
Cornelis Networks builds open interconnect systems for AI and high-performance computing.
cornelis.comCrawled Oct 15, 2026
The page covers the Cornelis CN5000 Omni-Path Director Class Switches.
The page says early AI infrastructure design prioritized compute horsepower, with performance tied to the number of GPUs connected.
Every major enterprise and cloud provider is racing to scale AI infrastructure.
The company aims to maintain the nation's competitive edge and keep pace with innovation coming out of Silicon Valley.
The page features news, product launches, and industry events from Cornelis, including innovations in AI and HPC networking solutions.
Cornelis’s leaders guide the development of AI and HPC networking solutions.
Cornelis's leaders guide innovation in high-performance AI and HPC networking solutions.
Cornelis works on open, high-performance networking for AI and HPC.
The page describes Cornelis's open, high-performance networking for AI and HPC and covers its mission, team, solutions, and global impact.
The company builds high-performance fabrics for HPC, HPDA, and AI in various industries.
Cornelis Networks launches its Active Compute Fabric architecture for AI scale-up and scale-out networks alongside a strategic collaboration with Qualcomm and a $205 million funding round led by IAG Capital Partners. The open fabric uses programmable compute and in-network acceleration to perform workload-specific operations, including assembling KV cache data and accelerating collective operations; Cornelis says pre-production simulations show network traffic reductions of up to 50%. The funding will scale production of its CN5000 and CN6000 switches and accelerate fabric deployment.
Sep 2026 · SiliconAngleNews article
Cornelis Networks and NEC Corporation expand their collaboration to advance AI and high-performance computing infrastructure in Japan. The work includes evaluation activity for Cornelis CN5000 and CN6000 systems.
Aug 2026 · Cornelis NetworksNews article
Cornelis announces a reference architecture pairing its CN6000 interconnect with AMD’s next-generation compute platforms for AI inference, training, and high-performance computing. The architecture is built for AMD 6th Gen EPYC and AMD Instinct MI400 Series processors.
Jul 2026 · Cornelis NetworksNews article
The Texas Advanced Computing Center has accepted a Cornelis CN5000 upgrade for Stampede3. The deployment involves more than 600 nodes.
Jun 2026 · Cornelis NetworksNews article