OpenAI Commits $1 Billion to Cybersecurity Defense as New AI Finds Software Flaws Independently

OpenAI announced it will provide $1 billion in subsidized access to its cybersecurity tools over the next six months, aiming to help organizations strengthen their defenses against increasingly sophisticated AI-driven cyberattacks. The initiative comes shortly after the company revealed that its newest AI model can independently discover previously unknown software vulnerabilities and convert them into functional exploits without human guidance.

The subsidy program covers discounted access to OpenAI’s existing cybersecurity platform, which offers a standard tier for routine defensive tasks and a more advanced tier reserved for approved organizations handling sensitive security work. Thousands of organizations are already using the platform. Priority access is being extended to sectors such as utilities, government agencies, community banks and open-source software maintainers.

Crypto and Blockchain Projects Notably Absent From Priority List

While the announcement does not specifically mention cryptocurrency exchanges or blockchain protocols, many widely used open source blockchain tools are maintained by small, often unpaid teams responsible for securing systems that manage billions of dollars in value.

This gap comes amid a difficult stretch for crypto infrastructure security. Recent months have seen several notable incidents, including a vulnerability in widely used bitcoin payment software that led to stolen funds, warnings issued to Lightning Network node operators after AI-assisted vulnerability discoveries, and a major hardware wallet theft that prompted urgent firmware updates.

Industry Pushes for Equal Access to Advanced AI Security Tools

These incidents led dozens of major crypto companies to jointly request earlier access to advanced AI models for defensive security research, arguing that malicious actors were already obtaining similarly powerful tools, leaving essential blockchain infrastructure maintainers at a disadvantage compared to potential attackers.