VAST Data has introduced a new confidential AI capability that lets enterprises run proprietary and open AI models against sensitive data without exposing either the data or the models to infrastructure operators.
Jeff Denworth, co-founder, VAST Data, said: “VAST Data – in partnership with NVIDIA – is moving the industry forward with a comprehensive approach to verifying previously untrusted computing environments and unlocking the ability to run any model against any data, anywhere.”
VAST DataEnclave
Called VAST DataEnclave, the capability is part of the VAST DataEngine and is built on NVIDIA Confidential Computing, ensuring enterprises in sectors such as financial services, healthcare, and government can process sensitive information securely without exposing data or models.
DataEnclave creates a hardware-isolated environment in which models and data can be processed while remaining protected in CPU and GPU memory. Before releasing encrypted assets, the environment is cryptographically verified through attestation, including NVIDIA GPU attestation.
The approach also separates control of the assets involved, allowing enterprises to retain control of their data encryption keys, while model builders maintain their own model keys and weights through independent key-management systems.
VAST said DataEnclave can operate in connected or fully air-gapped environments, including deployments using the open CNCF Trustee stack or Fortanix’s Confidential AI infrastructure.
The capability also extends to AI agents through VAST AgentEngine, providing isolated execution environments and policy controls over the data, systems and tools agents can access.
“VAST Data’s integration of NVIDIA Confidential Computing delivers protection for both enterprises and model builders, providing security, identity, permissions, governance and compliance as a foundation of the agent architecture,” said Justin Boitano, vice president, Enterprise AI, NVIDIA.
