The new DataEnclave service builds on NVIDIA’s confidential computing technology, extending VAST’s AI Operating System and DataEngine with a hardware‑isolated runtime and cryptographic attestation that secures both CPU and GPU memory during model execution.
In practice, encrypted models and data are released only after the enclave’s policy and environment are verified; customers retain control of their own data‑encryption keys while model providers keep the keys and weights, preventing administrators or cloud operators from accessing either while workloads run.
VAST positions DataEnclave for regulated sectors such as financial services, healthcare and government, where data‑privacy rules often prohibit moving information to external AI services and model owners are reluctant to place their software in untrusted environments.
The platform also supports sovereign and air‑gapped deployments, including installations that rely on open‑source attestation services or partner infrastructure, enabling organisations to keep AI compute on‑premises or in trusted clouds while maintaining full data and model isolation.
The launch is backed by a roster of technology partners, including NVIDIA, CrowdStrike, Cisco, Supermicro, Cohere and Australia‑based Sharon AI. Sharon AI CEO James Manning said the solution will help Australian and Asia‑Pacific customers run AI at full speed without compromising data sovereignty, allowing on‑shore hosting of frontier models in attested environments.
VAST founder and CEO Renen Hallak said the product reflects a broader shift toward treating AI models as operating‑system resources that must be governed and secured alongside data. Co‑founder Jeff Denworth added that model weights are becoming “the most valuable intellectual property in the world,” and that protecting them is essential as enterprises adopt AI more broadly.
NVIDIA’s Vice President of Enterprise AI, Justin Boitano, said the partnership provides a security framework that safeguards both enterprise data and model‑owner IP, delivering identity, permissions, governance and compliance as a foundation for future AI agent architectures.