CoreWeave has launched the CoreWeave Zero Trust AI Framework, a customer-controlled security model covering access to models, data and runtime environments, the company said on 21 October 2026.
Unveiled at NVIDIA GTC Berlin, the framework lets customers bring the security systems they already run to CoreWeave Cloud rather than adopt a provider-defined model. CoreWeave said running AI workloads in the cloud has typically meant trusting a provider that may administer encryption keys or hold privileged server access.
The framework has six layers of protection. They include Zero Key Access, which lets customers use keys from their own software key manager, hardware security module or another cloud's KMS, and Zero Node Access, which makes provider access to servers an explicit customer decision.
"Security for AI workloads cannot rest on a provider's promises about its own behavior," said Chen Goldberg, executive vice president of product and engineering at CoreWeave. "The Zero Trust AI Framework replaces assurances with architecture, so the controls that matter stay in the customer's hands."
The release did not give pricing or say whether every layer is generally available today.