# Liquid-Cooled Switching Doubles AI Network Bandwidth Per Rack

DevFeed: [Liquid-Cooled Switching Doubles AI Network Bandwidth Per Rack](<https://devfeed.tech/articles/liquid-cooled-switching-doubles-ai-network-bandwidth-per-rack-77995.md>)

Original publisher: [Read original article](<https://wf.coreweave.com/blog/liquid-cooled-switching-doubles-ai-network-bandwidth-per-rack>)

Author: CoreWeave

Published: 2026-10-02T14:12:41Z

Content type: article

Language: en

Sources: [Coreweave Blog Posts](<https://devfeed.tech/sources/coreweave-blog-posts.md>)

Topics: [NVIDIA Vera Rubin](<https://devfeed.tech/topics/nvidia-vera-rubin.md>), [NVIDIA Spectrum-X Ethernet](<https://devfeed.tech/topics/nvidia-spectrum-x-ethernet.md>), [flat network topology](<https://devfeed.tech/topics/flat-network-topology.md>), [Software-Defined Data Center](<https://devfeed.tech/topics/software-defined-data-center.md>), [AI Networking](<https://devfeed.tech/topics/ai-networking.md>)

Tags: [bandwidth](<https://devfeed.tech/tags/bandwidth.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [carbon-footprint](<https://devfeed.tech/tags/carbon-footprint.md>), [cluster](<https://devfeed.tech/tags/cluster.md>), [clusters](<https://devfeed.tech/tags/clusters.md>), [compute](<https://devfeed.tech/tags/compute.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [content-type](<https://devfeed.tech/tags/content-type.md>), [cooling](<https://devfeed.tech/tags/cooling.md>), [cooling-systems](<https://devfeed.tech/tags/cooling-systems.md>), [cost](<https://devfeed.tech/tags/cost.md>), [data-center](<https://devfeed.tech/tags/data-center.md>), [data-center-design](<https://devfeed.tech/tags/data-center-design.md>), [has](<https://devfeed.tech/tags/has.md>), [liquid](<https://devfeed.tech/tags/liquid.md>), [network](<https://devfeed.tech/tags/network.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [switching](<https://devfeed.tech/tags/switching.md>)

## AI overview

CoreWeave describes deploying liquid-cooled NVIDIA Spectrum-X SN6600-LD switches with Vera Rubin NVL72 systems. The article says the switches provide 102.4 Tb/s of capacity each and enable 1.64 Pb/s of switching capacity per rack, while integrating network monitoring, cooling controls, and lifecycle management. Compared with the air-cooled SN5600/SN5610, CoreWeave reports twice the switching performance per rack, a 62.5% reduction in rack footprint, and 30% better power efficiency for 16,000 Vera Rubin GPUs.

## Source excerpt

CoreWeave deploys liquid cooled switching for NVIDIA Vera Rubin NVL72 delivering 100% higher switching performance at 1.64 Pb/s per rack compared to air cooled switches.