Open architecture brings programmable compute across scale-up and scale-out networking for AI and HPC infrastructure

Summary:

• Cornelis introduces Active Compute Fabric, an open architecture for scale-up and scale-out networking with programmable compute built into the fabric.
• Qualcomm Technologies will join Cornelis at AI Infra Summit to share its perspective on the role of networking in efficient AI infrastructure.
• Cornelis has raised $205 million in funding to advance its next generation of scale-up and scale-out networking products.

Cornelis, a provider of high-performance networking solutions, today introduced Active Compute Fabric, an open architecture spanning scale-up and scale-out networking with programmable compute built into the fabric. The architecture marks the company’s entry into scale-up networking and extends technology already powering AI and HPC workloads in hundreds of data centers.

Cornelis also announced $205 million in funding to support its expansion and bring its next generation of products to market.

“AI infrastructure is reaching a point where faster endpoints alone are not enough. The fabric has to become an active part of the compute system,” said Lisa Spelman, CEO of Cornelis. “Customers want complete rack-scale solutions without being locked into a single vendor or architecture. We see a significant opportunity for Cornelis to deliver that choice with an open approach built for the demands of AI infrastructure.”

Making the fabric part of the compute system

As AI models and clusters grow, communication overhead and synchronization can leave expensive accelerators underutilized. While compute, storage, and memory have become more workload-aware, networking has largely remained focused on moving data between endpoints.

Active Compute Fabric moves beyond simply transporting data by combining lossless transport, in-fabric acceleration, and programmable compute. The architecture allows the network to operate on data as it moves through the system, adapt to changing workloads, offload collective operations, and take on new functions as AI algorithms and software evolve. The result is higher utilization of the accelerators customers already own, since work that would otherwise stall the GPU runs inside the fabric. The potential for savings is significant. In a 100,000-GPU system, Cornelis modeling of public data shows that roughly half of all GPU hours are spent waiting for data, worth about $1.68 billion a year in wasted capacity and 500 GWh of power, enough to supply nearly 48,000 U.S. homes.¹ Active Compute Fabric uses open industry standards, including UALink and ESUN for scale-up and Ultra Ethernet specifications for scale-out, and supports a broad range of accelerators.

Qualcomm Technologies joins Cornelis at AI Infra Summit

Cornelis and Qualcomm Technologies share a common view of the challenge facing AI infrastructure: passive fabrics that only move data leave accelerators underutilized, and that inefficiency shows up directly in the economics of inference. As rack-scale AI systems grow, both companies see the network becoming a first-order design decision rather than an afterthought.

“As AI systems continue to scale, moving data efficiently across the rack becomes just as important as the compute itself,” Tony Pialis, EVP and GM, Data Center, Qualcomm Technologies, Inc. “Improving utilization and AI economics will require a more integrated approach across compute, memory, and networking, and Cornelis’ vision for an open, programmable fabric aligns with that industry direction. Giving customers more choice and flexibility across the infrastructure stack will be critical as AI moves toward rack-scale architecture.”

Pialis will join Spelman during her keynote at AI Infra Summit to share Qualcomm Technologies' perspective on the role of the network in efficient AI infrastructure.

$205 million in funding backs next generation

Cornelis also announced $205 million in funding to expand into scale-up networking and bring its next generation of products to market. The funding is supporting the company’s efforts to scale production, deepen customer partnerships, and accelerate its go-to-market plans.

"Open-standard scale-up and scale-out networking for AI represents more than $55 billion of opportunity by 2030, and we believe the network will decide how much of the total AI build-out delivers real return," said Joel Whitley, Partner at IAG Capital Partners. "What gives us confidence in Cornelis is the combination of a differentiated vision and a team that has demonstrated it can execute. The company has brought multiple generations of technology to market and is well positioned to capitalize on the next phase of growth in AI infrastructure."

CN5000 is shipping today, while CN6000 is sampling with customers ahead of expanded availability expected in Q4 2026. Cornelis will also offer a first look at its next generation scale-up and scale-out roadmap at AI Infra Summit.

See it at AI Infra Summit

Spelman will share more on the Active Compute Fabric architecture during a keynote at AI Infra Summit on Sept. 15. Cornelis will demonstrate CN6000 and preview its next generation scale-up and scale-out portfolio at booth #830. More at cornelis.com/GoActive.

¹ Performance figures for the next generation roadmap are Cornelis projections based on pre-production simulation and modeling. Actual results will vary. Utilization and cost baselines reflect published industry data on large-scale training and inference workloads. Cornelis estimates based on a modeled 100,000-GPU system at $4.00 per GPU-hour, 8,400 annual operating hours, and approximately 50 percent unproductive GPU time, consistent with published industry data on large-scale training and inference workloads. Household energy comparison based on U.S. Energy Information Administration average residential consumption of approximately 10,500 kWh per year.

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. Learn more at www.cornelis.com.

FAQ

• What is Cornelis announcing?
  Cornelis is unveiling its Active Compute Fabric, an open architecture spanning scale-up and scale-out networking with programmable compute built into the fabric, marking its entry into scale-up networking. The company is also announcing a collaboration with Qualcomm Technologies, Inc. on validation toward future rack-scale AI data center designs and approximately $205 million in funding.
• What is Active Compute Fabric?
  Active Compute Fabric combines lossless transport, in-fabric acceleration, and programmable compute across scale-up and scale-out networks. It adapts to workloads, offloads collective operations, and performs functions on data as it moves through AI and HPC systems.
• Why is Cornelis entering the scale-up networking market?
  Expanding into scale-up brings Cornelis’ architecture closer to the accelerators inside rack-scale AI systems. The approach builds on the company’s existing scale-out technology and third-generation switch silicon while preserving accelerator choice through open industry standards.
• What is Cornelis doing with Qualcomm Technologies?
  Cornelis and Qualcomm Technologies, Inc. share a common view of the networking challenges facing increasingly large AI systems. Tony Pialis, EVP and GM, Data Center at Qualcomm Technologies, Inc., will join Cornelis CEO Lisa Spelman during her AI Infra Summit keynote to share Qualcomm Technologies’ perspective on the role of networking in efficient AI infrastructure.
• How much funding has Cornelis raised?
  Cornelis has raised approximately $205 million in funding to build out its scale-up networking capabilities, develop its next generation scale-up and scale-out roadmap, and support CN6000 manufacturing and customer deployments.

View source version on businesswire.com: https://www.businesswire.com/news/home/20260914323310/en/

“Customers want complete rack-scale solutions without being locked into a single vendor or architecture. We see a significant opportunity to deliver that choice with an open approach built for AI infrastructure demands.” Lisa Spelman, CEO, Cornelis