DriveNets Extends AI Networking with new AI Fabric platforms built for massive AI clusters, delivering higher performance, scalability, and efficiency.
DriveNets Extends AI networking capabilities with the launch of two new AI Fabric platforms designed to support some of the world's largest AI deployments. The company unveiled the DriveNets 2600SL and DriveNets 2601S, networking solutions aimed at powering AI clusters that can scale from hundreds to hundreds of thousands of XPUs, the processors that drive modern AI workloads.
The new platforms are powered by Broadcom's Tomahawk 6 chip and offer a total networking capacity of 102.4Tbps across 64 ports operating at 1.6Tbps. Industry analysts expect these high-speed connections to become a major part of future AI infrastructure as demand for larger AI models and faster processing continues to rise.
AI Growth Is Pushing Networks to Their Limits
As organizations invest heavily in AI training and inference systems, networking has become a crucial factor in overall performance. A slow or inefficient network can leave expensive computing resources underutilized, reducing productivity and increasing operational costs.
"The most expensive idle asset in the world today is an XPU waiting on the network," said DriveNets co-founder and CEO Ido Susan.
The new additions to their AI portfolio deliver industry-leading performance at massive cluster scale and let their customers maximize infrastructure utilization and power efficiency on any AI accelerator.
The company says the platforms are designed to support multiple network architectures and can handle both current and next-generation AI processing requirements. Flexible configurations allow operators to choose different port speeds depending on their deployment needs.
Why Are Cooling and Efficiency Becoming Critical for AI Data Centers?
One of the standout features is the availability of both air-cooled and liquid-cooled versions. The DriveNets 2600SL offers a fully liquid-cooled design, helping data centers manage the growing heat generated by large AI workloads while improving energy efficiency.
Beyond hardware, DriveNets is also focusing on software optimization. Its AI Fabric solution includes performance enhancements across network interface cards, drivers, operating systems, communication libraries, and network protocols. These optimizations are supported by the company's AI Cluster Orchestrator, which helps automate deployment, testing, and ongoing management of large AI environments.
Open Ethernet Gains Momentum in AI Infrastructure
DriveNets and Broadcom believe open Ethernet-based networking is becoming the preferred foundation for next-generation AI data centers. The approach enables organizations to build multi-vendor AI clusters while maintaining performance levels typically associated with proprietary solutions.
Analysts also note that AI networking is rapidly evolving into a major market segment as demand for larger and more efficient AI systems accelerates.
As Business Fortune observes, the arrival of these new platforms signals a future where AI infrastructure becomes faster, more scalable, and more energy efficient. As AI clusters continue expanding, advanced networking solutions are expected to play an increasingly central role in unlocking the full potential of next-generation artificial intelligence.
FAQs
What did DriveNets announce?
DriveNets introduced the 2600SL and 2601S networking platforms, expanding its AI Fabric portfolio for large-scale AI deployments.
What is the difference between the two new platforms?
The 2600SL uses a fully liquid-cooled design, while the 2601S is air-cooled, allowing deployment in different data center environments.
What technology powers these platforms?
Both platforms are built on Broadcom's Tomahawk 6 ASIC, delivering up to 102.4Tbps of networking capacity.
Who can benefit from these solutions?
Hyperscalers, AI model developers, cloud providers, and enterprises running large AI workloads can benefit from the new platforms.
Why is networking important for AI infrastructure?
Networking connects AI processors and data resources. High-performance networking improves utilization, reduces delays, and helps AI systems operate more efficiently at scale.
