Privacy-Preserving AI with Fully Homomorphic Encryption

Privasea is a Swedish startup focused on advanced privacy-computing infrastructure for AI and digital identity. At its core, it offers systems built around fully homomorphic encryption (FHE) so that organizations can run machine learning or inference tasks on encrypted data, ensuring that sensitive data remains inaccessible, even during computation.

One of the company’s offerings is the ImHuman app, which lets users verify their human identity (often via biometrics) such that their human-likeness is confirmed without exposing raw personal data. The verified identity is used in the broader ecosystem for bot prevention, credentialing, and trust frameworks.

Privasea’s architecture also includes a decentralized compute network (Privanetix) and a library of FHE tools (HESea) designed to enable encrypted model computations, combined with Web3/token incentive elements.

Privasea has raised a total of approximately $20 million across two rounds, starting with a $5 million seed round in March 2024. Investors include Binance Labs, Gate Labs, MH Ventures, K300, QB Ventures, Crypto Times, and industry angels such as Zakaria (zak) Awes, and Luke Sheng from Chainlink. The company announced a $15 million Series A in January 2025 from GSR, Amber, and Echo, as well as some previous investors.