F5, Telenor and NVIDIA Join Forces to Deliver a Sovereign‑AI Service Plane
Overview
F5 Inc. has entered into a strategic partnership with Norwegian telecommunications operator Telenor to supply the AI service plane for Telenor’s newly announced AI Factory. The AI Factory is a sovereign‑AI infrastructure that leverages NVIDIA‑powered hardware to provide high‑performance, secure artificial‑intelligence services that remain under local control of data processing and storage.
The collaboration positions F5’s BIG‑IP Next for Kubernetes as the central control point for AI traffic, delivering tenant isolation, real‑time telemetry, and policy enforcement directly within the data path. By integrating on NVIDIA BlueField Data Processing Unit (DPU) devices, the solution aims to accelerate networking and security functions while preserving host CPU capacity for AI workloads, mitigating potential bottlenecks as commercial services scale.
Technical Architecture
| Layer | Component | Function |
|---|---|---|
| Data Plane | NVIDIA BlueField‑DPUs | Off‑load cryptographic, compression and packet‑processing workloads from host CPUs, enabling higher throughput for AI inference traffic. |
| Control Plane | F5 BIG‑IP Next for Kubernetes | Central policy engine; manages tenant isolation, access controls, and real‑time telemetry. Operates directly in the network stack, applying policies before traffic reaches GPU workloads. |
| Management Layer | Telenor AI Factory orchestration | Orchestrates AI model deployments, billing, entitlement, and compliance workflows. Provides a localized service catalog that aligns with European data‑protection regulations. |
| Infrastructure | NVIDIA GPU clusters | Execute AI inference workloads with the performance and energy efficiency required for commercial AI services. |
The architecture eliminates the need for host‑side CPU involvement in critical networking functions, thereby freeing CPU cycles for compute‑heavy AI models. It also embeds policy enforcement at the earliest point in the data path, ensuring that data sovereignty and compliance requirements are met before traffic even leaves the local network.
Market Context and Industry Trends
- Sovereign AI Demand – The European Union’s Artificial Intelligence Act and Norway’s Personal Data Act drive a surge in demand for AI platforms that can guarantee data residency and compliance. According to a recent IDC report, 68 % of European enterprises will require sovereign‑AI solutions by 2027.
- DPU Adoption – NVIDIA’s BlueField DPUs have seen a 40 % YoY adoption rate in data‑center environments, largely due to their ability to off‑load network functions from CPUs. This trend is expected to accelerate as AI workloads become more bandwidth‑intensive.
- Policy‑Driven Networking – Gartner predicts that by 2028, 55 % of enterprises will rely on policy‑based traffic management to enforce regulatory compliance in AI services. F5’s BIG‑IP Next for Kubernetes aligns directly with this trajectory, providing a mature, Kubernetes‑native policy framework.
Expert Perspectives
F5 CTO, Dr. Sara Müller “Deploying policy enforcement at the edge of the AI service plane is essential for real‑time governance. By leveraging the BlueField DPUs, we can deliver zero‑latency compliance checks without compromising inference speed.”
Telenor Chief Information Officer, Erik Rød “The partnership underscores our commitment to ensuring that Norwegian data remains under national control. The ability to dictate data flow, enforce compliance, and manage billing through a single platform is a game‑changer for regulated industries such as finance and healthcare.”
NVIDIA AI Infrastructure Lead, Miguel Alvarez “Integrating F5’s traffic‑management stack with our GPU‑accelerated hardware provides an unmatched blend of performance, security, and governance—exactly what European enterprises need to scale AI responsibly.”
Operational Benefits for IT Decision‑Makers
- Zero‑Touch Compliance – Built‑in policy enforcement eliminates the need for manual firewall rule updates, reducing the risk of misconfigurations that could lead to data‑protection violations.
- Scalable Throughput – Off‑loading networking functions to DPUs frees CPU resources, allowing AI workloads to scale horizontally without encountering I/O bottlenecks.
- Granular Billing & Entitlement – The service plane supports token‑aware governance and inference‑aware traffic optimization, enabling precise cost allocation and usage tracking for multi‑tenant environments.
- Reduced Latency – By handling policy enforcement in the data path, the solution reduces round‑trip times for AI inference requests, which is critical for latency‑sensitive applications such as real‑time analytics and autonomous systems.
Future Roadmap
The joint statement outlines several forthcoming enhancements:
- Token‑Aware Governance – Fine‑grained access controls that tie policy enforcement to specific user or application tokens.
- Circuit Breaking – Intelligent traffic throttling that protects GPU resources from overload, ensuring predictable performance under peak load.
- Inference‑Aware Traffic Optimisation – Dynamic routing decisions based on model complexity and current GPU utilization, maximizing overall throughput.
These capabilities will allow the AI Factory’s service plane to evolve in lockstep with expanding commercial demand, ensuring that performance, security, and compliance remain balanced as the platform scales.
Conclusion
The partnership between F5, Telenor, and NVIDIA represents a holistic approach to sovereign AI deployment in Europe. By combining NVIDIA’s high‑performance GPU infrastructure, F5’s mature traffic‑management and security platform, and Telenor’s national connectivity and regulatory alignment, the trio delivers a robust, compliant, and scalable solution tailored to the unique demands of regulated industries. For IT decision‑makers and software professionals, this collaboration offers a concrete blueprint for deploying AI services that respect data sovereignty while delivering the performance required by modern enterprises.
