The European Telecommunications Standards Institute (ETSI) has officially released the technical specification ETSI TS 104 033 'Securing Artificial Intelligence (SAI); Security requirements for an Artificial Intelligence Computing Platform,' led by Huawei. This is ETSI's first international standard for AI computing platform security. The project was initiated by Huawei at an ETSI SAI meeting in October 2021 and gained support from international partners including BT, Qualcomm, Bosch, and Germany's BSI. The standard addresses security risks for AI computing platforms such as firmware and driver tampering, unauthorized access to computing resources, leakage of model and data assets, platform availability attacks, and fault injection, proposing mitigation measures. The Ascend security solution starts from a hardware root of trust to enable secure boot and measured boot, provides mandatory access control policies and a hardware-based confidential computing environment, and supports automatic recovery for checkpoint resumption. This standard, together with the ETSI EN 304 223 AI security baseline standard and ETSI TS 104 223 AI model and system security requirements, forms the ETSI SAI security standard system.