Rafay Systems and Fortanix Extend Confidential AI to Kubernetes Pods, Enhancing Data Security for AI Workloads

Rafay Systems and Fortanix have announced a joint solution that extends their confidential AI integration from virtual machines to Kubernetes pods. This new offering, unveiled as KubeCon + CloudNativeCon North America 2026 concludes, allows the Rafay Platform to orchestrate workloads running on CNCF Confidential Containers (CoCo) with Fortanix Confidential AI. This enables neoclouds, sovereign AI clouds, and enterprises to deliver confidential AI as a service. This development is highly significant for organizations dealing with sensitive data and intellectual property in their AI initiatives. The core benefit is the ability to run AI workloads within hardware-based Trusted Execution Environments (TEEs), where data and models remain encrypted even during computation. This directly impacts industries like financial services, healthcare, and the public sector, where regulatory compliance and data protection are non-negotiable. Practitioners can now deploy AI models that process confidential information on third-party GPU infrastructure with a much higher degree of assurance against unauthorized access or tampering. The broader trend in cloud-native computing is a relentless drive towards enhanced security and workload isolation, especially as AI adoption accelerates. Confidential computing, which protects data in use, is a natural evolution from encryption at rest and in transit. The integration with Kubernetes, the de facto standard for container orchestration, makes this advanced security accessible to a vast developer ecosystem. This move aligns with the increasing focus on securing AI pipelines, from data ingestion to model deployment and inference, particularly in multi-tenant or hybrid cloud environments. In practice, this means that developers and MLOps engineers can now design and deploy AI applications on Kubernetes with built-in, hardware-enforced confidentiality. They can leverage the Rafay Developer Hub to launch these workloads, streamlining the process and removing the need for complex manual configurations or support tickets. This offering provides a crucial layer of trust for enterprises hesitant to move sensitive AI workloads to the cloud or shared infrastructure due to security concerns. Practitioners should investigate how this solution integrates with their existing Kubernetes deployments and CI/CD pipelines, and evaluate its potential to unlock new AI use cases that were previously deemed too risky due to data privacy requirements.