1The Confidential-Computing Threat ModelConfidential computing protects data while it is being used, but only against specific attackers and only inside a defined boundary.Advanced 2TEEs, Enclaves & AttestationA trusted execution environment protects code and data while they run, but its real value depends on proof, design and limits.Advanced 3Confidential Inference & TrainingHow confidential computing protects AI data and models while they are being used, not only while they are stored or moved.Advanced 4Protecting Model Weights & DataA practical guide to keeping the learned parameters, prompts, and data behind AI systems out of the wrong hands.Advanced 5Deployment Trade-offsDeployment is where design meets risk: the right home for a workload depends on latency, control, cost, scale, and change.Advanced