads while using AI factory infrastructure more efficiently.
Open weights and open recipes are how enterprises take control of their AI deployments. What has been missing was a way to use the most sensitive data with these models without exposing it to the multi-tenant AI Factory, and without giving up the ROI. SGT for Nemotron 3 closes that gap.
Agents run within governed OpenShell workspaces, while SafeCLAW enforces policy-based routing of sensitive interactions to sovereign inference endpoints aligned with the Stained Glass Transform Proxy. HPE AI Services ensure organizations maintain a consistent user experience while ensuring end-to-end protection of sensitive data by anchoring access control at the inference layer.
Read the full Nemotron 3 SGT post
The joint Protopia AI and NVIDIA architecture for running Nemotron 3 on sensitive data, in depth.