Skip to main content
A

AurCore

AurCore provides high‑performance, open‑architecture switching platforms that eliminate fabric bottlenecks for AI, HPC, and latency‑sensitive workloads. Its lossless RoCEv2 fabric, ultra‑low‑latency (≈500 ns) switching, and modular AURC‑NOS deliver deterministic performance, hot‑patching upgrades, and carrier‑grade availability for GPU clusters, hyperscale data centers, enterprise campuses, and rugged industrial environments.

Lake Forest, United StatesFounded 202414500+ followers
Updated 2 months ago

Funding

Funding not disclosed

Funding rounds are not available yet.

Founders

Founder details are not available yet.

Product

Problem

Modern AI, HPC, and enterprise workloads generate massive east‑west traffic that overwhelms traditional data‑center and campus networks, causing packet loss, high latency, and under‑utilized GPU or compute resources.

Solution

AurCore delivers high‑performance, open‑architecture switching platforms designed to eliminate fabric bottlenecks for AI and other latency‑sensitive workloads. Its lossless RoCEv2 fabric and ultra‑low‑latency pathing keep GPUs and servers fully utilized, while modular AURC‑NOS provides deterministic performance and non‑disruptive upgrades. The portfolio spans AI‑optimized 100 G/400 G (with 800 G forthcoming) switches, carrier‑grade data‑center fabrics, secure enterprise campus solutions, and rugged industrial switches, all supporting both proprietary and open‑source SONiC operating systems. Integrated load‑balancing, high‑radix spine‑leaf designs, and flexible L2/L3 overlays enable scalable, secure, and cost‑effective deployments across hyperscale, enterprise, and industrial environments.

Target Audience

Primary customers are AI/ML data‑center operators, hyperscale cloud providers, and enterprises building large GPU clusters, as well as organizations needing high‑performance campus or industrial networking for IoT, Wi‑Fi 7, and automation.

Features

  • Zero‑loss RoCEv2 transport to prevent packet drops in GPU clusters
  • Ultra‑low latency switching (≈500 ns) for synchronized distributed training
  • High‑radix spine‑leaf architecture supporting 100 G, 400 G, and upcoming 800 G speeds
  • Falcon dynamic load balancing for real‑time “elephant flow” management
  • Modular AURC‑NOS with hot‑patching, multi‑threaded processing, and fault isolation
  • Open‑source SONiC compatibility for vendor‑neutral integration
  • Carrier‑grade high availability with sub‑50 ms failover and MLAG/ECMP routing
  • Rugged, fan‑less industrial switches with 6 kV surge protection and PoE++ support
This profile is AI-generated and may contain inaccuracies.