Cornelis Networks raises $205M and scales up and out with its new Active Compute Fabric

Cornelis Networks Inc., a data center infrastructure startup focused on specialized networking systems for AI and high-performance computing, has announced a significant step forward with the launch of its new Active Compute Fabric architecture. The launch coincides with a $205 million funding round led by IAG Capital Partners and a strategic collaboration with Qualcomm Technologies Inc. Cornelis Networks specializes in congestion-free data center networking systems for AI workloads, standing out from established providers like Cisco Systems Inc. and Arista Networks Inc. The Active Compute Fabric, an open architecture recently unveiled at the AI Infra Summit, integrates programmable compute directly into the fabric of the network, making it more flexible and efficient.

Cornelis' CEO, Lisa Spelman, explained that traditional networks are struggling to keep up with the growing sizes of AI model compute clusters, leading to expensive AI accelerators sitting idle for too long. The Active Compute Fabric addresses this by combining in-fabric acceleration with lossless transport and programmable compute, enabling the network to actively work with data as it traverses the system.

As a result, the network can adapt in real-time based on shifting workloads, offload complex collective operations, and enhance performance. The fabric's architecture allows for up to 50% reduction in overall network traffic and returns a meaningful share of the wasted accelerator capacity. Cornelis is expanding its focus to scale-up and scale-out networks to bring its architecture closer to AI accelerators and provide customers with more choices in building AI infrastructure.

Qualcomm, a strategic collaborator, shares this vision and is joining Cornelis at the AI Infra Summit to announce a collaboration aimed at shaping the future of networking for rack-scale AI data centers.