Duality Confidential Computing Platform is a secure data collaboration solution that enables organizations to run analytics and machine learning computations on sensitive data within hardware-backed Trusted Execution Environments (TEEs), also known as secure enclaves. The platform allows multiple parties to collaborate on sensitive datasets without exposing the underlying data during processing. Computations are performed inside TEEs hosted on major cloud providers, with hardware-backed proofs of execution providing verifiable confidentiality guarantees. Key capabilities include: - Training, tuning, and validating AI/ML models — including Large Language Models (LLMs) — on encrypted or sensitive data - Support for structured and unstructured data types (text, audio, images, etc.) - Built-in data management tools for schema alignment, transformation, and pre-processing for multiparty collaboration - Protection of model intellectual property (IP) when running computations on external or client infrastructure - Cross-border data analysis while satisfying data privacy and residency requirements - Enabling third-party analytics and ML models to be applied to sensitive data without exposing it The platform targets use cases such as medical multi-center studies, data enrichment, secure cloud migration of AI workloads, co-marketing and personalization, and data monetization. Performance is described as comparable to running computations in the clear, without the overhead typically associated with cryptographic privacy techniques.