A hardware security startup developing tamper-proof enclosures for AI chips to prevent physical attacks on AI hardware and enable international AI governance through verifiable compliance mechanisms.
A hardware security startup developing tamper-proof enclosures for AI chips to prevent physical attacks on AI hardware and enable international AI governance through verifiable compliance mechanisms.
People
Updated 05/18/26Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- $1,500,000
- Funding Raised to Date
- $461,000
Org Details
Updated 05/18/26TamperSec is a hardware security startup on a mission to secure AI hardware against physical tampering, protecting sensitive models and data from advanced attacks and enabling international governance of AI. The company was founded by Jonathan Happel and co-founded by Riccardo Varenna, with the team working full-time on this mission since September 2024. The core problem TamperSec addresses is that adversaries can manipulate AI hardware at the nanometer scale, allowing them to extract confidential data, steal AI model weights, or disable on-chip governance mechanisms. These attacks can be executed at scale and remain both undetectable and unpreventable with current technology. TamperSec's solution is a secure physical enclosure for AI hardware that is retrofittable to existing servers. The device is capable of identifying nanometer-scale modifications and automatically deletes a secret attestation key upon detecting tampering. This technology is closely related to the concept of Flexible Hardware-Enabled Guarantees (FlexHEGs), which are mechanisms designed to enable verification and automated compliance guarantees that are multilateral, privacy-preserving, and trustworthy for agreements regarding AI development and use. The company was incubated by Impact Academy and participated in the Winter 2024/25 Catalyze AI Safety Incubation Program in London, where it was one of eleven new AI safety organizations introduced. TamperSec has also been supported by the Survival and Flourishing Fund through their FlexHEGs funding round, receiving $461,000, as well as by Entrepreneur First's def/acc program. Jonathan Happel brings a decade of product development experience, with a background as a mechanical engineer and expertise in hardware security and risk management in medical devices. Riccardo Varenna, as co-founder and CEO, brings over nine years of experience scaling startups and previously worked with Entrepreneur First, AI Safety Collab, Control AI, and helped build Impact Academy from the ground up. TamperSec is structured as a Delaware C Corp with plans to restructure as a Public Benefit Corporation. The company has completed two hardware prototype iterations and has been actively hiring for electronic engineers, embedded systems engineers, and business development roles with offices in London and San Francisco.
Theory of Change
Updated 05/18/26TamperSec believes that effective international AI governance requires trustworthy hardware-level verification mechanisms. Without physical tamper protection, adversaries can manipulate AI chips at the nanometer scale to steal model weights, extract confidential data, or disable governance mechanisms, undermining any software-based compliance frameworks. By developing secure, retrofittable enclosures that detect physical tampering and destroy attestation keys upon breach, TamperSec aims to provide the hardware foundation for FlexHEG (Flexible Hardware-Enabled Guarantee) mechanisms. These mechanisms could enable privacy-preserving verification of compliance with international AI safety agreements, making it possible to enforce compute limits, conduct model evaluations, and ensure safety protocol adherence at the hardware level. The causal chain runs from tamper-proof hardware to trustworthy compliance verification to enforceable international AI governance treaties, ultimately reducing the risk of uncontrolled AI development.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.