AMD's AI Security Push: TRACE & Confidential Computing

Advanced Micro Devices, or AMD, has introduced TRACE, a new open specification for verifiable AI runtime evidence, now under Linux Foundation governance. This initiative aims to create a vendor-neutral, hardware-enforced standard for trusted AI workloads. Separately, Anjuna Seaglass confidential computing is rolling out on AMD EPYC CPUs, utilizing AMD Secure Virtualization to protect sensitive data. Both efforts target secure AI and confidential computing, which are crucial for regulated industries and wider AI infrastructure adoption. AMD, with a reported market value of about $745.6 billion, is a US semiconductor company supplying processors for data centers and AI. These security-focused developments support enterprises seeking more control over their AI models. TRACE gives AMD a role in setting an open, hardware-enforced evidence standard for AI agents and confidential workloads. This aligns with how enterprises and governments want to verify AI systems are running as intended. The practical test will be how quickly TRACE and Anjuna Seaglass deployments show up in usage metrics and customer disclosures through 2027, especially in regulated sectors. This matters because it shows how AMD is addressing security and policy themes critical to its AI strategy.