Homomorphic Encryption for Confidential AI

Blyss delivers a confidential AI platform built to protect data not only at rest or in transit, but even while it is actively being processed. Using GPU- and CPU-based secure enclaves, the company enables encrypted-in-use computation for highly sensitive workloads, allowing developers to store proprietary information and retrieve it privately through a key-value architecture that conceals access patterns from everyone — including Blyss itself.

Founded in 2022 in San Francisco and backed by Y Combinator and Unit Ventures, Blyss aims to democratize advanced cryptographic protections long reserved for large tech firms. Its mission is to make strong privacy infrastructure accessible to any organization that needs to run AI on confidential or regulated data, offering a practical path toward building systems where security is not bolted on, but fundamentally woven into the act of computation itself.