Cornelis, spun off from Intel in 2020 to build networking tech that helps AI chips communicate more effectively, raised $205M led by IAG Capital
Context & Ripple Effects
Cornelis emerged from Intel in 2020 with a focus on the network layer connecting AI chips. Its financing arrives after Upscale’s $190M AI-networking round and follows funding for Enfabrica’s networking chips for GPU data centers, showing that investors are backing multiple approaches to scaling AI clusters.
The common constraint is no longer only the processor: the links among processors and memory are becoming a distinct infrastructure market, as illustrated by Celestial AI’s optical interconnect funding.
First-order effects
- Cornelis gains $205M, led by IAG Capital, to advance and commercialize its technology for communication among AI chips.
- IAG Capital becomes the lead financial backer of an Intel spin-off competing for AI-cluster networking deployments.
Second-order effects
- Upscale and Enfabrica face a better-funded peer in the contest to supply the networking layer of AI data centers, increasing pressure to differentiate their architectures and win deployments.
- AI-system buyers gain another financed specialist supplier, making network design a more explicit procurement choice alongside accelerators and memory.
Third-order effects
- If funding continues to concentrate in interconnect specialists, AI infrastructure spending will spread beyond compute chips toward the fabric that determines how efficiently clusters operate.
- The market is trending toward a more modular AI stack, in which networking vendors can compete for strategic control of cluster performance rather than serving as interchangeable components.
The trend: AI-infrastructure investment is broadening from processors into the networking and interconnect technologies needed to turn large numbers of chips into usable AI systems.