AI Networking Firm Cornelis Raises $205M to Challenge Nvidia's Dominance
Cornelis, a company building networking technology for AI chips, has raised $205 million and unveiled a new product aimed at reducing GPU idle time. The firm is pushing an open architecture to compete against Nvidia's market dominance.
Cornelis, a company that develops networking technology to help AI chips communicate more efficiently, announced Monday that it has raised $205 million in a funding round led by IAG Capital Partners.
Tackling Wasted GPU Time
Alongside the funding news, the company unveiled a new product called Active Compute Fabric. This networking technology addresses a common inefficiency in AI computing: GPUs frequently sit idle while waiting for data to arrive. Cornelis's solution allows chips to process and transmit information simultaneously, aiming to cut down on that wasted time.
Competing with Nvidia's Ecosystem
Cornelis, which spun off from Intel in 2020, is positioning itself as a competitor to Nvidia by offering an open architecture. This means customers can pair a variety of GPU and accelerator hardware from different manufacturers with Cornelis's networking fabric, rather than being locked into one vendor's ecosystem. While Nvidia's chips can technically operate on other networking systems, they are optimized for Nvidia's own software stack, which makes adopting the company's full suite of products more convenient and appealing for customers.
Cornelis is part of a growing group of AI infrastructure companies working to erode Nvidia's dominant position in the market piece by piece. The company has already begun shipping its current product and is now developing a next generation, which is expected to launch later this year.
