AI Cloud Protect: Next-Gen Enterprise Security by Check Point & NVIDIA

Businesses racing to implement artificial intelligence now have a powerful new security option with AI Cloud Protect, a joint solution from Check Point and NVIDIA designed specifically for on-premises enterprise AI environments. This integrated platform safeguards the entire AI lifecycle, including model development, agentic AI applications, and inference workloads, all while maintaining performance integrity. The solution is validated on NVIDIA RTX PRO Servers, allowing companies to deploy secure AI operations from their data centers to the cloud, leveraging the underlying power of NVIDIA BlueField technology.

Nataly Kremer, Chief Product Officer at Check Point, emphasized the critical need for this protection. “In the rush to build AI-driven innovation, enterprises simply cannot tolerate security blind spots,” she stated. “Our collaboration with NVIDIA ensures that AI factories are secure by design, protecting valuable models, data, and the underlying infrastructure without putting the brakes on innovation.”

The urgency for such a solution is clear. Over half of all enterprise networks currently utilize AI tools, making them attractive targets for cybercriminals. According to Check Point’s own data, approximately one out of every 80 prompts entered into generative AI systems risks exposing sensitive corporate information. Supporting this, a recent Gartner report revealed that 32% of organizations have already experienced an AI attack involving prompt manipulation, with another 29% reporting direct attacks on their generative AI infrastructure within the last year. Protecting the complete AI pipeline, from initial development to full-scale production, is no longer a secondary concern but a primary business imperative.

AI Cloud Protect is engineered to secure the entire AI factory, the cloud infrastructure that powers model development, training, and commercial deployment. By operating on NVIDIA BlueField-3 Data Processing Units (DPUs), the solution provides full-stack security without consuming valuable CPU or GPU resources. This architecture completely sidesteps the performance compromises typically associated with conventional security software. It employs network-level protection to guard AI infrastructure against unauthorized access, thereby mitigating risks related to data poisoning and model theft. At the host level, the platform utilizes the unique direct memory access capabilities of the NVIDIA DOCA Argus security framework. This provides complete visibility into every process running on an AI node, enabling Check Point to detect and block host-level malicious activities, including those originating from publicly downloaded Large Language Models (LLMs).

David Reber, chief security officer at NVIDIA, commented on the partnership, stating, “Robust security is a foundational requirement for the next generation of AI infrastructure. NVIDIA is collaborating with Check Point to integrate BlueField acceleration and the DOCA Argus runtime security framework directly into the AI Cloud Protect platform, empowering enterprises to deploy their AI initiatives with confidence.”

Looking ahead, AI Cloud Protect is also planned to run on the next-generation NVIDIA BlueField-4 DPU, which promises a substantial performance uplift with up to six times the compute power and double the network throughput for AI inference and storage processing. This will enable businesses to scale their secure AI operations at unprecedented speeds.

Beyond the core AI Cloud Protect offering, Check Point provides a comprehensive security portfolio that allows enterprises to safeguard the entire AI supply chain. This includes cloud infrastructure, enterprise applications, and end-users. For instance, the Check Point CloudGuard Web Application Firewall (WAF), enhanced with integrated Lakera capabilities, protects AI-powered enterprise applications from threats like prompt injection and jailbreaking. Simultaneously, Infinity GenAI Protect facilitates the safe enterprise adoption of generative AI tools. It identifies which GenAI applications are in use and enforces security policies to prevent sensitive data leaks and ensure regulatory compliance. Together, these solutions create an end-to-end security blanket across infrastructure, application, and user layers, giving businesses the assurance to deploy AI confidently at every stage.

Check Point is currently piloting AI Cloud Protect with financial services clients and strategic partners like World Wide Technology (WWT). These initial deployments are concentrated on AI data centers that support large language model (LLM) development and prompt-based applications, guaranteeing enterprise-grade protection for sensitive data and intellectual property. With the NVIDIA RTX PRO Server validation already complete, enterprises can immediately deploy AI Cloud Protect onto their existing AI infrastructure.

Chris Konrad, Vice President of Global Cyber at World Wide Technology, confirmed the solution’s value. “As enterprises construct AI server factories at scale, the combination of Check Point’s AI Cloud Protect and NVIDIA BlueField acceleration delivers the enterprise-grade protection required for sensitive AI workloads, from model training all the way to inference, all without sacrificing the high performance that modern AI applications demand.”

AI factories introduce a distinct set of security hurdles, including expansive attack surfaces created by massive compute and data pipelines, emerging threats like prompt injection and model poisoning, and data leakage risks stemming from ungoverned employee use of AI tools.

AI Cloud Protect directly addresses these challenges by providing comprehensive protection that spans network, host, and application layers. It achieves this with zero performance impact, thanks to its NVIDIA BlueField-accelerated security, and offers unified management and automation for seamless security orchestration across thousands of AI nodes.

(Source: ITWire Australia)