What happened
VAST Data unveiled DataEnclave, a Confidential AI runtime inside its VAST AI Operating System. The feature lets leading AI models from Cohere, CrowdStrike, Deepgram, Factory, Fundamental, NVIDIA, TwelveLabs and others run on enterprise data without exposing customer data or model IP. DataEnclave works across data center, cloud, or edge locations, tying together storage and compute in a trusted environment. It also underlines VAST's push to combine data storage with AI compute for secure workflows.
Why it matters
The setup targets privacy and IP protections in enterprise AI. Organizations can use external AI models without moving or exposing sensitive data. Model providers get a controlled channel to access enterprise data for inference or evaluation, potentially expanding use cases. It reflects a broader move toward secure, hybrid AI setups that blend trusted infrastructure with modern models, important for regulated industries. This approach may influence where companies place AI work, mixing on‑prem and cloud deployments.
What to watch
Who adopts DataEnclave and in which industries. How the system enforces privacy and IP protection in practice (controls, audits). Any cost or performance trade-offs, and effects on data-center storage and networking demand. Future partnerships or broader model support, and whether Nvidia’s ecosystem benefits indirectly. Look for case studies, audits, or security attestations.