• News
  • North America

Rafay and Fortanix Launch Confidential AI for Kubernetes

Joint service extends confidential computing to Kubernetes workloads

11/19/2026
•Ali Abounasr El Alaoui
Back to News

Rafay Systems and Fortanix launched a joint confidential AI service for Kubernetes workloads. The service lets cloud operators and enterprises run AI workloads inside hardware-isolated trusted execution environments. Rafay's platform now orchestrates workloads built on CNCF Confidential Containers with Fortanix Confidential AI.


How the service works

Each Kubernetes pod runs in a lightweight confidential virtual machine built on Kata Containers. Memory is encrypted by the CPU, and supported GPUs sit inside the trust boundary. Fortanix attests the environment and releases encryption keys only after checks pass.

Target users

The companies target cloud providers, sovereign AI clouds, and regulated enterprises. Financial services, healthcare, and public sector teams can protect customer, patient, and financial data during processing. Model providers can also deploy proprietary models on third-party GPU infrastructure with reduced risk.

Rafay's role

Rafay turns the infrastructure into a governed, multi-tenant service. Operators onboard GPU infrastructure once and package confidential workloads as self-service SKUs. End users launch workloads from Rafay's Developer Hub without administrator provisioning.


Rafay support for Confidential Containers with Fortanix Confidential AI is available now. The offering joins the companies' existing confidential virtual machine integration. Fortanix announced Fortanix Confidential AI for CoCo on October 19, 2026.