Check Point Launches AI Defense Plane and Security Blueprint

Check Point Software Technologies Ltd. introduced two major security initiatives on March 23, 2026, to address the transition of AI from simple assistants to autonomous agents. The company launched the AI Defense Plane, a unified control plane designed to govern the deployment and operation of AI. This platform uses an AI-native security engine to provide adaptive protection in under 50 milliseconds across more than 100 languages, integrating technology from ThreatCloud AI and acquisitions including Lakera and Cyata.

The AI Defense Plane consists of three modules: Workforce AI Security and AI Application & Agent Security, both available immediately, and AI Red Teaming, which is currently in limited release. The company plans to demonstrate its red teaming capabilities through a showcase called Gandalf: The Agent Gauntlet at the RSA Conference 2026.

Simultaneously, the company released the AI Factory Security Architecture Blueprint to secure private AI infrastructure from hardware to the application layer. Developed with NVIDIA, the blueprint uses BlueField DPUs and the DOCA software platform to protect GPU clusters and training pipelines against prompt injection, model theft, and data poisoning. The framework aligns with CISA's Secure by Design principles and the NIST AI Risk Management Framework to help organizations meet regulatory requirements such as the EU AI Act.