Arrcus: Scaling the Network Fabric for the AI and Hyperscale Era

Headquartered in San Jose, California, Arrcus, Inc. is a pioneering, high-growth technology leader specializing in hyperscale networking software. Founded by industry veterans Keyur Patel and Derek Yeung, and led by CEO Shekar Ayyar, the company’s core mission is to democratize networking. By replacing rigid, monolithic, and proprietary hardware-locked systems with open, programmable, and cloud-native software, Arrcus delivers high-performance routing and switching at a fraction of traditional costs.

For software, systems, and network protocol engineers, launching a career at Arrcus means building the underlying virtualized architecture that powers modern data centers, 5G edge infrastructure, multi-cloud platforms, and the physical backbone of the global AI inference revolution.

1. The Core Architecture: The “Power of One” Philosophy

Historically, network scaling forced organizations into costly proprietary ecosystems. Arrcus dismantles this vendor lock-in through the Arrcus Connected Edge (ACE) platform, which delivers a unified networking environment across switches, routers, and servers.

• ArcOS (The Cloud-Native Operating System): A first-principles, 64-bit, highly programmable, microservices-based network operating system. It features modular, resilient software running on open merchant silicon (e.g., Broadcom StrataDNX) and natively supports advanced routing paradigms like IPv4/IPv6, MPLS, Segment Routing (SR-MPLS/SRv6), and BGP-EVPN.
• ArcRR (Virtualized Route Reflector): A high-performance, scale-out route reflector designed to eliminate scale barriers in multi-tenant BGP networks.
• ArcIQ (Deep Visibility & Analytics): An AI-driven, real-time analytics platform providing network operations centers (NOCs) with proactive network health monitoring, telemetry ingestion, and threat mitigation.

2. Redefining AI Infrastructure with Distributed Network Fabrics

In the era of massive AI workload scaling, physical network latency and bandwidth bottlenecks can stall training and inference cycles. Arrcus has positioned itself at the vanguard of AI-driven infrastructure:

• Arrcus Inference Network Fabric (AINF): Designed to support physical and virtual AI inference setups, AINF optimizes throughput and traffic routing based on real-time computational loads.
• Deep Hardware Integrations: Arrcus works in tight coordination with global leaders. Recent integrations include aligning the Arrcus networking stack with NVIDIA’s AI infrastructure (incorporating BlueField-3 DPUs and Spectrum-X platforms) and collaborating with Fujitsu to build secure sovereign AI workloads running on the energy-efficient FUJITSU-MONAKA Arm-based processor ecosystem.

3. Massively Scalable Virtualized Routing (VDR)

The Virtualized Distributed Router (VDR) platform from Arrcus changes the economics of network engineering.

• Eliminating Chassis Costs: Instead of relying on expensive, legacy chassis-based router hardware, VDR transforms physical routing into a virtualized, scale-out control plane cluster managing distributed, low-cost merchant silicon data planes.
• Hyperscale Elasticity: A single virtualized cluster can seamlessly scale from 4 Terabits per second (Tbps) up to 768 Tbps of forwarding capacity, offering unmatched resiliency and carrier-grade 99.999% uptime.

4. Engineering Culture: Collaborative, High-Impact, and Open

Arrcus maintains a highly technical, fast-paced environment where software engineers, QA developers, and system architects directly shape active standards.

• Be One Team: A culture focused on collaborative growth and continuous evolution. The Arrcus team is highly influential in the global networking community, having shipped over 875 products, secured more than 215 patents, and contributed over 290 papers to the Internet Engineering Task Force (IETF).
• High Accountability & Visibility: At a highly focused scale-up, developers are not isolated inside micro-segments. You have end-to-end visibility of code pipelines and work closely with senior leaders, global telecom carriers, and hyperscale cloud providers.
• Open-Standards Driven: Engineers actively work on open-source paradigms, routing security systems (such as RPKI-based Route Origin Validation), and standards-based API integrations to keep the global internet secure and open.