FHE

Fully homomorphic encryption allows arbitrary computation directly on encrypted data, so a cloud service can process your information without ever seeing it in plaintext. First realized by Craig Gentry in 2009, FHE was long considered impractical, but modern schemes and hardware acceleration have made real workloads feasible. It is a cornerstone of privacy-enhancing technology for outsourced analytics, private machine learning, and confidential queries.