Rafay Systems and Fortanix launch confidential AI service

Rafay Systems and Fortanix Inc. have introduced a joint solution that enables neoclouds, sovereign AI clouds, and enterprises to offer confidential AI as a service. The approach keeps sensitive data and proprietary models encrypted while AI workloads run on infrastructure operated by others.

Workloads in financial services and healthcare often process customer records, patient information, and financial transactions. These workloads frequently run on GPU infrastructure operated by a third party, and encryption that protects data at rest and in transit does not cover data while it is processed. Research from Fortanix and Atomik Research found that 93% of AI leaders understand this gap, yet only half can fully verify protection during inference.

CNCF Confidential Containers applies confidential computing to the Kubernetes pod, running each pod in its own lightweight confidential virtual machine with CPU-encrypted memory and, on supported hardware, GPUs inside the trust boundary. Fortanix Confidential Computing Manager performs composite attestation across CPU and GPU trusted execution environments, firmware, and software before Fortanix Data Security Manager releases encryption keys. Failed attestation leaves keys protected and prevents the workload from accessing sensitive data.

The Rafay Platform lets operators deliver this infrastructure to their own customers. Rafay blueprints install and synchronize the Fortanix CoCo add-ons across confidential clusters, while operators onboard GPU infrastructure once and assign each customer an isolated tenant. Administrators package confidential workloads as SKUs in Rafay SKU Studio, and end users launch them from the Rafay Developer Hub without filing a support ticket. Governance covers policy, access controls, and quotas across tenants and clusters.

“Fortanix verifies the environment through cryptographic attestation and controls the secure key release. Rafay turns that into a governed, multi-tenant service, so a neocloud or sovereign cloud can offer confidential AI as a SKU, and its customers can run sensitive workloads on infrastructure they don't operate.” said Mohan Atreya, chief product officer at Rafay.

“Fortanix Confidential AI releases keys only to attested workloads, so model owners, enterprises, and operators, each keep control of what they own, every time it runs.” said Anuj Jaiswal, Chief Product Officer at Fortanix. Rafay support for CoCo-based deployments with Fortanix Confidential AI is available now, alongside the existing confidential VM integration. Fortanix announced Fortanix Confidential AI for CoCo on Oct. 19, 2026.