# Fortanix Extends Confidential AI Protections to Kubernetes Workloads

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/34922
Published: 2026-10-19T17:37:25.000Z
Updated: 2026-10-19T17:37:25.000Z
Section: Technology

> The design checks a container’s security policy and its CPU and GPU before releasing decryption keys.

Fortanix Inc. announced Oct. 19 that its Confidential AI platform will support Kubernetes workloads running through Confidential Containers, using hardware isolation and security checks to control access to decryption keys.

The integration connects those workloads to Fortanix’s security managers. Confidential Computing Manager checks the container and its hardware through remote attestation, while Data Security Manager handles key release. The hardware review covers the CPU and GPU together. Keys are released only after the container meets its security policy.

The controls address a point in AI inference, when a model processes input to produce an output. Model weights and input data may need to be decrypted in memory during that work. Traditional encryption protects data at rest and in transit, while confidential computing is designed to protect data during processing. Confidential Containers runs cloud-native workloads in hardware-isolated environments designed to protect data during computation.

“Moving AI into containers doesn’t change the threat,” said Anand Kashyap, Fortanix co-founder and CEO.

Fortanix’s open-source plugins connecting Confidential Containers with its security managers remain in development. A Sept. 17 update listed application support in an Early Customer Access Program, and a Sept. 18 deployment guide for Confidential Containers on Intel TDX carried a beta label.

Confidential Containers advanced to the Cloud Native Computing Foundation’s incubating tier on July 22, 2026, following approval by its Technical Oversight Committee.