These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.
nvTrust
nvTrust is NVIDIA’s open-source repository of supporting software for confidential computing on NVIDIA graphics processing units (GPUs). It provides tools to implement and validate trusted computing, with a focus on attestation. Attestation is the process of cryptographically proving that a GPU and its software environment are genuine, correctly configured and untampered with before sensitive data or AI models are entrusted to them. It supports organisations that run AI training and inference on sensitive data in confidential computing environments, where data and model weights must stay protected even from the cloud or infrastructure operator.
The repository groups its tools into guest tools, used inside the confidential virtual machine, and host tools, used on the server. The guest tools include an Attestation software development kit (SDK) with application programming interfaces for adding GPU and NVSwitch attestation to applications. The SDK supports both local verification and remote verification through an NVIDIA attestation service. The verification process checks the GPU’s signed evidence, including firmware and driver measurements, against NVIDIA’s certificate chain and reference values, and confirms the device is in a trusted state before workloads run.
A Protected PCIe (PPCIE) verifier extends attestation to multi-GPU systems, where GPUs communicate over NVLink while the confidential virtual machine’s security is preserved. The host tools configure GPU confidential computing modes and provide sample scripts for launching confidential virtual machines. The original Python SDK and local GPU verifier are now deprecated in favour of a newer C++ SDK and command-line interface. Quick-start and deployment guides support adoption from proof of concept to production.
About the tool
You can click on the links to see the associated tools
Tool type(s):
Objective(s):
Purpose(s):
Country/Territory of origin:
Lifecycle stage(s):
Type of approach:
Maturity:
Usage rights:
Target users:
Stakeholder group:
Risk management stage(s):
Technology platforms:
Use Cases
Would you like to submit a use case for this tool?
If you have used this tool, we would love to know more about your experience.
Add use case
Partnership on AI
