Arrcus, a provider of distributed networking infrastructure solutions, has announced a proof-of-concept (PoC) initiative with TELUS to evaluate the deployment of the Arrcus Inference Network Fabric (AINF) as the networking foundation for sovereign, distributed artificial intelligence inferencing across Canada. The initiative is aimed at enabling secure, low-latency AI services for mission-critical applications while ensuring that sensitive data remains within national borders.
As AI adoption accelerates across industries, computing architectures are shifting from centralized model training toward distributed inferencing, where intelligence is delivered closer to the point of data creation and decision-making. This transition is driving demand for advanced networking platforms capable of supporting real-time AI workloads across geographically dispersed environments.
The PoC will explore how AINF can help TELUS deliver distributed AI services across public safety networks, emergency response systems, enterprise environments and government operations. By integrating policy-based intelligence directly into the network, the platform is designed to route AI requests dynamically based on factors such as latency requirements, data residency rules, model availability, security policies and compute capacity.
For organizations operating in regulated sectors, data sovereignty has become a strategic priority. Applications such as real-time video analytics, predictive threat monitoring, emergency response coordination and critical infrastructure management require AI systems that combine ultra-low latency with stringent data protection measures. The collaboration seeks to demonstrate how sovereign AI services can be deployed at scale while maintaining regulatory compliance and operational performance.
According to TELUS, the initiative supports its broader vision of building trusted AI infrastructure capable of serving public institutions, businesses and communities across Canada. The networking fabric is intended to provide secure connectivity, deterministic performance and policy-driven governance for distributed AI workloads operating across multiple locations.
At the heart of the solution is Arrcus AINF, an AI-aware networking platform developed specifically for distributed inferencing environments. Unlike conventional network architectures, the platform evaluates business and operational policies in real time, directing inference traffic to the most appropriate location while maintaining performance and compliance requirements.
The proof-of-concept also incorporates advanced infrastructure technologies, including NVIDIA BlueField-3 data processing units and NVIDIA Spectrum-4 Ethernet switches. Together, these technologies create a secure and scalable environment spanning edge locations, data centers and cloud infrastructure. The platform additionally supports integration with leading AI frameworks and orchestration tools, enabling flexible deployment across diverse environments.
The initiative represents an important step in the evolution of AI infrastructure, demonstrating how intelligent networking can become a critical enabler of secure, sovereign and distributed artificial intelligence services. By combining advanced networking intelligence with scalable compute resources, Arrcus and TELUS aim to establish a framework capable of supporting the next generation of AI-driven applications while strengthening national digital resilience and data sovereignty.
Arrcus is a leading provider of high-performance networking solutions, delivering advanced routing, switching, and AI networking technologies that help enterprises and service providers build scalable, resilient, and secure digital infrastructure. Its flagship ACE platform, powered by the ArcOS network operating system, provides a flexible, software-driven alternative to traditional networking architectures across data center, core, and edge environments.
To address the evolving requirements of artificial intelligence workloads, Arrcus offers purpose-built Ethernet fabric solutions for both AI training and inference. Its networking platforms are designed to support high-bandwidth data movement, low-latency communications, and large-scale distributed computing. The company’s Arrcus Inference Network Fabric (AINF) extends intelligent, policy-based networking across edge, data center, and cloud environments, improving inference performance, reducing token latency, enhancing throughput, and lowering the cost of AI operations. These capabilities also support sovereign AI deployments, enabling organizations to modernize network infrastructure for next-generation AI applications. Headquartered in San Jose, California, Arrcus serves customers worldwide across enterprise, cloud, telecommunications, and AI-driven industries.
