Vast Data's vice president of systems engineering, Vaughn Stewart, said at the AI Data Pipeline Forum that agentic pipelines assembled from best-of-breed components struggle to keep vector embeddings, user access controls and source data in sync, turning data sovereignty into an architectural problem for AI inference. He said roughly half of the world's data still sits in corporate data centers or colocation facilities, much of it legally unable to reach public-cloud frontier models. Vast's answer is DataEnclave, a confidential AI runtime announced in September 2026 with partners including Nvidia and Cohere that runs models in secured virtual machines, requires remote cryptographic attestation before releasing keys and logs access to a tamper-proof database; the report says Vast raised about $1 billion at a $30 billion valuation in April 2026.

Vast Data·DataEnclave·Nvidia·+1