Crypto glossary
Fully homomorphic encryption (FHE)
Fully homomorphic encryption (FHE) allows computations on encrypted data without decrypting it, so results come out encrypted for the data's owner.
With ordinary encryption, data must be decrypted before anyone can process it. FHE schemes let a server add and multiply encrypted values directly, producing an encrypted result that only the key holder can read. The first fully homomorphic scheme was demonstrated by Craig Gentry in 2009; years of research since have made it much faster, though still far slower than computing on plain data.
In crypto, FHE is being explored to bring confidentiality to public blockchains. Projects build virtual machines where token balances, votes or bids stay encrypted on-chain while smart contracts still process them, enabling confidential tokens, private voting and sealed-bid auctions. FHE is often combined with zero-knowledge proofs and multi-party computation to manage keys and prove correct behaviour.
Performance remains the main barrier: FHE operations can be thousands of times slower than normal ones and need specialised hardware for practical throughput. Someone must hold the decryption key, so many designs split it across a committee using threshold cryptography, which reintroduces trust. Bugs in such young systems could expose supposedly private data.
Example
In an FHE-based auction contract, each bidder submits an encrypted bid; the contract computes the highest one without decrypting any, and only the winning amount is revealed at the end.