With the flourishing development of 5G networks and IoT devices, the number of cyber-attacks exploiting vulnerabilities in IoT devices is increasing. To address these threats, this project will focus on implementing a secure module software-hardware integration architecture using a security chip mode, divided into three parts: security chip trust root and security detection defense technology, next-generation zero-trust technology and architecture design, and AI security, data, and model protection technology. In the first part, we will develop security chip trust root and security detection defense technology, including hardware modules and software security modules. The research content includes trust root design, chip Trojan detection technology, side-channel attack detection technology, and fault injection attack detection technology, which will help develop domestic security chips and cooperate with the IoT device industry to create more competitive secure IoT devices. In the second part, we will implement a zero-trust security architecture, which includes identity verification for both person entities and non-person entities. Authentication and authorization activities will occur in multiple places across the enterprise but will be centralized at the same authentication execution point. Records generated at each authentication execution point will be sent to SIEM for analysis to establish trust levels. Device and user trust levels are independently scored and aggregated where appropriate to execute policies. If the trust level score of a non-person entity or user is higher than the measurement threshold, they are authorized to access the requested data. During this process, data will be protected by Data Loss Prevention (DLP). The third part will explore AI security and data protection technology by transplanting the HP SL platform and precision medical data to develop OurChain, which can be used for both public and private use to achieve better security and privacy performance while designing AV applications.