Fortinet is further expanding its partnership with Nvidia to better secure AI environments. The security vendor is integrating its FortiAIGate platform with Nvidia’s AI infrastructure and software, with the goal of giving organizations greater control over AI workloads, autonomous AI agents, and sensitive data within data centers and cloud environments.
According to Fortinet, there is a growing need among businesses for security solutions that can monitor AI processes without compromising performance or scalability. The combination of FortiAIGate with Nvidia technology is designed to provide real-time security for AI applications while meeting requirements for governance and data sovereignty.
Security for AI Runtime
FortiAIGate focuses on securing AI environments during runtime. The platform monitors traffic to and from large language models and AI agents. Among other things, the solution is designed to detect attacks via malicious prompts and prevent the unintended sharing of sensitive information.
Fortinet states that the integration utilizes Nvidia Blackwell and Hopper GPUs in combination with Nvidia Dynamo, a framework for distributed inference. This enables the security layer to operate with minimal additional latency, even in large-scale AI applications.
Fortinet positions FortiAIGate as a zero-trust solution for AI infrastructures. The platform monitors all interactions between users, devices, and AI models and logs suspicious prompts and AI responses for analysis.
Additionally, the solution is designed to prevent AI models from generating unwanted or harmful content. According to Fortinet, the system learns the context of individual AI models, enabling it to detect anomalous behavior more quickly.
Focus on data sovereignty
A key aspect of the partnership revolves around data sovereignty and local control over AI systems. Fortinet offers options to fully self-host AI environments, reducing organizations’ dependence on foreign AI vendors.
According to the company, this helps organizations comply with local regulations such as the GDPR and keep training and inference data within national borders. FortiAIGate utilizes Nvidia Nemotron security models for this purpose.
The solution supports deployments in data centers, cloud environments, edge locations, and hybrid infrastructures. FortiAIGate will be available as a physical appliance with GPU support, as well as a virtual appliance or container environment on Nvidia-certified systems.
Fortinet also states that GPU-accelerated security can manage energy and hardware usage more efficiently than traditional CPU-based security solutions. This should reduce the operational costs of AI security, especially in large-scale multi-tenant AI environments.