top of page
mithrilAI Web Header v2.png

Trusted AI needs trusted hardware

Do you trust your AI hardware? What security and reliability guarantees does your AI hardware provide? What happens if your adversaries or competitors get their hands on your AI hardware?

Hardware-Level Defenses

Consulting Services

Military & Civilian Applications

Unhackable AI Models

Our Latest Innovative Product

Our provably secure ML chip.jpg

We have designed hardware that will run your AI tasks slower with a larger chip. 

 

But wait, there is a catch: if you can bear these costs, we can ensure your valuable AI models are unhackable even when your adversaries get their hands on your systems!

About mithrilAI

mithrilAI is an NC State University startup company. We pioneer innovation at the intersection of AI and secure hardware. We have developed the world's first silicon-proven artificial intelligence circuits designed to thwart hardware attacks, including side-channel and fault injection threats. Our cutting-edge technologies ensure the confidentiality, integrity, and availability of AI models, safeguarding critical systems from vulnerabilities.

We specialize in partnering with AI hardware providers focused on safety- and security-critical applications. Our expertise supports military systems in aviation, autonomous driving, cybersecurity, and critical infrastructure management. Semiconductor companies and system integrators serving the defense sector trust mithrilAI Corp. to fortify their products against evolving threats.

Looking forward, mithrilAI Corp. is committed to expanding its impact into civilian applications where safety and security are paramount, ensuring robust protection in an increasingly AI-driven world.

Join mithrilAI Corp. as we redefine security standards for artificial intelligence hardware.

Our Technology

Our solution utilizes randomized multi-party computation to make individual computations within hardware computationally intractable while maintaining essential functionalities. This cryptographic-level defense significantly enhances the privacy, security, and reliability of AI applications, offering substantial advantages across diverse industries.

In protecting real-world AI/ML workloads, our technology extends hardware masking, shuffling, and fault-injection resilience concepts to encompass large and intricate networks. To ensure high throughput and low overhead designs that meet real-world performance demands, we are developing a new architecture with custom-tuned and scalable hardware defenses.

Here is our newest paper related to our technology
mithrilAI Deck Bkgd.png

A Full-Stack Approach for Side-Channel Secure ML Hardware

Machine learning (ML) has recently emerged as an application with confidentiality needs. A trained ML model is indeed a high-value intellectual property (IP), making it a...

Our Team

Aydin Aysu headshot.png

Dr. Aysu is currently an associate professor and University Faculty Fellow at the Electrical and Computer Engineering Department of North Carolina State University, where he leads HECTOR: Hardware Cybersecurity Research Lab. He got his M.S from Sabanci University in Istanbul, Turkey, and his Ph.D. from Virginia Tech. Before joining NC State, he was a post-doctoral researcher at the University of Texas at Austin. Dr. Aysu's interests are broadly in hardware security research and cybersecurity education. His hardware security research has won NSF CAREER, NSF CRII, Google RSP, and Goodnight Innovation Fellow, and Bennett Faculty Fellow awards, six best paper nominations (IACR TCHES, IEEE HOST, DATE, GLS-VLSI), two best paper awards, two hardware security top picks (IEEE CEDA), and one publicity paper award (DAC). He is an IEEE senior member and a co-founder and president of mithrilAI Corp.

  • LinkedIn
  • Website Icon A

Aydin Aysu

FOUNDER & PRESIDENT

Furkan Aydin headshot_2.png

Furkan received his Ph.D. degree in Computer Engineering from North Carolina State University in 2024, where he dedicated five years to researching hardware security for emerging applications. He has authored many scientific papers, focusing on exposing and mitigating side-channel attacks on hardware with innovative solutions. Notably, he published the first side-channel attacks on homomorphic encryption. He received his M.S. and B.S. degrees from Ozyegin University in Istanbul, Turkey. He also has over eight years of hardware design experience. His research interests are generally in hardware security research and secure AI/ML hardware implementations.

  • LinkedIn

Furkan Aydin

HEAD OF ENGINEERING

Latest News

Contact Us

Please complete the form below and a mithrilAI team member will be in touch as soon as we are able.

bottom of page