Map / Model Security / Mithril Security
Mithril Security
Quick answer: Mithril Security is an AI security company based in France and founded in 2021, working in Model Security and Sandboxing. It has raised $1.3M from investors including Auriga Cyber Ventures, Berkeley SkyDeck and IT-Farm. It was acquired by H Company in 2024.
mithrilsecurity.io ↗ · LinkedIn ↗ · Open on the map
Capabilities
Confidential AI inference allowing models to run on private data inside hardware-based trusted execution environments.
Sensors / integration
Open-source SDK and inference server deployed into TEE/secure enclaves (Intel SGX, GPU TPMs).
Protections
Guarantees data and model confidentiality during inference so providers cannot access user data and users cannot access model weights.
Differentiator
Open-source confidential AI tooling (BlindAI, BlindLlama) using hardware-backed enclaves with attestation.
Team & funding
- Founders
- Daniel Huynh, Mehdi Bessaa, Raphaël Millet
- Total raised
- $1.3M
- Investors
- Auriga Cyber Ventures, Berkeley SkyDeck, IT-Farm, Polytechnique Ventures
- Acquired by
- H Company · 2024
Other Model Security companies
ActiveFence · Advai · Adversa AI · AI Aramid · Alice · AptaSentry · Arize AI · Attestable · Bosch AIShield · Citadel AI · Cranium AI · DataKrypto · See all →
Last updated · Get the “Listed on” badge