Why did Palo Alto scan its entire codecase?

Palo Alto Networks turns AI into a full-code vulnerability sweep

Palo Alto Networks said it scanned its entire codebase using frontier AI models, a move it frames as the start of a “vulnpocalypse” where vendors increasingly rely on AI to find software flaws faster and at greater scale.

The core change is operational: instead of limiting AI use to narrow tasks (such as triaging alerts or suggesting fixes), the company applied frontier models across its full repository to uncover vulnerabilities and multiply patch efforts. Palo Alto previously said it typically finds only a small number of vulnerabilities per month, but the new approach signals a step-change in throughput—meaning more bugs can be surfaced sooner, and more teams may need to respond simultaneously.

The practical impact is twofold:

• Security teams may see more findings at once, because AI-assisted scanning can widen the net compared with traditional tooling.
• Patch volume can rise quickly, increasing the burden on engineering and release processes and potentially affecting timelines for mitigations.

For the industry, the broader takeaway is that the “AI for security” trend is shifting from experiments to production-like workflows. As more vendors adopt similar scanning strategies, the vulnerability discovery rate—and the pressure to keep patch pipelines healthy—could become a defining challenge for 2026 security operations.

In that context, the term “vulnpocalypse” underscores not just fear of more vulnerabilities, but the operational reality that AI may surface them faster than organizations can absorb and remediate.