F5 and Telenor Advance Sovereign AI Services with NVIDIA-Powered Infrastructure
F5, a global provider of application delivery and security solutions, has announced that Telenor AI Factory has selected F5 BIG-IP Next for Kubernetes to serve as an intelligent AI service plane for its NVIDIA-powered sovereign artificial intelligence (AI) infrastructure. The collaboration aims to help organizations across Norway and Europe access secure, high-performance AI services while maintaining control over data processing, storage, access, and governance.
As businesses and governments accelerate AI adoption, digital sovereignty is becoming an important consideration in infrastructure planning. Organizations increasingly need assurances that sensitive information remains under appropriate jurisdictional control and that AI workloads operate within established security and compliance requirements.
Telenor AI Factory addresses these requirements by combining Telenor’s national infrastructure and connectivity capabilities with NVIDIA-powered AI computing infrastructure and F5’s traffic management, security, and policy enforcement technologies. The initiative is designed to provide a foundation for delivering AI services to Norwegian and European enterprises, including organizations operating in highly regulated industries.
The collaboration also reflects a broader industry shift toward building AI infrastructure that combines computing performance with operational controls, tenant isolation, and real-time visibility.
Strengthening Sovereign AI Infrastructure
Sovereign AI involves more than locating computing infrastructure within a particular country. It also requires control over how data moves through systems, how AI workloads interact with applications, how customers are separated from one another, and how security policies are enforced.
For enterprises operating in regulated sectors, these requirements can influence where AI applications are deployed and which infrastructure providers they can use.
Telenor AI Factory is designed to help Norwegian and European organizations access advanced AI capabilities while maintaining greater control over their digital operations. By integrating F5’s technology into its infrastructure, the company aims to establish a service management layer that governs AI-related traffic and enforces policies across connected applications and services.
F5 BIG-IP Next for Kubernetes is intended to provide centralized controls for traffic entering and leaving AI environments, application programming interface (API) interactions, operational telemetry, and inference-related communications.
These capabilities are designed to help service providers apply consistent security and governance measures as they introduce new AI services and expand their customer base.
F5 BIG-IP Next for Kubernetes and NVIDIA BlueField-3
The deployment uses F5 BIG-IP Next for Kubernetes on NVIDIA BlueField-3 data processing units (DPUs). This architecture is designed to support traffic management and selected networking and security functions while separating those operations from the main host computing workloads.
The approach can help preserve host central processing unit (CPU) resources for AI applications while strengthening the boundaries between infrastructure operations and individual tenants.
Several capabilities support Telenor AI Factory’s objectives.
Secure multi-tenancy and network isolation: The platform enables multiple customers to share sovereign GPU infrastructure while maintaining separation between their environments. This helps address risks associated with tenant interference, inconsistent security policies, and unauthorized data exposure.
Policy-based traffic control: The system supports policy enforcement across ingress and egress traffic, API interactions, and operational telemetry. These controls are intended to help align network activity with sovereignty, security, and compliance requirements.
Real-time observability and telemetry: Operational visibility enables service teams to monitor traffic, investigate activity, and generate information needed for auditing and compliance reporting. These capabilities can help organizations maintain a clearer understanding of how AI services operate.
Planned token-aware governance: As commercial AI services expand, Telenor plans to introduce additional controls for monitoring token usage and enforcing AI consumption policies. These capabilities are intended to support entitlement management, reporting, and billing processes.
Planned inference-aware traffic optimization: Future enhancements are expected to improve traffic management across changing AI inference workloads, helping service providers use available infrastructure efficiently while maintaining consistent performance.
The planned capabilities will extend the service plane as customer requirements and commercial AI workloads evolve.
Extending Digital Sovereignty Across AI Operations
Digital sovereignty increasingly depends on how organizations manage the complete flow of information through their AI environments.
Even when computing infrastructure is located within a particular jurisdiction, organizations must consider how customer data, system telemetry, API requests, and operational policies are handled. These considerations become more complex when multiple customers share computing resources or when AI services interact with external applications.
F5 BIG-IP Next for Kubernetes is designed to provide policy enforcement within the data path, allowing Telenor AI Factory to apply controls directly as traffic moves through its infrastructure.
This architecture supports consistent management of incoming and outgoing connections, API interactions, observability data, and AI inference traffic.
For organizations with strict governance requirements, these controls can help strengthen operational oversight and support the implementation of security policies across connected services.
The approach also provides a foundation for introducing additional AI-specific capabilities without requiring the entire infrastructure architecture to be redesigned as new services become available.
Preparing for Commercial AI Growth
As AI infrastructure moves from experimental deployments toward commercial services, providers must manage increasingly complex workloads and customer requirements.
AI inference, the process through which trained models generate responses or perform tasks, can involve changing traffic patterns and varying levels of resource consumption. Service providers therefore need mechanisms to monitor activity, enforce usage policies, and manage demand across shared infrastructure.
Telenor plans to extend its AI service plane with capabilities such as token-aware usage visibility, circuit breaking, and more advanced inference-aware traffic management.
Token-aware controls are intended to provide greater visibility into AI consumption and help providers implement usage limits, manage customer entitlements, and support billing and reporting workflows.
Circuit breaking can help protect systems when services encounter excessive demand or operational problems by interrupting or limiting certain requests under defined conditions.
Inference-aware traffic management is intended to improve how AI requests are distributed and handled as workloads change.
Running selected traffic-management functions on NVIDIA BlueField-3 DPUs can also help isolate networking and security operations from host workloads. This may preserve CPU resources for AI processing and help reduce data-path bottlenecks as service demand increases.
The effectiveness of these capabilities will depend on implementation details, workload characteristics, and operational requirements as the platform develops.
Leadership Perspectives on the Collaboration
Kaaren Hilsen, chief executive officer of Telenor AI Factory, said the company’s objective is to provide Norwegian and European enterprises with trusted, secure, and high-performance AI services.
Hilsen emphasized that sovereignty must extend beyond the physical location of infrastructure to include how AI services are operated and governed. She explained that F5 provides a policy and traffic-control layer that Telenor can build upon as it expands its services for customers.
Kunal Anand, chief product officer at F5, highlighted the importance of real-time governance as AI factories develop into commercial service platforms.
Anand said providers need mechanisms to govern AI traffic, tenant environments, telemetry, and policies as their services scale. He described BIG-IP Next for Kubernetes as a foundation for expanding sovereign AI services, including a future path toward token-aware governance and DPU acceleration.
Ash Bhalgat, senior director of AI Ecosystem and Marketing at NVIDIA, said combining NVIDIA-powered AI infrastructure with F5 BIG-IP Next for Kubernetes on BlueField-3 can help service providers deliver high-performance AI services while maintaining the governance, control, and security required for sovereign AI.
The statements reflect the companies’ shared focus on combining advanced AI computing with the operational controls needed to support enterprise deployments.
About Telenor AI Factory and F5
Telenor AI Factory is designed to provide Norwegian and European organizations with access to secure, high-performance AI infrastructure while supporting data sovereignty and operational control. Its approach combines Telenor’s infrastructure and connectivity capabilities with NVIDIA-powered computing and technologies from partners such as F5.
F5, Inc. (NASDAQ: FFIV) provides application delivery and security solutions designed to help organizations deliver and protect applications and APIs across on-premises infrastructure, cloud environments, edge deployments, and hybrid or multicloud systems.
The company’s Application Delivery and Security Platform brings together capabilities intended to support application performance, availability, and security. F5 works with enterprises and technology partners to help organizations deliver digital services across complex infrastructure environments.
Additional information about F5 and its technologies is available at https://www.f5.com/.
The collaboration between F5 and Telenor AI Factory highlights the growing importance of integrating security, traffic management, and policy enforcement into AI infrastructure. By combining NVIDIA-powered computing with an intelligent service plane, the companies aim to support the development of sovereign AI services that can scale while maintaining operational control, security, and accountability.
Future enhancements, including token-aware governance and advanced inference traffic optimization, are planned as Telenor AI Factory expands its commercial offerings. Their availability and benefits will depend on further development and implementation.
Source Link: https://www.businesswire.com/