# Belfort

> Belfort builds hardware acceleration for Fully Homomorphic Encryption (FHE), making it practical to compute on encrypted data at scale. Spun out of KU Leuven's COSIC lab, with offices in Leuven (Belgium) and San Francisco (USA).

Belfort makes encrypted compute real. FHE lets you run computations directly on encrypted data, so sensitive information never has to be decrypted, not even during processing. The problem has always been speed: FHE is thousands of times slower than normal computing. Belfort solves this with purpose-built hardware, delivering the world's first hardware accelerator designed for encrypted compute.

The technology comes from over a decade of research at KU Leuven's COSIC lab, including contributions to DARPA's DPRIVE program. Belfort raised a $6M seed round in September 2025, led by Vsquared Ventures, with angels including Jeff Dean (Google) and Naval Ravikant. The company also holds an EIC Transition grant to develop its FHE ASIC.

Belfort's acceleration is available through the cloud or on-premises, and supports popular FHE schemes such as TFHE and CKKS. Example results: encrypted sanctions screening for Swift went from 4 minutes to under 1 second, and TFHE workloads run over 7,000x faster than a CPU baseline.

## Key pages

- [Home](https://belfortlabs.com/): What Belfort does and who is behind it
- [Technology](https://belfortlabs.com/technology): How FHE and Belfort's acceleration work, cloud and on-premises options
- [Blog](https://belfortlabs.com/blog): Announcements, technical deep dives, and customer stories
- [GitHub](https://github.com/belfortlabs): Open-source repos, including FPGA acceleration for TFHE-rs and GPU libraries

## Company

- Founded: 2025, spin-off from KU Leuven COSIC
- Team: Laurens De Poorter (CEO), Michiel Van Beirendonck (CTO), Furkan Turan (Chief Engineer), Ingrid Verbauwhede (Chief Scientist)
- Locations: Leuven, Belgium and San Francisco, USA
- Contact: via belfortlabs.com

## Use cases

- Finance: real-time encrypted sanctions screening, confidential analytics
- Healthcare: computing on patient data without exposing it
- Government and defence: processing classified or sensitive data securely
- AI: running machine learning inference on encrypted inputs