Cornelis and NextSilicon have announced a collaboration to develop next-generation reference architectures for AI and high-performance computing, targeting the critical bottlenecks that leave expensive systems underutilized. By pairing Cornelis' congestion-free networking with NextSilicon's reconfigurable compute platform, the joint effort aims to deliver validated system blueprints for OEM partners and customers. The initiative, unveiled at ISC High Performance 2026, addresses both data movement and compute efficiency in a unified design.
Cornelis and NextSilicon are building joint reference architectures for AI and HPC.
The collaboration combines the CN5000 fabric with the Maverick-2 dataflow accelerator.
Joint evaluations aim to eliminate networking congestion and compute idle time.
OEM partners will receive proven system blueprints instead of untested component lists.
Future work will extend to 800 Gbps CN6000 fabric and disaggregated AI inference.
The partnership targets emerging workloads like Mixture of Experts and agentic AI.
The collaboration officially launched at ISC High Performance 2026 in Hamburg, where the companies detailed their plan to validate integrated solutions for AI and HPC environments. The initial phase pairs the 400 Gbps CN5000 fabric, which debuted in 2025, with the Maverick-2 compute platform that began volume shipping later that same year. This first stage focuses on verifying how the fabric and compute perform together across various configurations, ensuring that OEM partners can start from proven combinations rather than untested parts lists. The companies have already outlined plans to extend testing to the 800 Gbps CN6000 fabric, expected in the second half of 2026.
Each company targets a different bottleneck that has historically plagued high-performance systems. Standard Ethernet was not built for the small, latency-sensitive messages that AI inference and HPC simulation generate at scale, leading to congestion that leaves expensive compute capacity sitting idle while waiting for data. The CN5000 is specifically engineered to eliminate that idle time by providing a lossless, congestion-free fabric. On the compute side, the traditional von Neumann model shuttles data between memory and a fixed execution unit, stalling on irregular, data-dependent workloads that now dominate AI and HPC. NextSilicon built Maverick-2 on its Intelligent Compute Architecture (ICA), a software-defined dataflow design that reconfigures to each workload at runtime and runs existing code without modification. Pairing the two addresses both limits at once: a fabric that keeps data moving and an accelerator that keeps compute busy, giving OEM partners a comprehensive blueprint for systems they can build and bring to market.
“Operators keep telling us their most expensive systems sit idle, waiting on the network," said Lisa Spelman, CEO of Cornelis. "We built the CN5000 to end that wait. NextSilicon challenges the same kind of assumption on the compute side, so this collaboration is a natural fit. Together we can show partners and customers what a congestion-free fabric and a workload-driven compute architecture deliver as one design.”
"For decades, software had to bend to fit the processor," said Elad Raz, founder and CEO of NextSilicon. "Maverick-2 makes the processor adapt to the software. Cornelis takes the same approach to the network. Evaluating our architectures together is the first step toward giving customers and OEM partners a faster, more efficient foundation for AI and HPC."
About Cornelis
Cornelis delivers high-performance scale-out and scale-up networking solutions that accelerate AI and HPC workloads. Cornelis technology enables lossless, congestion-free networking that reduces training time, improves inference, and maximizes compute utilization. From foundation model training to complex climate modeling and real-time analytics, Cornelis solutions power the most demanding workloads across commercial, academic, government, and cloud environments. With a focus on performance, scalability, and efficiency, Cornelis helps organizations achieve faster insights and greater return on infrastructure investments.
About NextSilicon
NextSilicon builds computing infrastructure for algorithmically complex workloads. The company's Maverick-2 accelerator uses a runtime reconfigurable dataflow architecture to deliver up to 10x performance over leading GPUs at less than half the power, with no requirement to rewrite existing applications. Maverick-2 is in production at customer sites across HPC, AI, and national security computing environments. NextSilicon is headquartered in Tel Aviv, Israel, with offices in Minneapolis, MN, in the United States.
