FundraiseSep 14, 2026, 16:59 UTC

Cornelis Networks raises $205M to speed up AI networks

What's the deal? Cornelis Networks has raised $205 million in a late-stage round led by IAG Capital Partners. The AI networking startup announced the funding alongside a new Active Compute Fabric architecture and a strategy collaboration with Qualcomm Technologies.

What does the company do? Cornelis builds specialised, congestion-free data center networking systems for AI and high-performance computing. It competes with more established providers such as Cisco Systems and Arista Networks.

Why now? The rapid growth in AI cluster sizes has created bottlenecks in data center networks, leaving expensive AI accelerators idle while they wait for data. Compute, memory, and storage have become more "workload-aware," but traditional networks have not, said chief executive Lisa Spelman.

What's the product? Announced at the AI Infra Summit, the Active Compute Fabric integrates programmable compute directly into the network, so it can process data in transit rather than just moving it. It combines in-fabric acceleration with lossless transport and is built on open standards including Ethernet, UALink, and Ultra Ethernet.

What's the endgame? The fabric performs workload-specific operations as data moves — assembling KV cache data, coordinating expert dispatch for mixture-of-experts models, and compressing gradients in transit. "The payload does not arrive the way it left," said chief marketing officer Brandon Draeger.

Cornelis says this cuts network traffic by up to 50%, based on pre-production simulations. Accelerator use in large AI deployments often sits near half of installed capacity, and the architecture aims to recover a share of that waste, Draeger said.

The signal: The round ranks in the 96th percentile among all-time late-stage VC deals in US big data, a sign of how much capital is chasing the plumbing beneath the AI boom — not just the models and chips, but the networks feeding them.