Cato Networks unveils GPU-powered SASE with native AI security controls

Cato Networks this week launched two additions to its secure access service edge (SASE) platform that the company says will address security challenges enterprises are facing now: protecting the AI tools end users rely on, while also using AI to defend against sophisticated threats. Cato Neural Edge deploys Nvidia GPUs across the 85+ points of presence (PoPs) of Cato’s global private backbone to enable real-time traffic inspection, threat detection, and policy enforcement. Cato is also launching Cato AI Security, which combines AI governance and runtime protection capabilities the company acquired with AIM Security into the Cato SASE Platform. “AI is changing both sides of the equation: the threats we face and the defenses we need. With Cato Neural Edge and Cato AI Security, we are empowering enterprises to strengthen AI-driven defense and govern enterprise AI without sacrificing performance or adding operational complexity,” said Matan Getz, vice president of AI security at Cato Networks, in a statement. According to Gartner, by 2028, more than 75% of enterprises will rely on AI-amplified cybersecurity products for the majority of use cases, up from less than 25% in 2025. This reality is reshaping cybersecurity, according to Cato Networks, and driving the need for enterprises to “evolve both their security controls and the infrastructure powering them.” Cato Neural Edge embeds a GPU-powered enforcement layer directly within the company’s global PoP network, enabling Cato’s platform to execute intelligence and enforcement within the PoP itself, the company says. Cato Neural Edge can enable: * High-frequency execution of AI/ML models inline * Real-time semantic and behavioral inspection * Scalable analysis across global traffic flows * Deterministic performance without external processing layers Cato AI Security is designed to govern employee use of AI tools, secure homegrown AI applications, and enforce guardrails for autonomous AI agents, according to the company. It can operate as a standalone solution or with additional Cato SASE Platform capabilities, including SD-WAN, SSE, and universal ZTNA. The capabilities can be managed via a unified control plane and policy engine that shares context across the platform to deliver faster detection and response, Cato says. One early adopter shared their impressions of the Cato AI Security capability: “One of the biggest advantages for us is that AI security isn’t another console or separate enforcement layer. It’s built directly into the Cato SASE Platform,” said Marc Crudgington, vice president of cybersecurity and IT infrastructure at global logistics company Crane Worldwide Logistics, in a statement. “We can govern AI usage, secure homegrown AI applications, and manage agent workflows using the same policy engine and data lake that already protect our network and cloud environments. That unified architecture reduces complexity and ensures consistent enforcement everywhere AI operates.” Cato SASE Cloud Platform runs on a private global backbone of more than 85+ points of presence (PoP) connected via multiple SLA-backed network providers. The PoPs software continuously monitors the providers for latency, packet loss, and jitter to determine in real-time the best route for every packet. Cato applies optimization and acceleration to all traffic going through the backbone to enhance application performance and the user experience. To ensure all locations benefit, Cato optimizes traffic from all the edges and toward all destinations, on-premises and in the cloud. Cato AI Security is generally available to customers worldwide. Cato Neural Edge has been deployed across the Cato SASE Platform.