Descriptions:
Palo Alto Networks CEO Nikesh Arora revealed on the All-In Podcast that the company used Mythos, an AI-powered code vulnerability analysis tool, to find security flaws in its own codebase in six weeks that would have taken five to seven years through conventional security review — at a total compute cost in the “low millions.” Arora noted that Palo Alto considers itself among the top percentile of companies that rigorously test their code, making the magnitude of the finding particularly significant when extrapolated to the broader universe of organizations that write software with less security focus.
The interview covers the compound threat this capability represents going forward. Arora estimated that Mythos-level vulnerability discovery could become available in open-source or Chinese AI models within three months, turning the same analysis techniques into potential offensive tools. He identified smaller organizations — medical practices, dental offices, small businesses — as the most vulnerable targets, not major infrastructure, citing the Change Healthcare ransomware attack as a template for the kind of economic disruption these scenarios can produce at scale.
Arora also offers a business strategy perspective on where AI value is concentrating. He argues that profit pools in AI are shifting toward applications — pointing to Codex and Claude Code as examples of where OpenAI and Anthropic are actually capturing revenue — rather than raw model API usage, and predicts that AI intelligence will eventually be purchased as a utility priced by task complexity, with different “IQ levels” commanding different price points. Palo Alto Networks’ market cap has grown from $17 billion to $238 billion during Arora’s eight-year tenure as CEO.
📺 Source: All-In Podcast · Published June 08, 2026
🏷️ Format: Interview
