No article in the knowledge graph for Desilo yet.
Who is Desilo?
Desilo is a privately held Korean company that develops homomorphic encryption and privacy-preserving computation software, enabling encrypted data to be processed without first being decrypted in enterprise and AI workloads.
- Homomorphic encryption software for computation on encrypted data
- Privacy-preserving machine learning and AI model inference
- Data security for organizations handling sensitive or regulated information
- Secure collaborative analytics across distinct data owners
- Cryptographic components for integration with enterprise data platforms
Show more
More About Desilo
Desilo positions its offerings for enterprise environments in which sensitive data must remain protected during processing. Typical use cases include analysis of combined datasets from several organizations that cannot share raw records, training and running machine learning models on encrypted inputs, and supporting compliance obligations for personal and financial information.
The technical approach is based on lattice-based cryptography, including approximate homomorphic encryption schemes that support arithmetic on real-number vectors. These operations are common in statistical analysis and machine learning. Implementations are generally exposed through programming interfaces and integrated with existing data pipelines and cloud or on-premises infrastructure. Lattice-based constructions also underpin several post-quantum cryptography standards, which places this class of technology within broader quantum-resistance planning.
Homomorphic encryption belongs to a solution category distinct from conventional encryption at rest and in transit. Traditional methods protect data while it is stored or moved, but data typically must be decrypted before it can be processed. Trusted execution environments and secure multiparty computation address related problems through different trade-offs in performance, trust assumptions, and hardware requirements. Organizations evaluating these approaches generally weigh computational overhead against the degree of data exposure they need to eliminate.
The business relevance of these offerings lies in enabling data collaboration and AI development that would otherwise be restricted by privacy regulation or data-sharing agreements. Their operational value depends on workload performance, the cost of encrypted computation, and the ability to fit into existing analytics and machine learning environments.
In the Decision Insights directory, Desilo is positioned within the cybersecurity category, specifically encryption, key management, and post-quantum cryptography, with additional activity in AI infrastructure and privacy-enhancing computation. This positioning reflects the company's current solution areas in data protection and secure computation.
Our description of Desilo. Updated October 2026.
