10 switches running Larch Hardened SONiC and Dorado Cruz orchestrating and monitoring solution allow creation of accessible AI systems.
Larch Networks, a provider of enterprise-grade SONiC-based networking systems, and Dorado, a provider of orchestration and automation solutions, announced a joint demonstration at the OCP Global Summit. The demonstration will showcase a fully managed, low-power low-latency AI fabric built with ready-to-use, commercially available Ethernet components, offering a clear alternative to heavily constrained proprietary AI networks.
Today, advanced AI infrastructure is largely isolated within hyperscaler environments. For the broader market, traditional proprietary AI networking technologies like InfiniBand are hindered by extreme supply shortages, multi-year lead times, and complex end-use authorizations. This leaves mainstream enterprises and service providers struggling to deploy the AI clusters required to remain competitive.
The joint solution from Larch Networks and Dorado directly addresses this availability crisis. By leveraging existing Ethernet fabrics powered by the Marvell® TeraLynx® 10 switch architecture, organizations can bypass proprietary bottlenecks. This approach not only ensures immediate availability but also delivers the lower power consumption and ultra-low latency required for demanding AI workloads and reducing total cost of ownership. When combined with a flexible choice of alternate AI accelerators, this open architecture democratizes AI infrastructure, bringing advanced capabilities closer to real users and real markets. The demonstration fabric features TeraLynx 10 switches in a spine-leaf topology running Larch Hardened SONiC, seamlessly connecting GPU servers equipped with ConnectX-7 network adapters.
While open networking gives AI infrastructure operators’ flexibility and freedom from vendor lock-in, these high-performance environments must also be easy to deploy, operate, and troubleshoot at scale. Dorado Cruz provides centralized network orchestration, observability, automation, and tuning. The platform delivers automated device discovery, configuration and provisioning, real-time health monitoring, telemetry, and workload-aware tuning across the fabric from a single pane-of-glass. This lets operators run a high-performance, open AI network with the standard operational tooling they expect from traditional enterprise systems.
OCP Global Summit attendees can view the live, fully operational AI fabric demonstration during the exhibit hall days October 13-15.
Want to arrange a meeting in advance? https://meetings.hubspot.com/imurzina/larch-networks-at-ocp-
About Dorado
Dorado helps organizations accelerate AI infrastructure deployment, simplify operations, close specialized skills gaps, and move toward intelligent network operations with built-in AI governance and security. With AI network fabric and compute orchestration, AI networking solutions, and intelligent infrastructure operations, Dorado gives modern data centers an end-to-end path to production AI.
Dorado’s Cruz AI portfolio orchestrates network fabrics, compute clusters, and rack-scale architectures across deployment, configuration, operations, telemetry, and lifecycle management. With Cruz’s native agentic AI capabilities, teams can interact with live infrastructure in natural language and, with human approval, take and verify actions.
Dorado combines proven infrastructure technologies with its Cruz platform, professional services, and ongoing technical support to help organizations meet their AI business goals faster.