Cornelis, which spun off from Intel in 2020 to build networking tech that helps AI chips communicate more effectively, raised $205M led by IAG Capital Partners
Context & Ripple Effects
Cornelis enters a better-funded AI-networking field: Upscale raised $200 million in January and then a $190 million follow-on round in June, while Enfabrica had raised $125 million in 2023 for networking chips designed to scale Nvidia GPU data centers. Cornelis’s Intel spinout lineage gives the financing added significance as a bid to turn specialized chip-to-chip communication technology into an independent infrastructure business.
First-order effects
- Cornelis gains $205 million, led by IAG Capital, to fund development and commercialization of networking technology aimed at improving communication among AI chips.
- IAG Capital becomes a major financial backer of Cornelis, tying its investment to demand for AI data-center interconnects.
Second-order effects
- Cornelis joins Upscale and Enfabrica in competing for AI-infrastructure capital and customer attention, making networking a more visible spending category alongside AI chips.
- AI data-center operators evaluating large chip clusters gain another funded specialist focused on the communication layer rather than the processors themselves.
Third-order effects
- If funding continues to flow to specialists such as Cornelis, Upscale, and Enfabrica, AI infrastructure investment will extend beyond accelerators toward the networking bottlenecks that determine how effectively clusters operate.
- The field may increasingly split into specialist interconnect suppliers and the larger compute-platform vendors they must interoperate with, raising the strategic value of open or broadly compatible networking technology.
The trend: AI infrastructure financing is broadening from compute chips to the networking layers needed to turn large numbers of AI processors into usable clusters.