Check Point introduces new security solutions for AI factories and against vulnerabilities in AI systems. The solutions aim to bridge the growing gap between innovation and cybersecurity.
Organizations are investing heavily in infrastructure such as GPU clusters and data lakes, but these systems are proving particularly vulnerable to advanced cyberattacks. Traditional security solutions cannot adequately protect the complex architecture of AI factories. Check Point is responding with a three-pronged approach: a blueprint for data centers, an AI Defense Plane, and a platform for managing AI agents.
The new approach comes at a time when organizations worldwide are facing AI-driven attacks that are faster and more cunning than ever before. The integration of AI into business processes creates new attack surfaces, ranging from prompt injection to data leaks. Check Point aims to offer a complete framework for preventive cybersecurity in the AI era.
Security plan
AI factories combine powerful GPU clusters, distributed training pipelines, and real-time inference APIs, making traditional security models inadequate. Just as physical factories need a security plan, Check Point is designing this for AI factories. Check Point’s new AI Data Center Security Architecture Blueprint provides security across four critical layers.
The perimeter layer is secured by the Check Point Maestro Hyperscale Firewall, which enables zero trust and scalable policy enforcement. At the application and LLM layer, a WAF protects against prompt injection and API abuse.
At the workload and container layer, integration with Illumio ensures microsegmentation in Kubernetes clusters, preventing lateral movement within the network. For the hardware layer, Check Point is collaborating with Nvidia to embed security directly into Bluefield DPUs. This allows for inspection without consuming GPU resources.
AI Defense Plane
With the rise of AI agents and autonomous systems, a new risk category is emerging that surpasses traditional security measures. Check Point’s AI Defense Plane offers a unified architecture for managing AI agents at three levels. For employees, the platform provides visibility into the use of AI tools, with policy enforcement to prevent data leaks without disrupting the user experience.
For AI applications and agents, the platform expands capabilities following the acquisition of Cyata. Security teams gain insight into which agents are active, which tools they have access to, and what actions they perform. Additionally, a cloud-based AI Red Teaming function is available, simulating attacks to uncover vulnerabilities in reasoning and workflows before attackers can exploit them. This feature is currently in controlled availability.
Jonathan Zanger, CTO of Check Point, emphasizes that AI security must be integrated from day one. Check Point offers a robust framework that protects the infrastructure as well as users and agents.
