Six layers of customer-controlled security, from encryption keys to runtime, designed to keep CoreWeave out of the customer's trust chain

BERLIN--(BUSINESS WIRE)-- CoreWeave, Inc. (Nasdaq: CRWV), The Essential Cloud for AI®, today announced the CoreWeave Zero Trust AI Framework™, a customer-controlled security model that gives customers authority over access to their models, data, and runtime environments. Unveiled at NVIDIA GTC Berlin, the framework lets customers bring the security systems they already operate to CoreWeave Cloud rather than rebuilding around a provider-defined security model.

Agentic AI has changed what enterprises need from cloud security. As agents retrieve context, call tools, and trigger actions across other systems, sensitive data no longer sits behind a single application boundary. It moves through prompts, embeddings, weights, traces, outputs, and runtime systems. Running those workloads on a cloud has typically meant extending trust to the provider, which may administer encryption keys, hold privileged access to servers, manage the execution environment, or operate its own AI models that may learn from customer interactions to improve provider-owned AI systems.

"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. "Our customers are running their most valuable models against their most sensitive data. They should not have to take our word for how we handle either. The Zero Trust AI Framework replaces assurances with architecture, so the controls that matter stay in the customer's hands."

The framework brings together six layers of protection:

• Zero Key Access. With Remote Key Encryption, customers use keys from the key infrastructure they already operate, whether a software key manager, a hardware security module, or a KMS in another cloud. Customers retain control of the keys protecting their data at rest without transferring key authority to CoreWeave, and without rebuilding key management workflows, policy controls, or governance processes to run sensitive training and inference workloads.
• Zero Node Access. Support Access Management makes provider access an explicit customer decision. Customers authorize access for a defined operational need rather than treating it as always available, and access ends when the authorized activity does.
• Zero Trust Infrastructure. CoreWeave attestation services allow the infrastructure executing a workload to be verified using hardware roots of trust. Attestation can also drive automated trust decisions, so a workload, key-management system, or orchestration layer can require valid evidence before releasing secrets, decrypting data, or allowing execution to proceed.
• Zero In-Use Access. CoreWeave intends to extend NVIDIA’s third-generation Confidential Computing across CPUs, GPUs, and the NVIDIA NVLink fabric, including on NVIDIA HGX Rubin NVL8. HGX Rubin NVL8 is coming to CoreWeave Cloud in Q4 2026 and is built for agentic AI, so customers can run agents at scale while keeping models, data, and code protected while in use. CoreWeave plans to begin with select bare metal environments and extend to additional deployment models, including Kubernetes-based AI workloads. Additionally, CoreWeave intends to support the NVIDIA OpenShell Reference Architecture, which combines Rubin GPUs and Vera CPUs running the NVIDIA OpenShell secure runtime.
• Zero Model Data. As an AI cloud, not a competing model provider, CoreWeave does not use customer prompts, training data, model weights, or outputs to train or improve its own models.
• Zero Trust Network Encryption. CoreWeave uses IEEE 802.1AE MACsec to encrypt data moving across its backbone, making protected transport part of the network architecture rather than a burden on every application team.

Together, these six controls minimize CoreWeave's role in the customer trust chain. Each one addresses a point where a provider would otherwise sit inside the customer's security boundary. Rather than depending on policy statements or operational assurances, customers can rely on architecture that limits access, verifies infrastructure, separates responsibilities, and preserves ownership.

“As enterprises move increasingly sensitive and regulated data into AI workflows, zero trust must extend beyond identity and network security to the entire AI infrastructure stack,’ said Dave McCarthy, group vice president, Cloud and Datacenters, Enterprise Infrastructure, IDC. “CoreWeave’s Zero Trust AI Framework addresses this requirement by giving customers control over encryption keys and administrative access while introducing verifiable infrastructure, confidential execution, protected data movement, and customer data controls that do not rely on using customer data to train or improve the cloud provider’s own AI models. This shifts trust from provider promises to technical controls, a critical step for enterprises deploying proprietary AI at scale.”

To learn more, visit the CoreWeave Trust Center.

Built on Proven AI Infrastructure

CoreWeave consistently delivers industry-leading performance, demonstrated by record-breaking MLPerf benchmark results in inference and training, and its position as the only AI cloud to earn the SemiAnalysis ClusterMAX™ Platinum ranking three times in a row.

About CoreWeave

CoreWeave is The Essential Cloud for AI®. Built for pioneers by pioneers, CoreWeave delivers a fully integrated platform of technology and teams that enables innovators to move at the pace of innovation, building and scaling AI with confidence. Trusted by 9 of the 10 leading foundation model providers, CoreWeave serves as a force multiplier by combining superior infrastructure performance with deep technical expertise to accelerate human discovery. Established in 2017, CoreWeave completed its public listing on Nasdaq (CRWV) in March 2025.

View source version on businesswire.com: https://www.businesswire.com/news/home/20261021867955/en/

Media Contact

[email protected]

Source: CoreWeave, Inc.