NEAR AI has implemented independent verification

$NEAR AI Cloud has integrated Intel Trust Authority to provide independent verification of confidential AI computations. Clients can now obtain a signed Intel attestation token that confirms their tasks are executed within the secure Intel TDX environment.

$NEAR AI Cloud uses Intel TDX to isolate AI tasks within secure virtual machines. The memory of these machines is encrypted and inaccessible to the host operating system, hypervisor, and infrastructure operator. $NEAR AI already verifies hardware attestation and provides raw data on which the security of the environment can be verified.

The new integration adds independent verification from Intel to this. Intel Trust Authority verifies cryptographic signatures, matches secure environment parameters against Intel's reference values, and applies defined security policies. Following the verification, the service issues a signed JWT token confirming the result.

$NEAR AI Cloud is designed for confidential processing of AI queries and operation of AI agents. Intel develops hardware and software solutions for confidential computing and enterprise infrastructure, and Intel Trust Authority provides independent verification of secure computing environments.

The obtained token can be used in other systems to control access to protected resources. For example, key management systems and information security platforms may require such a token before granting access to confidential data or encryption keys. For companies, this also serves as an additional independent confirmation of security that can be used during audits and for compliance with regulatory requirements.

Users of $NEAR AI Cloud include Brave, Venice AI, Abound, the Government of Bermuda, OpenMind, and Corbits.

Image: Magnific