Palo Alto Networks CEO Nikesh Arora (니케시 아로라) said the spread of artificial intelligence has put companies worldwide in a situation where they must overhaul existing cybersecurity systems on a large scale.

On Sept. 1, CNBC reported that Arora said about $1 trillion worth of ageing security infrastructure built for pre-AI environments cannot handle automated threats.

Arora said no system built 7 or 10 years ago is ready to handle the machine speed of AI. He said cyber architecture must be rethought. That means companies need to redesign security systems first if they are to fully adopt AI in work and services.

That view was reflected in Palo Alto Networks' results. In its fiscal fourth-quarter earnings released that day, the company topped market expectations and offered strong guidance for the new fiscal year. Arora said AI also gives attackers faster ways to find and exploit vulnerabilities, increasing the burden on companies that maintain existing defences.

At the earnings briefing, he also said automated threats operate immediately, so the world needs to modernise about $1 trillion in cybersecurity debt to defend against them. In this context, cybersecurity debt refers to legacy infrastructure built for past environments and delayed demand for security investment.

Market views are also changing. Earlier this year, investors were concerned that advanced AI models could disrupt existing security software. Cybersecurity stocks, including Palo Alto Networks, were at one point under pressure. But as it became clear that attackers can also weaponise the same technology, AI has begun to be seen as a factor that instead increases demand for security.

Arora said the change brought a sharp reversal in investor sentiment. He said 9 months ago it felt as if AI would take away breakfast, lunch and dinner. He said it does not seem that way now, and that the company will also get to meet that opportunity with them.

He pointed to Anthropic's 'Mythos' model released earlier this year as a turning point. Arora said companies began to take cybersecurity more seriously after it became known that the model could be easily used to exploit software vulnerabilities. He said he had told customers for 8 years that they were not prepared, but a single event of Dario Amodei releasing Mythos changed that perception. Palo Alto Networks shares have surged 113 percent since April 7. The stock had been weak so far in 2026 before then.

Palo Alto Networks formally launched the Frontier AI Critical Defense Program in August to target that demand. The programme tests customers' defence systems with advanced AI models, finds vulnerabilities and supports security infrastructure modernisation. Arora said the company has discussed the programme with about 2,000 companies.

He said related spending is unlikely to be executed all at once in the short term. Arora said not everything happens next quarter, but that AI has fundamentally expanded the size and duration of the opportunity for the cybersecurity industry. He said the cybersecurity sector as a whole, not just Palo Alto Networks, is expected to keep looking for medium- to long-term benefits as AI adoption advances.