Arrcus Triples Bookings in 2025, Launches AI-Policy Aware Network Fabric
Event summary
- Arrcus reported a 3x increase in bookings in 2025, driven by demand across datacenter, telco, and enterprise customers.
- The company introduced the Arrcus Inference Network Fabric (AINF), designed to optimize AI inferencing applications.
- AINF aims to reduce Time to First Token (TTFT) by 60%, improve Tokens per Second (TPS) by 15%, and cut End-to-End Latency (E2EL) by 40%.
- AINF integrates with popular inference frameworks like vLLM, SGLang, and Triton, and supports Kubernetes-based orchestration.
The big picture
Arrcus' 3x bookings growth underscores the accelerating demand for flexible, cost-effective networking solutions in AI-driven environments. The launch of AINF positions the company at the forefront of a critical shift toward AI-aware networking, addressing bottlenecks in inferencing speed, model diversity, and edge decision-making. As AI inferencing becomes the fastest-growing segment, Arrcus' ability to dynamically route traffic with policy control could redefine network infrastructure for distributed AI workloads.
What we're watching
- Market Adoption
- The pace at which enterprises and network operators adopt AINF to address AI inferencing challenges.
- Competitive Response
- How incumbent networking vendors react to Arrcus' policy-aware fabric, which addresses gaps in traditional hardware-defined solutions.
- Strategic Partnerships
- The extent to which Arrcus leverages partnerships with Broadcom, Fujitsu, and others to scale AINF deployment.
Related topics