Check Point & NVIDIA launch AI Cloud Protect for secure deployments

Check Point has announced a partnership with NVIDIA to deliver an integrated security solution designed for AI factory environments.

The new solution, called AI Cloud Protect, is now available for on-premises enterprise deployments. It is positioned to safeguard AI model development, agentic AI applications and inference workloads, promising zero impact on performance. The solution has been validated on NVIDIA RTX PRO Servers and is built on NVIDIA BlueField technology, allowing enterprises to deploy AI infrastructure securely in both data centre and cloud settings.

Addressing cyber threats

The growing adoption of AI within enterprise networks has prompted significant security concerns. According to Check Point, over half of enterprise networks are now leveraging AI tools, turning them into prime targets for cyberattacks.

Company data reveals that one in every 80 generative AI prompts exposes sensitive data. A recent Gartner report found that 32% of organisations reported AI attacks involving prompt manipulation, and 29% experienced attacks on their generative AI infrastructure in the past year.

As enterprises accelerate the deployment of AI, there is heightened pressure to protect the end-to-end AI pipeline, from development through to production and commercial implementation. AI Cloud Protect is intended to secure the infrastructure powering AI model development and deployment, addressing these needs directly.

"As enterprises race to build AI-driven innovation, they can't afford blind spots," said Nataly Kremer, Chief Product Officer at Check Point. "With NVIDIA, we're making AI factories secure by design-protecting models, data, and infrastructure without slowing innovation."

AI Cloud Protect operates on NVIDIA BlueField-3 Data Processing Units (DPUs), delivering security across the stack without using additional CPU or GPU resources. This approach aims to eliminate performance trade-offs commonly experienced with traditional security solutions. The product uses network-level protection to reduce risks of data poisoning and model exfiltration through prevention of unauthorised access. Internally, it utilises the direct memory access provided by NVIDIA DOCA Argus, offering visibility into all processes running on AI nodes. This capability allows Check Point to detect and prevent malicious host-level processes, including threats arising through large language models acquired from public sources.

David Reber, Chief Security Officer at NVIDIA, commented on the collaboration:

"Security is essential for the next generation of AI infrastructure," said David Reber, chief security officer at NVIDIA. "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."

Future capabilities and further protection

Check Point confirmed that AI Cloud Protect will be supported on upcoming NVIDIA BlueField-4 DPUs, which will offer up to six times the compute power and double the network throughput for inference and storage processing over previous versions. This increased capacity is aimed at supporting scalable and secure AI operations at high speed.

Beyond AI Cloud Protect, Check Point has also launched other tools to address AI security needs. The CloudGuard Web Application Firewall (WAF) integrates Lakera capabilities to defend against threats including prompt injection and jailbreaking. Infinity GenAI Protect allows companies to monitor generative AI application use, apply policies to prevent sensitive data leakage, and ensure compliance. Both offerings aim to provide comprehensive protection across infrastructure, applications, and users throughout the AI value chain.

Pilot deployments

AI Cloud Protect is now being piloted with financial services customers and strategic partners, including World Wide Technology (WWT). Initial implementations focus on securing AI data centres that support large language model development and prompt-based applications with an emphasis on protecting sensitive data and intellectual property. NVIDIA RTX PRO Server validation gives enterprises the option to deploy AI Cloud Protect into their AI infrastructure immediately.

Chris Konrad, Vice President, Global Cyber, World Wide Technology, commented on the deployment process:

"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," said Chris Konrad, Vice President, Global Cyber, World Wide Technology.

AI challenges and solutions

AI factories present a wide attack surface due to increased compute capabilities and vast data pipelines. New threats such as prompt injection, jailbreaking, and model poisoning are on the rise, while ungoverned employee use of AI poses a risk of data leakage. Check Point's AI Cloud Protect applies a unified security approach, providing protection at the network, host, and application layers. The solution is designed for full security orchestration across thousands of AI nodes, automated to manage the complexity of large-scale enterprise AI deployments.

Check Point's collaboration with NVIDIA underlines a trend where cybersecurity providers and technology firms are increasingly focusing on the intersection of AI infrastructure and operational security in response to evolving threats and enterprise adoption of AI.