Check Point & NVIDIA launch AI Cloud Protect for secure AI use

Check Point has announced the availability of AI Cloud Protect, a security solution for AI infrastructure developed in collaboration with NVIDIA.

The new AI Cloud Protect product is designed to provide protection for AI model development environments, agentic AI applications, and inference workloads, available to enterprises for on-premises deployment. With validation on NVIDIA RTX PRO Servers and leveraging NVIDIA BlueField technology, the tool aims to secure AI operations from data centre to cloud without impacting system performance.

AI security needs

AI adoption in enterprises has seen a rapid uptick, introducing new and complex security considerations. Check Point reports that AI is now in use on more than half of enterprise networks, which has increased the risk profile of these organisations.

The company notes that their data indicates one in every eighty generative AI prompts can expose sensitive data. This is corroborated by a recent Gartner report, cited in the announcement, which found that 32% of organisations experienced prompt manipulation attacks on their AI systems, while 29% reported attacks targeting their generative AI infrastructure in the preceding year.

Nataly Kremer, Chief Product Officer at Check Point, commented on the need for integrated protection, stating:

"As enterprises race to build AI-driven innovation, they can't afford blind spots. With NVIDIA, we're making AI factories secure by design-protecting models, data, and infrastructure without slowing innovation."

Product features

AI Cloud Protect works within AI factories-the infrastructure powering model development, training, and deployment-running on NVIDIA BlueField-3 data processing units (DPUs). The security solution is designed to operate independently of CPU and GPU resources, aiming to eliminate the trade-off between security and performance typical of previous systems.

The solution deploys security at the network level to restrict unauthorised access, reducing risks such as data poisoning and model exfiltration. On the host level, Check Point utilises NVIDIA DOCA Argus for direct memory access visibility across all active processes within the AI environment. This allows the detection and mitigation of attacks or malicious workloads, including threats hidden within publicly obtained large language models.

David Reber, Chief Security Officer at NVIDIA, emphasised the joint initiative:

"Security is essential for the next generation of AI infrastructure. NVIDIA is working with Check Point to integrate BlueField acceleration and the NVIDIA DOCA Argus runtime security framework into the AI Cloud Protect platform to help enterprises deploy AI confidently."

Performance and deployment

AI Cloud Protect will extend support to the upcoming NVIDIA BlueField-4 DPU, which is expected to provide up to six times the compute power and double the network throughput over previous generations, facilitating secure and scalable AI operations. The product is currently being piloted by financial services firms and strategic partners including World Wide Technology.

Chris Konrad, Vice President, Global Cyber at World Wide Technology, described the value proposition for enterprises:

"As enterprises build AI server factories at scale, the combination of Check Point's AI Cloud Protect and NVIDIA BlueField acceleration delivers enterprise-grade protection for sensitive AI workloads from model training to inference without compromising the performance modern AI applications demand."

Broader security portfolio

Check Point's AI security approach extends beyond AI Cloud Protect, including other products such as CloudGuard Web Application Firewall, which incorporates Lakera capabilities to defend against threats like prompt injection and jailbreaking in enterprise AI applications. The Infinity GenAI Protect tool is designed to oversee the use of generative AI tools within organisations, enforcing security policies to limit data leakage and support compliance needs.

The set of solutions offered by Check Point is intended to secure the full spectrum of the AI pipeline, encompassing cloud infrastructure, enterprise applications, and end-user activities. This approach aims to provide what the company describes as end-to-end security for enterprises as they deploy and scale AI technologies.

Industry challenges

Security for AI factories involves addressing broad attack surfaces, with risks stemming from extensive data flows and high-performance computing pipelines. New threat types, including prompt injection, jailbreaking, and data leakage from unmonitored AI use, have been identified as major concerns. Check Point claims AI Cloud Protect addresses these issues through network, host, and application-level defence, NVIDIA hardware-accelerated security, and scalable security orchestration.

AI Cloud Protect is now available to enterprises deploying AI infrastructure, with the intention of securing sensitive data and intellectual property across both private and cloud-based AI deployments.