Enhancing AI Security: Anthropic Integrates Opaque Systems’ Tech

Securing sensitive enterprise data remains a central challenge in deploying artificial intelligence across regulated industries. Anthropic announced a technology licensing and talent agreement with confidential AI startup Opaque Systems. Standard encryption methods secure information during storage and transmission, but processing requires unencrypting data within system memory, exposing potential vulnerabilities. Opaque addresses this vulnerability by processing workloads inside hardware-enforced trusted execution environments. These systems generate cryptographic attestations, offering verifiable proof that computations execute securely according to specific policies. Originating from UC Berkeley RISELab research, the platform serves organizations with strict compliance requirements, including healthcare and financial institutions. Integrating hardware-sealed privacy mechanisms allows model developers to strengthen compliance and security frameworks. Ultimately, adopting verifiable confidential computing provides independent mathematical assurance, helping establish robust governance standards for high-stakes enterprise applications.