Building a Trusted Foundation for Confidential AI
Cisco, Thursday, September 24th, 2026
Cisco and VAST describe a confidential AI architecture that processes sensitive data and proprietary models without exposing either.
Cisco outlines the challenge of protecting AI data while it is being processed, not just stored or in transit, for organizations like banks, healthcare providers, and government agencies that can't send sensitive data to external AI services.
VAST DataEnclave, a confidential AI capability within the VAST DataEngine, provides a secure runtime and attestation service that runs AI workloads inside hardware-isolated environments, letting models come to where sensitive data already lives instead of moving data to an external model.
Cryptographic attestation checks run before a workload begins, spanning compute, networking, data, security, attestation, and observability so neither the data nor the model provider's IP is exposed to the infrastructure host.