Akamai & NVIDIA extend AI factory security integration

Akamai and NVIDIA have expanded their security collaboration to integrate Akamai Guardicore Segmentation with NVIDIA's Vera BlueField-4 STX storage architecture, using the NVIDIA DOCA software platform to bring Zero Trust security controls directly into AI factory infrastructure.

The integration is designed to protect AI workloads, data, context memory and autonomous software agents that increasingly underpin enterprise artificial intelligence deployments. It aims to provide security enforcement at the infrastructure layer without consuming processing resources needed for AI training and inference workloads.

The move reflects growing concerns about securing large-scale AI environments as organisations accelerate investment in AI factories and high-performance computing infrastructure.

The combined architecture is intended to allow operators to apply workload-aware segmentation, monitor AI agent behaviour and contain cyber threats across AI environments.

Under the arrangement, Akamai Guardicore Segmentation provides visibility across workloads, applications and data operating in hybrid environments spanning data centres, cloud infrastructure, Kubernetes deployments and edge systems.

The platform maps communications between systems and uses workload identity, application context and runtime behaviour to establish security policies. Rather than relying on static network addresses, policies are tied to the characteristics and behaviour of workloads.

NVIDIA Vera BlueField-4 STX and the DOCA platform then enforce those policies within the infrastructure fabric. Security controls operate in the data path and within dedicated silicon rather than on host systems.

This approach is intended to reduce the performance overhead traditionally associated with host-based security tools in AI environments, where computing resources are often heavily utilised.

The companies said the architecture is based on a model in which visibility and policy definition are separated from enforcement.

Guardicore Segmentation continuously maps communication relationships across AI training pipelines, inference services, orchestration platforms and data ingestion systems. The system operates without requiring agents on protected workloads.

Security policies are then established around workload identities and permitted communications. For example, a data preprocessing node can be restricted to specific datasets and training services while being prevented from accessing unrelated systems.

NVIDIA DOCA applies those policies through BlueField-4 infrastructure components. Functions such as segmentation, telemetry, anomaly detection and system isolation are executed within the infrastructure layer.

The companies said this design allows compromised workloads to be isolated without affecting the wider AI environment. If a threat is detected, the affected workload can be contained within a defined segment while other systems continue operating.

The collaboration builds on a broader security architecture relationship announced by the companies last year as AI infrastructure deployments continue to expand.

Industry demand for AI factories has increased rapidly as organisations deploy larger AI models and autonomous software systems. Those environments often require significant investments in graphics processing units, storage systems and networking infrastructure.

As enterprises scale AI deployments, security has become a growing concern. AI environments often involve large datasets, distributed computing resources and autonomous agents that create additional attack surfaces for cyber threats.

"AI factories are becoming critical assets that must be designed for containment, especially as frontier LLM-driven attacks increase the speed and scale of cyber threats," said Ofer Wolf, Senior Vice President, Enterprise Security, Akamai.

"In environments where every clock cycle matters, traditional host-based security tools behave like a speed bump on a racetrack. By moving workload-aware segmentation onto NVIDIA Vera BlueField-4 STX and DOCA, we are enforcing Zero Trust at the speed of AI workloads themselves, helping organizations contain threats before they spread across high-performance environments," said Wolf.

Akamai said the Guardicore Segmentation integration with NVIDIA BlueField and NVIDIA DOCA is expected to become available during the second half of 2026.

Integration with NVIDIA Vera BlueField-4 STX is expected to be available through storage and infrastructure partner platforms during the first half of 2027.

"Data is the foundation of agentic AI factories, powering the intelligence behind autonomous decision-making and making robust protection more critical than ever for enterprises," said Kevin Deierling, Senior Vice President, Networking, NVIDIA.

"The Akamai Guardicore enterprise security platform and NVIDIA Vera BlueField-4 STX brings a Zero Trust layer directly into the infrastructure fabric, helping protect enterprise data by intelligently controlling how AI workloads communicate at scale," said Deierling.