Palo Alto Networks CEO: $5 Trillion AI Buildout Needs a New Security Stack

Nikesh Arora, chief executive of Palo Alto Networks, is warning that the wave of investment in artificial intelligence will force an equally large wave of spending on cybersecurity. In his view, simply building data centers and AI models is not enough - without a new security stack, companies will be exposed to attacks that move faster than humans can respond to them.

What Arora Said

Arora's comments came alongside Palo Alto Networks' fourth-quarter fiscal 2026 earnings, released on Tuesday, September 1. The company beat analyst expectations on both revenue and earnings per share, while also issuing a strong forecast for the coming fiscal year.

Speaking with Jim Cramer on CNBC's Mad Money, the Palo Alto Networks chief laid out a case meant to illustrate the scale of the problem. If a combined $5 trillion flows into data centers and AI computing capacity over the next five years, Arora argues, artificial intelligence must already be generating enormous value for enterprises to justify spending on that scale.

Nothing that was deployed seven or ten years ago is prepared or ready to handle AI moving at machine speed - Nikesh Arora, CEO of Palo Alto Networks

A Trillion Dollars of Outdated Infrastructure

Arora's central argument concerns the state of existing protection systems. The Palo Alto Networks CEO estimates that about $1 trillion of the world's existing cybersecurity infrastructure is outdated and unfit to defend against AI-driven threats. These are tools designed for an era when attacks were planned and carried out by humans, not by autonomous systems capable of scanning, testing, and exploiting vulnerabilities within seconds.

Arora argues that AI is working in attackers' favor faster than in defenders' favor, because it lowers the barrier to entry for criminals - sophisticated intrusion techniques are becoming accessible to a far wider pool of people than before. In his view, the industry's only sensible response is to build AI models directly into defensive tools and to consolidate scattered, often incompatible security systems into a single platform.

Quarterly Results and Acquisition Appetite

Arora's statements are not just sales rhetoric detached from the numbers. Palo Alto Networks reported fourth-quarter fiscal 2026 revenue of $3.41 billion versus an expected $3.35 billion, while adjusted earnings per share came in at $1.02 against analyst forecasts of $0.98. The next-generation security segment, made up of products built around AI and cloud platforms, now accounts for annual recurring revenue of about $9.1 billion, growing 63 percent year over year, with nearly $1 billion in new ARR added in the quarter alone.

For the next fiscal year, the company forecasts revenue in the range of $14.10-14.20 billion, above earlier market estimates of $13.79 billion. Palo Alto also continues a string of acquisitions bolstering its AI offering - the Console platform recently joined its portfolio, intended to support building automated, agentic security workflows that detect and remediate issues on their own.

What It Means for Businesses

For IT and security teams, Arora's message boils down to a simple warning: deploying AI without simultaneously modernizing security widens the attack surface faster than the ability to defend it grows. Every new AI agent, every system connected to a language model, and every automated machine-to-machine action is a potential new intrusion vector that classic, manually configured firewalls and detection systems cannot keep up with monitoring in real time.

In practice, this means growing pressure on security budgets at companies that are simultaneously investing in automation and AI agents - the cost of protection rises in proportion to the pace of deployment. The cybersecurity market is responding with consolidation: instead of buying more narrowly specialized tools, enterprises are increasingly moving to integrated platforms from a single vendor, which also explains the recent string of acquisitions by Palo Alto Networks.

For Polish companies and institutions still building their AI adoption strategies, Arora's remarks are a warning about sequencing. Investing in AI agents or process automation without first auditing and modernizing the security layer risks turning new systems into the weakest link in the infrastructure before they ever deliver the savings they were meant to bring.