COTI Foundation deployed a production-ready garbled circuits implementation on its mainnet in March 2025, claiming performance advantages of up to 3,000 times faster than fully homomorphic encryption and 250 times more efficient than alternative privacy solutions.
Garbled circuits are a cryptographic protocol that allows multiple parties to jointly compute a function on private data without any party learning the others' inputs. The implementation, called the gcEVM, is fully compatible with the Ethereum Virtual Machine and was built in partnership with Soda Labs, according to a post on X by COTI Foundation.
The technology originates from computer scientist Andrew Yao's 1982 paper "Protocols for Secure Computations," which framed the problem as two millionaires determining who is wealthier without either disclosing their net worth. The protocol remained largely confined to academic research for four decades because hardware capabilities and algorithmic optimizations necessary for blockchain deployment did not exist at production scale.
COTI's implementation marks the first time garbled circuits have been deployed as an optimized, live blockchain application. The gcEVM executes privacy-preserving computations while maintaining EVM compatibility, meaning developers can port existing smart contracts to privacy-enhanced versions without redesigning core logic.
The performance metrics compare garbled circuits against established privacy technologies. Fully homomorphic encryption, which allows computation on encrypted data without decryption, requires substantially more computational work per operation. Alternative privacy solutions—including zero-knowledge proofs and other secure computation methods—incur larger proof sizes and verification overhead. COTI's benchmarks reflect these differences across latency and on-chain footprint.
The mainnet deployment follows testing phases and represents COTI's move to integrate privacy infrastructure directly into a live blockchain environment. Soda Labs contributed optimization work to the implementation. The partnership aligns with a broader industry focus on practical privacy solutions that do not sacrifice performance or developer experience.
