95% of critical infrastructure leaders fear Frontier AI-powered attacks: Palo Alto Networks

A survey of more than 1,600 security leaders across 11 countries finds organisations struggling with poor OT visibility, legacy systems and fragmented defences as AI-driven attacks speed up the exploitation of vulnerabilities.

Nearly all security leaders responsible for critical infrastructure are worried about attacks powered by Frontier AI, according to new research from Palo Alto Networks.

The company has published its 2026 State of Critical Infrastructure Cybersecurity Report, based on a survey of more than 1,600 security leaders across 11 countries and five critical sectors. The study looks at the challenges of securing increasingly connected IT and operational technology (OT) environments, including limited visibility, legacy infrastructure and the growing threat of Frontier AI-powered attacks.

Global findings: breaches and blind spots

The report found that 95% of respondents are concerned about Frontier AI-powered attacks. Some 59% said they had suffered a significant security breach in the past year, and half of those organisations cited physical safety concerns as a result.

Visibility into OT networks remains a weak point, with 68% saying they lack complete, real-time visibility into all assets connected to those networks. Legacy infrastructure is another major risk, with 42% naming legacy, unpatchable OT assets as their biggest cybersecurity risk. In addition, 74% have not fully integrated their IT and OT security operations.

Asia-Pacific: operational impact and tool sprawl

Based on 620 respondents in the region, the Asia-Pacific findings show a similar picture. Some 58% experienced a security incident that affected operations in the past 12 months, while 74% lack complete, real-time visibility into connected OT assets. Tool sprawl adds to the complexity, with 64% using six or more security tools across IT and OT environments.

Concern about Frontier AI is high in the region, with 92% worried about such attacks targeting critical infrastructure. At the same time, 90% expect AI security solutions to help defend against these threats. IT and OT convergence still has a long way to go, as 80% have not fully integrated their security operations.

Tom Scully, Industry Principal, Asia Pacific & Japan at Palo Alto Networks, said: “Critical infrastructure is entering a period where the systems organisations rely on most are becoming increasingly connected, while the time available to respond to security risks continues to shrink. This makes visibility and resilience across IT and OT environments increasingly important. Organisations need to move toward a more coordinated approach that can identify which systems are most critical, contain threats before they spread and put protection in place without disrupting essential operations.”

Priorities for leaders

The report urges organisations to move beyond fragmented security measures towards a coordinated approach. This begins with complete visibility into connected OT assets, prioritising vulnerabilities by operational risk and relying less on disconnected tools. More automation in threat detection and response, and closer integration between IT and OT security teams, will be critical to building resilience against increasingly sophisticated AI-powered threats, it said.

Protecting against evolving risks

Palo Alto Networks said its PAN-OS 12.2 Ceres introduces Frontier Virtual Patching, which helps protect against vulnerabilities at the network level. It also includes ruggedised next-generation firewalls designed for critical infrastructure environments.

The company’s Frontier AI Critical Defense Program extends this approach through collaboration with organisations across OT, healthcare, commercial software and open-source communities. The aim is to identify emerging vulnerabilities earlier and put protections in place before attackers can exploit them, reducing reliance on traditional patching cycles.