Exabits and Phala Network Enhance AI Security with TEE-Enabled Infrastructure

Exabits has partnered with Phala Network to enhance AI security by offering trusted execution environment (TEE) enabled GPU clusters, announced in a press release. This collaboration aims to ensure that personal data is processed within a secure enclave, providing heightened data protection for AI applications.

The partnership leverages Exabits' GPU clusters and Phala Network's trustless Web 3.0 computation platform to create a TEE-enabled infrastructure. This setup ensures that sensitive user data remains secure across hardware, software, and cloud systems, particularly when serving large language models (LLMs).

Exabits and Phala Network's joint solution includes TEE-enabled Nvidia H200 clusters, which allow developers to run applications like DeepSeek R1 within a secure domain. This advancement provides an isolated environment for processing data safely, keeping sensitive information protected from third-party servers.

The collaboration also integrates security features from Intel and Nvidia's TEE technology stack, along with Phala's Redpill and private-ml-sdk. These solutions ensure user data is protected within a tamper-proof, encrypted environment, preventing cloud-based AI applications from storing data externally.