Semiconductor Insiders Notify me• Jun 12, 2026• 15:56Interview

Podcast EP350: The Growing Threat of Hardware Security Breaches and What to do About it with Dr. Andreas Kuehlmann

From Semiconductor Insiders

Dr. Andreas Kuehlmann(General Manager of Security Solutions, Arteris, guest)

Get the full transcript next time Semiconductor Insiders releases an episode

Summary, key quotes, top claims, and the searchable transcript - emailed automatically. No card needed.

Executive Summary

Continue your research

Keep pulling the thread on Dr. Andreas Kuehlmann.

AI as a Potent Vulnerability Discovery Engine The Hardware Security Lag

10 quotes

Concerns Raised

AI models will soon automate the discovery and exploitation of hardware vulnerabilities at an unprecedented scale.
The semiconductor industry is not yet prepared for this new class of AI-driven threats.
Hardware vulnerabilities are fundamentally harder and more expensive to fix than software, often requiring physical replacement of chips.
The potential for 'blended attacks' combining software and hardware exploits creates complex and highly damaging threat scenarios.

Opportunities Identified

Companies that proactively adopt robust security practices like 'shift-left' verification and supply chain transparency can gain a significant trust advantage.
The development of a 'Hardware Bill of Material' (HBOM) standard can improve security and visibility across the entire semiconductor supply chain.
There is a window of opportunity for the industry to prepare its defenses before AI-driven hardware attacks become widespread.

Key Themes

Research Findings10

The AI model Mythos successfully compromised Apple's M5 chip memory defense systems within a few days.

The AI model Mythos discovered hundreds of zero-day vulnerabilities in major software platforms, including the Firefox web browser, Linux kernel components, cryptographic libraries, and smartphone firmware chains.

The public release of the AI models Mythos and GPT-5.5 has been delayed due to their discovery of numerous zero-day vulnerabilities, with access currently restricted to selected entities for remediation.

GitHub hosts 400 million public software repositories, providing a vast source of training data for AI models.

The open-source hardware ecosystem is significantly smaller than the software ecosystem, with platforms like OpenCourseWare hosting just under 900 designs.

The 2018 Spectre and Meltdown vulnerabilities affected processors from major designers including Intel, AMD, and ARM.

The number of publicly disclosed hardware vulnerabilities (CVEs) has grown dramatically since 2018, following the discovery of Spectre and Meltdown.

An AI model from OpenAI disproved the 80-year-old 'planar unit distance problem,' a mathematical conjecture by Paul Erdős.

The European Cyber Resilience Act is expected to impose significant compliance costs on companies.

Dr. Andreas Kuehlmann advocates for the semiconductor industry to adopt a 'Hardware Bill of Material' to improve supply chain security, similar to the 'Software Bill of Material' in the software industry.

Topics

Cybersecurity Artificial Intelligence Hardware Security Semiconductor Industry Zero-Day Vulnerabilities AI Safety Mythos (AI Model)GPT-5.5 Spectre and Meltdown Supply Chain Security Hardware Bill of Material (HBOM)European Cyber Resilience Act Shift-Left Security Vulnerability Discovery Open Source Hardware

Processed Jun 16, 2026Daily intelligence brief → yt-dlp + mlx-whisper + Gemini

Podcast EP350: The Growing Threat of Hardware Security Breaches and What to do About it with Dr. Andreas Kuehlmann

Continue your research

Concerns Raised

Opportunities Identified

Key Themes

AI as a Potent Vulnerability Discovery Engine

The Hardware Security Lag

The Inevitability of AI-Driven Hardware Attacks

A Proactive Security Framework for Semiconductors

Research Findings10

Topics