AI infrastructure startup Cornelis announced Monday that it has raised $205 million in a funding round led by IAG Capital Partners. Spun off from Intel in 2020, the company designs hardware and software networking fabrics engineered to improve data communication efficiency between AI accelerators and reduce idle GPU time.
Alongside the funding, Cornelis unveiled its Active Compute Fabric technology, which allows hardware components to process and transfer data simultaneously. By leveraging an open architecture, Cornelis enables enterprise customers to pair its interconnect fabric with heterogeneous GPU and accelerator chips, posing a direct challenge to Nvidia’s proprietary networking ecosystem.
Why it matters
Open networking fabrics offer data center operators an alternative to Nvidia’s proprietary hardware stack, potentially lowering AI cluster buildout costs.
Enterprise buyers gain flexibility to integrate heterogeneous GPU and accelerator hardware without suffering data transfer bottlenecks.
Source: techcrunch.com
