FROM RISK TO RESILIENCE: CONFIDENTIAL COMPUTING WITH AMD AND OPAQUE WHITE PAPER FROM RISK TO RESILIENCE: CONFIDENTIAL COMPUTING WITH AMD AND OPAQUE | JANUARY 2026 1

EXECUTIVE SUMMARY As the use of artificial intelligence becomes more common in business, the need to protect confidential data also grows. Information travels constantly between on-premises data centers and the cloud. While data is often encrypted when stored or in transit, it must typically be decrypted for use, creating a significant security gap when deployed on shared resources. Confidential computing closes this gap. What is confidential computing? According to the Confidential Computing Consortium, confidential computing (CC) is the protection of data in use by performing computation in...

THE MODERN DATA CHALLENGE IN AI The lifecycle of an AI workload — from data acquisition to model deployment — involves nearly continuous data manipulation across different environments. Historical data might be moved from mainframes to the cloud for processing, or new data could stream in from remote sensors. With traditional security methods, encryption protects data when it’s stored or in transit. When it’s in use, however, it must be decrypted, which introduces security challenges. Organizations working with sensitive information face three fundamental barriers to AI adoption: 1. THE COMPLEXITY...

SECURING DATA WITH TRUSTED EXECUTION ENVIRONMENTS Confidential computing mitigates data risks by performing computations in a hardware-based TEE, helping ensure the integrity of code execution and data protection within the environment. Before sharing data for processing in a TEE, an attestation process verifies that the hardware and software environment is genuine, secure, and compliant with required standards. By leveraging TEEs, organizations can securely process fragmented data across silos, without exposing it to unauthorized access, and help ensure compliance with data sovereignty laws...

OPAQUE SYSTEMS: ENABLING CONFIDENTIAL AI While hardware provides a secure foundation, organizations need a flexible platform to manage and orchestrate these capabilities for complex AI workflows. OPAQUE Systems unblocks AI value by making trust verifiable, enabling enterprises to use their most sensitive data to power AI innovation while adhering to security, compliance, and privacy guidelines. OPAQUE’s platform delivers verifiable runtime policy enforcement, backed by cryptographic audit logs. It leverages transparent confidential computing and TEEs, enabled by AMD SEV, that protect data throughout...

REAL-WORLD DEPLOYMENT: PROTECTING SENSITIVE CONSUMER INFORMATION For a practical look at confidential AI in action, consider a large U.S. credit management company. As a leading debt buyer, it routinely collaborates with hundreds of debt settlement companies (DSCs), which involves sharing sensitive consumer debt information, including Social Security numbers, often in CSV files. The old manual workflow for analyzing these files was slow and inefficient. This introduced several business challenges: • Data was processed in the clear, leaving sensitive information largely unencrypted and vulnerable....

THE OPAQUE AND AMD SOLUTION IN ACTION The implemented solution created an automated and compliant workflow with strong security capabilities that transformed the organization’s operations. It is a cloud-based, transparent, confidential computing environment powered by OPAQUE and AMD confidential VMs (CVMs) within the company’s Azure environment. Here’s an overview of how it works: 1. ATTESTATION: Remote attestation is the foundation of OPAQUE’s security model. Before any work begins, the OPAQUE platform cryptographically verifies that it has been deployed within a transparent confidential computing...

together we advance_confidential computing “Confidential AI is about turning sensitive data into advantage. AMD is thrilled to power partners like OPAQUE who help customers do that securely and at scale.” The joint solution delivered significant improvements: • STREAMLINED WORKFLOWS THROUGH AUTOMATION: Automating manual, multi-step workflows can reduce operational friction and shorten turnaround time. This helps shift staff effort from routine review of raw outputs (e.g., CSV files) toward higher-value analysis and strategic work. • ENHANCED SECURITY POSTURE: By accepting encrypted files from...

THE FUTURE OF CONFIDENTIAL COMPUTING This collaboration among a leading U.S. credit management company, OPAQUE, and AMD shows what’s possible when innovative hardware and software solutions come together to tackle real-world problems. By leveraging AMD SEV technology within the OPAQUE platform, the company has transformed its data handling processes, implementing robust security capabilities, enhancing efficiency and compliance. As AI continues to evolve, confidential computing will become an essential foundation for any organization looking to innovate responsibly with sensitive data. Learn...