Mad Money

Nikesh Arora on how AI turned cybersecurity from near-death to a feast

Nikesh Arora· Chairman and CEO of Palo Alto Networks at Palo Alto Networks
·~9 min·English·CNBC
AI SafetyAgentsAI InfrastructureBusiness StrategyAI Company
TL;DR

Palo Alto Networks CEO Nikesh Arora explains why the AI he once feared would kill cybersecurity has instead made it existential, inverting the industry from chasing customers to being chased by them.

01Core Reframe

From Near-Death to the Feast

Nine months ago Arora expected AI to kill the cybersecurity business; now he says the industry gets to feast alongside AI instead of being eaten by it.

Nine months ago, I said we were guilty and convicted of near death because AI was going to eat our lunch, breakfast and dinner. It seems like that's not the case. Seems like we're going to have to have the feast with them.

— Nikesh Arora, Mad Money
Key Insight
The economic role flipped: AI first looked like a substitute for security vendors, then revealed itself as a complement that makes their product mandatory. The same force that threatened to automate the industry away is what now forces every buyer to spend more on it.
02The Catalyst

Dario Did It in One Event

Arora spent eight years failing to convince customers they were unprepared, and says a single frontier-lab launch made every CEO ask the question for him.

I've been trying for eight years to tell customers they're not ready. And Dario did it in one event just by launching Mythos.

— Nikesh Arora, Mad Money
Key Insight
When the person warning you profits from the warning, it lands as a sales pitch; when a third party demonstrates the danger, it lands as fact. Arora's most effective salesman turned out to be a frontier lab's product launch, which he credits to Anthropic's Dario Amodei.
03The Market Thesis

A Trillion Dollars of Hackable Legacy

Arora estimates about a trillion dollars of security infrastructure is deployed today, and says the gear installed 7 to 10 years ago was never built to fight AI at machine speed.

Nothing that was deployed 7 or 10 years ago is prepared or ready to handle AI at machine speed.

— Nikesh Arora, Mad Money
Key Insight
The obsolescence is a step-change, not a slow fade. A defense tuned to human-speed attacks is not merely old; once the attacker operates at machine speed it is the wrong category of tool, which is why Arora frames this as rearchitecting rather than upgrading.
04The Surprising Number

Governments Spend 78%. Enterprises Spend 4%.

Governments already devote most of their IT budgets to security while enterprises spend a rounding error, which is why Arora says the private sector, not government, is the unprepared one.

Government spends 78% of their budgets of IT on cyber. Enterprises spend 4 to 5% of their budgets on IT, on cyber.

— Nikesh Arora, Mad Money
Key Insight
Arora's number flips the usual assumption: the least-protected targets are not federal agencies with air-gapped systems but ordinary companies putting a far smaller share of their IT budget into security (4-5% versus 78%). The AI attack wave will find the enterprise first.
05The Contrarian Bet

The Punches That Paid Off

Three years ago analysts attacked Palo Alto's bet to bundle security into one platform, but Arora argues fragmented tools simply cannot defend at machine speed.

When multiple systems are fragmented, they cannot talk to each other at machine speed.

— Nikesh Arora, Mad Money
Key Insight
The platform bet is really a bet about tempo. A human team can juggle separate consoles, but automated defense correlating an attack cannot wait for tools that do not share state. Once attacks move at machine speed, integration stops being a convenience and becomes the only architecture that can respond in time.
06The Open Risk

Customers Aren't Ready for the Kill Switch

Arora says customers are not fully ready to deploy a kill switch for rogue agents, so his near-term promise is narrower: identify and govern an agent, then intercept and stop it midstream.

Even if you don't have a kill switch, you definitely need the ability to intercept agents and stop their activity.

— Nikesh Arora, Mad Money
Key Insight
The tell is what Arora does not promise. A kill switch implies a clean off button; conceding customers are not ready for one, he reframes the goal as interception, detecting an agent mid-action and halting it. That is the harder engineering problem, because it means watching every agent's behavior in the runtime path and deciding, live, which actions to stop.
07The Human Element

CEOs Remember 1999

The buying urgency is driven less by any specific threat than by fear: CEOs who watched the internet race leave companies behind are resolute not to be the has-beens this time.

I remember the 1999, we had the internet race, and a lot of companies were left behind because the internet created a whole new categories of companies. I think this time we're all resolute not to become the has beens.

— Nikesh Arora, Mad Money
Key Insight
Fear, not a breach, is the growth engine. The demand is adoption-driven rather than incident-driven: CEOs are buying security because they are racing to deploy AI, and being unprepared for AI feels like being unprepared for the internet in 1999, when nobody wanted to explain to a board why they were left behind.