Akamai and NVIDIA Embed Zero Trust Security Inside AI Factories

AI factories are scaling faster than they can be secured, and traditional host-based controls cannot keep pace with GPU-driven workloads.

As enterprises adopt agentic AI, the infrastructure powering training pipelines, inference services, and autonomous agents has become a critical asset. Yet every additional security layer running on the host risks slowing down the GPUs, CPUs, and storage processors that AI factories depend on. Akamai and NVIDIA are addressing this gap by embedding Zero Trust segmentation directly into the silicon fabric of the AI factory itself.

What was announced

Akamai (NASDAQ: AKAM) and NVIDIA announced an expansion of their security collaboration to bring Akamai Guardicore Segmentation to the NVIDIA Vera BlueField-4 STX storage architecture, powered by the NVIDIA DOCA software platform. The integration builds on the architecture agreement the two companies introduced in February and is designed to layer Zero Trust architecture into the AI factory, protecting data, context memory, and autonomous agents.

The combined solution enables AI factory operators to enforce workload-aware segmentation, monitor agent behavior, and contain threats at the infrastructure layer without consuming GPU, CPU, or storage cycles.

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. 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

Why this matters

For CIOs, CISOs, and infrastructure leaders, the integration removes a long-standing trade-off between fast AI and secure AI. Host-based security agents add overhead that erodes the value of accelerated computing, while perimeter-only controls cannot see how AI workloads communicate inside the cluster. By moving enforcement into the BlueField data path, security becomes a property of the infrastructure rather than an additional product layer.

Procurement and platform teams gain a model where research environments are explicitly separated from production inference, preprocessing nodes are restricted to defined datasets, and a compromised workload is confined to a small, identified segment without disrupting the rest of the AI factory.

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

Technical details

• Visibility: Akamai Guardicore Segmentation continuously maps communication relationships across data centers, cloud, Kubernetes, and edge systems, observing training pipelines, inference services, data ingestion, and orchestration platforms agentlessly.
• Policy: Rules are defined by workload identity, application context, and runtime behavior, not by static network addresses. Pods can scale and services evolve without weakening the policy boundary.
• Enforcement: NVIDIA DOCA applies policies in BlueField-4 silicon, in the data path, at line speed, running segmentation, telemetry, anomaly detection, and isolation inside the infrastructure fabric.
• Containment: When a workload is compromised, the blast radius is limited to an identified segment while the rest of the AI factory continues to operate.
• Availability: Akamai Guardicore Segmentation integrated with NVIDIA BlueField and DOCA is expected in the second half of 2026; integration with NVIDIA Vera BlueField-4 STX is expected on storage and infrastructure partner platforms in the first half of 2027.

Softprom and Akamai

Softprom is the official distributor of Akamai. Our team supports enterprise customers and partners in designing Zero Trust architectures, deploying Akamai Guardicore Segmentation, and aligning security with high-performance AI infrastructure projects.

This content was prepared as part of the Softprom DistriFlow project — an automated system for monitoring and adapting vendor news. Original source: original article.