Arrcus on AI, Inference and Network Fabric

Why It Matters

The AINF delivers the latency and power efficiency essential for real‑time AI decisions, positioning Arrcus as a critical enabler of next‑generation edge AI and data‑center interconnects.

Key Takeaways

• Arrcus achieved three‑fold booking growth in 2025 year
• New Arcus Inference Network Fabric targets ultra‑low latency AI inference
• Fabric unifies data‑center and inter‑data‑center networking for consistency
• Power‑efficient routing addresses AI’s massive energy consumption concerns
• Growth driven by core networking, not solely LLM market hype

Summary

Arrcus highlighted its explosive 2025 performance, reporting a three‑fold increase in bookings and expanding adoption across data‑center providers, telecom carriers, and large enterprises. The company introduced the Arcus Inference Network Fabric (AINF), a purpose‑built layer designed to deliver ultra‑low latency, high throughput, and power‑optimized connectivity for AI inference workloads, bridging both top‑of‑rack switches inside data centers and inter‑data‑center links.

Key insights include AINF’s focus on “time‑to‑first‑token” performance for edge‑critical applications such as autonomous vehicles and industrial rigs, and its policy‑driven routing that enforces compliance, data sovereignty, and energy constraints. Arrcus positions the fabric as a unified operating environment that simplifies security and operational policies across the entire network stack.

Executive Steve emphasized that the company’s growth “had very little to do with the LLM boom,” likening today’s networking evolution to the early internet era, and underscored the strategic importance of power‑efficient routing as AI workloads threaten unprecedented electricity consumption.

The announcement signals Arrcus’s ambition to become the foundational infrastructure for real‑time AI inference, offering enterprises a differentiated solution that remains valuable regardless of AI hype cycles, and potentially reshaping data‑center and edge networking economics.

Original Description

AI growth is reshaping cloud and data center infrastructure — but inference, latency and power efficiency are now taking center stage.

In this discussion, Shekar Ayyar, Chairman and CEO of Arrcus, explains how the company tripled bookings growth in 2025 and where it plays across data centers, interconnect and global routing environments. He outlines how Arrcus supports top-of-rack, spine-and-leaf architectures, as well as data center interconnects that link distributed pools of AI capacity.

Ayyar also introduces the Arrcus Inference Network Fabric (AI&F), designed to deliver low latency, high throughput and optimized power consumption for AI inference nodes. The conversation explores how networking supports agentic AI, industry 4.0 use cases, sovereignty requirements and compliance boundaries — and why infrastructure decisions increasingly shape AI performance at the edge.

Featuring Shekar Ayyar of Arrcus in conversation with Steve Saunders.

#AIInfrastructure #DataCenter #Networking #Arrcus