Securing NVIDIA AI cloud data centers with the next wave in real-time runtime security

The explosive growth of AI is transforming enterprises and cloud providers alike, creating unprecedented demand for secure, high-performance AI infrastructure. At NVIDIA GTC 2025, Check Point will share how it is integrating the NVIDIA Cybersecurity AI platform to secure the next generation of AI cloud infrastructure at scale.

 The Unique Security Challenges of AI Cloud Data Centers

AI Cloud Data Centers represent a fundamental shift from traditional data centers to accelerated computing. These AI Cloud Data Centers are sometimes called “AI factories” because they focus specifically on the creation, development, and delivery of AI capabilities that power innovation across industries.

AI Cloud Data Centers deal with the same requirements and attack vectors as other data centers, but with added challenges. They require massive computational resources, handle extremely valuable intellectual property, and process sensitive data at unprecedented scale. For cloud service providers, this creates a dual challenge: delivering the performance their customers demand while ensuring ironclad security for these critical workloads.

Enterprises are leveraging AI to transform core business operations, from detecting fraud to designing new products to driving earlier disease detection. Across all sectors, the security of AI infrastructure has a direct impact on business outcomes and customer trust. Stakes are particularly high in AI cloud environments. Training runs can last weeks or months using petabytes of sensitive data, and the resulting models often represent substantial investment and competitive advantage. Similarly, inference processes, where trained AI models conclude from new data, must be protected. A security breach doesn’t just compromise their data – it can erase months of work and put essential intellectual property at risk.

AI Cloud Protect: Security without Compromising Performance

Through our The explosive growth of AI is transforming enterprises and cloud providers alike, creating unprecedented demand for secure, high-performance AI infrastructure. At NVIDIA GTC 2025, Check Point will share how it is integrating the NVIDIA Cybersecurity AI platform to secure the next generation of AI cloud infrastructure at scale.
The Unique Security Challenges of AI Cloud Data Centers
AI Cloud Data Centers represent a fundamental shift from traditional data centers to accelerated computing. These AI Cloud Data Centers are sometimes called “AI factories” because they focus specifically on the creation, development, and delivery of AI capabilities that power innovation across industries.
AI Cloud Data Centers deal with the same requirements and attack vectors as other data centers, but with added challenges. They require massive computational resources, handle extremely valuable intellectual property, and process sensitive data at unprecedented scale. For cloud service providers, this creates a dual challenge: delivering the performance their customers demand while ensuring ironclad security for these critical workloads.
Enterprises are leveraging AI to transform core business operations, from detecting fraud to designing new products to driving earlier disease detection. Across all sectors, the security of AI infrastructure has a direct impact on business outcomes and customer trust. Stakes are particularly high in AI cloud environments. Training runs can last weeks or months using petabytes of sensitive data, and the resulting models often represent substantial investment and competitive advantage. Similarly, inference processes, where trained AI models draw conclusions from new data, must be protected. A security breach doesn’t just compromise their data – it can erase months of work and put essential intellectual property at risk.
AI Cloud Protect: Security without Compromising Performance
Through our AI Cloud Protect solution, we are delivering a new standard for securing NVIDIA AI factories and cloud infrastructure. This solution enables cloud service providers and enterprises to offer robust security that matches the need for speed in AI, with zero impact on performance.
That is because all the security functions run natively, as part of the AI infrastructure on the NVIDIA BlueField-3 DPU, meaning that CPU and GPU resources can be fully dedicated to AI workloads. It’s an approach that transforms security from a potential bottleneck into a business enabler, helping cloud service providers (CSPs) and enterprises create secure cloud platforms, protecting their most valuable assets.
Comprehensive Protection for AI Development, AI Training and Inference
Check Point’s AI Security solutions address critical vulnerabilities throughout the AI development and training stages for CSPs and enterprises. Check Point AI Cloud Protect secures AI infrastructure from unauthorised access to reduce the risk of data poisoning and model exfiltration using network-level protection. At the host level, AI Cloud Protect leverages the unique direct memory access of NVIDIA DOCA App Shield to provide visibility into all running processes on the AI node. This enables Check Point to detect and prevent host-level malicious processes and workloads including those in publicly downloaded Large Language Models (LLMs), powered by NVIDIA BlueField DPUs.
Join Us at NVIDIA GTC 2025
Join us at GTC 2025 to discover how we are helping organisations confidently deploy and scale their AI initiatives. Learn how cloud providers and enterprises can use this solution to secure their AI infrastructure while accelerating innovation and growth.
Don’t miss this essential session on how to Power the Next Wave of AI with Real-Time Runtime Security [S71642] on Thursday March 20 at 2pm PDT (and on demand after) at GTC 2025. Check Point is redefining how organisations protect their AI infrastructures powered by NVIDIA Cybersecurity AI capabilities in an increasingly complex threat landscape.
solution, we are delivering a new standard for securing NVIDIA AI factories and cloud infrastructure. This solution enables cloud service providers and enterprises to offer robust security that matches the need for speed in AI, with zero impact on performance.

That is because all the security functions run natively, as part of the AI infrastructure on the NVIDIA BlueField-3 DPU, meaning that CPU and GPU resources can be fully dedicated to AI workloads. It’s an approach that transforms security from a potential bottleneck into a business enabler, helping cloud service providers (CSPs) and enterprises create secure cloud platforms, protecting their most valuable assets.

Comprehensive Protection for AI Development, AI Training and Inference

Check Point’s AI Security solutions address critical vulnerabilities throughout the AI development and training stages for CSPs and enterprises. Check Point AI Cloud Protect secures AI infrastructure from unauthorised access to reduce the risk of data poisoning and model exfiltration using network-level protection. At the host level, AI Cloud Protect leverages the unique direct memory access of  NVIDIA DOCA  App Shield to provide visibility into all running processes on the AI node. This enables Check Point to detect and prevent host-level malicious processes and workloads including those in publicly downloaded Large Language Models (LLMs), powered by NVIDIA BlueField DPUs.

Join Us at NVIDIA GTC 2025

Join us at GTC 2025 to discover how we are helping organisations confidently deploy and scale their AI initiatives. Learn how cloud providers and enterprises can use this solution to secure their AI infrastructure while accelerating innovation and growth.