Cornelis, a company specializing in networking technology designed to enhance communication among AI chips, announced a $205 million funding round led by IAG Capital Partners. The company introduced its Active Compute Fabric, a networking solution intended to address inefficiencies where GPUs spend significant time waiting for data transfer. This fabric allows AI chips to process and transmit data concurrently, improving overall performance.
Founded as a spin-off from Intel in 2020, Cornelis is positioning itself against Nvidia by providing an open architecture. This approach enables customers to integrate various GPU and accelerator hardware with Cornelis’ networking fabric, unlike Nvidia’s ecosystem, which is optimized for its own software stack. While Nvidia GPUs can operate on other networking fabrics, their full capabilities are best realized within Nvidia’s proprietary environment, making it challenging for competitors to gain traction.
Cornelis represents a growing group of AI infrastructure companies aiming to reduce Nvidia’s market dominance by offering alternative solutions. The company has already begun shipping its current product and plans to release an updated version later this year.