AUTO-UPDATED

We’ve spent billions defending software. It’s time to protect execution

Cybersecurity experts are advocating for a shift toward hardware-based oversight to monitor processor instructions, aiming to secure critical infrastructure against vulnerabilities that software-only defenses currently fail to prevent.

Key Points

  • The United States recorded a record 3,322 data breaches in 2025, with 80% attributed to cyberattacks.
  • Memory-safety vulnerabilities, such as buffer overflows, account for approximately 70% of reported security bugs at companies like Microsoft and Google.
  • Verizon’s 2026 Data Breach Investigations Report identified vulnerability exploitation as the leading initial access vector, responsible for 31% of all breaches.
  • Anthropic reported that AI models, such as Claude Opus 4.6, are now capable of identifying and validating hundreds of high-severity vulnerabilities in open-source software.
  • Implementing an independent hardware layer allows processors to enforce security rules and stop prohibited behaviors before they result in successful exploits.

Why it Matters

Moving the security boundary from software to hardware provides a necessary backstop for critical systems that cannot rely on perfect code. This approach helps mitigate the risks posed by AI-accelerated attacks, which currently allow hackers to identify and exploit vulnerabilities faster than human teams can patch them.
The Next Web Published by Jothy Rosenberg
Read original