CoreWeave launches six-layer zero trust framework for AI workloads
CoreWeave unveils its Zero Trust AI Framework, a six-layer model giving customers authority over access to their models, data and runtime environments.
By GennnAi··1 min read
Unveiled at GTC Berlin, the framework lets customers use their own encryption keys, authorise provider access case by case and verify infrastructure through attestation using hardware roots of trust. CoreWeave says it does not use customer prompts, training data, weights or outputs to train its own models, and encrypts its backbone with MACsec. It plans to extend NVIDIA's Confidential Computing, with HGX Rubin NVL8 coming in Q4 2026.
Source: CoreWeave, published on October 21, 2026.