Nikesh AroraChairman & CEO, Palo Alto Networks, Inc. (PANW)
10 recorded appearances
Arora argued the window between a vulnerability being discovered and exploited is collapsing, so patching has to move from a ~55-day industry baseline to hours. He also explained why he signed Jensen Huang's open-source AI letter and why autonomous agents force enterprise security to be rebuilt.
Launching an AI-driven platform that compresses detection-to-protection from weeks to hours, Arora told Ed Ludlow he expects enterprises to eventually spend as much as 15% of operating expenses on AI.
Arora said AI token pricing needs to fall roughly 90% before large-scale enterprise adoption is economic, while insisting demand for AI consumption is effectively infinite and that his own capex will decline over time as efficiency improves.
Arora laid out a breadth-versus-depth critique of frontier models, said enterprises are still deploying AI wrong, and warned that current AI guardrails are not robust enough for agentic workloads.
Arora described AI surfacing years' worth of vulnerabilities in Palo Alto's own codebase in weeks, argued analytical SaaS is dead, and discussed where value accrues if models become a utility, plus the company's M&A playbook toward a $1T ambition.
Arora defended the quarter and made the case that platform consolidation plus AI integration is what lets Palo Alto grow while the rest of the security stack fragments.
Arora told Aditya Agarwal that AI is simultaneously the threat and the only viable defense, and that the same models writing code today will soon expose decades of latent vulnerabilities faster than IT teams can respond. He also recounted arriving at Palo Alto from Google with no security background.
Arora opened the company's keynote with the strategic framing for the agentic-AI era, ahead of the launch of Prisma AIRS 3.0 and the Prisma Browser for Business and the roadmap to an autonomous SOC.
Speaking with Deirdre Bosa, Arora tied rising cyber risk to geopolitical conflict and AI disruption, and explained the rationale for partnering with Databricks.
Arora's pitch that security controls for AI have to be built into the platform layer, otherwise enterprises will choose speed over safety and deploy agents ungoverned.