Fr0ntierX is proud to introduce the second installment of our Polaris Secure Container Series: Polaris AI and Polaris LLM, designed to provide superior protection for AI (including AI agents), machine learning (ML) models, and large language models (LLMs).

Powered by Confidential Computing and Trusted Execution Environments (TEEs), the Polaris Secure Container series ensures that sensitive data remains fully encrypted throughout its lifecycle – whether at rest, in transit, or in use – and is fully isolated from cloud providers, internal IT resources, and even us.

Polaris AI enables secure deployments of AI and ML models within a TEE, encrypting model weights and sensitive data in transit, at rest, and in use. Utilizing Confidential Virtual Machines (CVMs) and Confidential GPUs based on the NVIDIA Hopper architecture, Polaris AI provides maximum security with minimal overhead.

Polaris LLM includes the protection of large language models and related applications, such as retrieval-augmented generation (RAG). Like Polaris AI, Polaris LLM secures these large language models within a TEE, encrypting both model weights and context documents. By leveraging CVMs and Confidential GPUs, Polaris LLM ensures complete isolation, with encrypted data storage and decryption handled within the TEE.

Polaris AI and Polaris LLM employ transparent encryption for HTTP requests and responses, ensuring no changes to AI models or common inference servers are required.

Available now on Azure Marketplace, Polaris AI and Polaris LLM offer robust data protection for AI, machine learning, and LLM applications, providing security without compromising performance or scalability.