CAMBRIDGE, Mass., June 02, 2026 (GLOBE NEWSWIRE) -- Akamai (NASDAQ:AKAM) is collaborating with NVIDIA to bring an advanced security architecture into AI factories — the foundation of the agentic era.
The two companies today 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 collaboration is designed to layer Zero Trust architecture into the AI factory itself, protecting the data, context memory, and autonomous agents that increasingly run modern enterprises.
This novel security integration will enable AI factory operators to enforce workload-aware segmentation, monitor agent behavior, and contain threats at the infrastructure layer, operating at the speed of accelerated computing without taxing the GPU, CPU, or storage cycles that AI workloads depend on.
"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 at 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."
"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 at 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."
Security at the speed of accelerated computing
AI factories are being built faster than they can be secured. Until now, there has been a trade-off between fast AI and secure AI. But that compromise that's no longer tenable.
The expanded Akamai–NVIDIA integration, which builds on the architecture agreement that two companies introduced last February, is designed to remove that trade-off.
Akamai Guardicore Segmentation, which protects some of the world's largest and most sensitive organizations, provides the intelligence layer, continuously mapping how workloads, applications, and data interact across hybrid environments that include data centers, cloud infrastructure, Kubernetes clusters, and edge systems. Policies are defined by workload identity, application context, and runtime behavior, not by static network addresses. Visibility extends across the full lifecycle of AI workloads, surfacing abnormal patterns and unauthorized access to sensitive data.
NVIDIA Vera BlueField-4 STX, programmable through NVIDIA DOCA, provides the threat detection and enforcement layers in silicon. Security policies are applied in the data path at line speed, inside the infrastructure fabric rather than on the host. Enforcement moves closer to the workload itself, so it won't trip up the GPUs, CPUs, and storage processors that AI factories depend on. Together, the two layers establish identity-based Zero Trust as a property of the infrastructure rather than an additional product.
How the integration works
The combined solution operates on the principle that intelligence must precede enforcement:
- Visibility. Akamai Guardicore Segmentation continuously maps communication relationships across data centers, cloud, Kubernetes, and edge systems. Its agentless architecture observes AI workloads, including training pipelines, inference services, data ingestion systems, and orchestration platforms, without interfering with them.
- Policy. Workload identity, application context, and runtime behavior are used to define explicit communication policies. A preprocessing node may access a dataset and a training service but nothing beyond that scope. Research environments are explicitly separated from production inference. Pods can scale and services can evolve without weakening the policy boundary.
- Enforcement. NVIDIA DOCA applies those policies in BlueField-4 silicon, in the data path, at line speed. Security functions, including segmentation, telemetry, anomaly detection, and isolation of compromised systems, run inside the infrastructure fabric rather than on the host.
- Containment. When a workload is compromised, the blast radius is limited to a small, identified segment of the environment. The rest of the AI factory continues operating uninterrupted.
Availability
Akamai Guardicore Segmentation integrated with NVIDIA BlueField and NVIDIA DOCA is expected to be available in the second half of 2026 for implementing workload-aware segmentation in AI factories. Akamai's integration with NVIDIA Vera BlueField-4 STX is expected to be available on storage and infrastructure partner platforms in the first half of 2027.
About Akamai
Akamai is the cybersecurity and cloud computing company that powers and protects business online. Our market-leading security solutions, superior threat intelligence, and global operations team provide defense in depth to safeguard enterprise data and applications everywhere. Akamai's full-stack cloud computing solutions deliver performance and affordability on the world's most distributed platform. Global enterprises trust Akamai to provide the industry-leading reliability, scale, and expertise they need to grow their business with confidence. Learn more at akamai.com and akamai.com/blog, or follow Akamai Technologies on X and LinkedIn.
Contacts
Akamai Media Relations
akamaipr@akamai.com
Akamai Investor Relations
invrel@akamai.com
This press release was published by a CLEAR® Verified individual.
