NEAR AI introduced two new privacy-focused products aimed at bringing verifiable confidentiality to AI workloads, expanding the push for secure, compliant infrastructure in an increasingly regulated environment. The rollout includes NEAR AI Cloud, a confidential computing platform, and NEAR Private Chat, an encrypted interface for interacting with open-source AI models. Both products rely on hardware-based isolation and cryptographic proof to ensure that sensitive data is processed securely.
The cloud platform allows companies and developers to run models within trusted execution environments using technologies such as Intel TDX and NVIDIA Confidential Computing. Each request generates a verifiable proof showing how and where it was executed, offering enterprises a higher degree of assurance compared with standard cloud environments. Early partners include Brave, Phala, and OpenMind, which are integrating the system into privacy-first and data-sensitive applications.
NEAR Private Chat provides a secure channel for users seeking confidential interaction with AI models, with conversations kept encrypted and excluded from future model training. The launch also supports migration through compatibility with the OpenAI API and SDKs for major programming languages. NEAR AI plans to extend its roadmap to include encrypted memory, agent frameworks, and decentralized coordination infrastructure supported by the NEAR blockchain.
