First NVIDIA Vera Rubin NVL72 benchmarks show 10X improvement over Grace Blackwell

Q: How much improvement did CoreWeave report for Vera Rubin NVL72 versus Grace Blackwell NVL72 on DeepSeek-R1?

CoreWeave reported a 10X improvement for Vera Rubin NVL72 versus Grace Blackwell NVL72 on the DeepSeek-R1 benchmark. This 10X figure refers to tokens per second per megawatt, an efficiency and power-budget-based metric.

Q: What metric is NVIDIA using when it cites the 10X improvement (tokens per second per what)?

NVIDIA is using tokens per second per megawatt as the metric for the 10X improvement. This measures throughput relative to power consumption, comparing how effectively Vera Rubin scales versus Grace Blackwell under the same power budget and workload.

Q: How does NVIDIA’s Spectrum-X contribute to Vera Rubin’s performance gains?

Spectrum-X alleviates bandwidth bottlenecks by providing extremely high switching capacity for Vera Rubin racks, enabling a reported 1.64 Pb/s per rack. CoreWeave deploys the Spectrum-X Ethernet SN6600-LD, built on the 102.4 Tb/s Spectrum-6 switch chip with a liquid-cooled, dense design that delivers 100% more capacity than previous-generation air-cooled switches, and this bandwidth increase is cited as a key part of the 10X tokens-per-second-per-megawatt improvement.

Q: What switching hardware and throughput did CoreWeave deploy for Vera Rubin NVL72?

CoreWeave deployed the NVIDIA Spectrum-X Ethernet SN6600-LD as the switching fabric for Vera Rubin NVL72, built on the 102.4 Tb/s Spectrum-6 switch chip with a liquid-cooled design. This setup delivers 1.64 Pb/s of switching throughput per rack, offering 100% more capacity than previous-generation air-cooled switches.

CoreWeave ran the same DeepSeek-R1 benchmark on both Vera Rubin NVL72 and Grace Blackwell NVL72 systems and found a 10X improvement in one key area.

VIEW GALLERY - 2

Kosta Andreadis

Senior Editor

Published Jul 22, 2026 12:26 AM CDT

1-minute read time

TL;DR: CoreWeave's DeepSeek-R1 runs show Vera Rubin NVL72 delivers tenfold tokens-per-second-per-megawatt versus Grace Blackwell NVL72. Vera Rubin racks-deployed across partners-use Spectrum-X switching (1.64 Pb/s per rack) and liquid cooling, improving bandwidth, power efficiency, and reducing water use.

Voice: Kosta AndreadisSpeed

0:00 / 2:56

"Vera Rubin NVL72 production is ramping up with racks running at partners CoreWeave, Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure," NVIDIA confirms as its rack-scale Vera Rubin platform rolls out to multiple sites in 30 countries. And when it comes to performance, we've got CoreWeave's first benchmark on DeepSeek-R1.

First NVIDIA Vera Rubin NVL72 benchmarks show 10X improvement over Grace Blackwell 1

VIEW GALLERY - 2 IMAGES

And yes, a 10X improvement over Grace Blackwell NVL72 definitely falls into the "order-of-magnitude performance leap" that NVIDIA attributes to it, with this number referring to tokens per second per megawatt. As an efficiency and power-budget-based metric, it directly correlates to how effectively AI infrastructure built with Vera Rubin scales when dealing with the same power budget and workload as Grace Blackwell systems.

This 10X figure comes from CoreWeave running the same DeepSeek-R1 benchmark on both Vera Rubin NVL72 and Grace Blackwell NVL72. A key part of the 10X improvement and impressive evolution is how Vera Rubin leverages NVIDIA Spectrum-X to deal with bandwidth bottlenecks. We're talking about a jaw-dropping 1.64 Pb/s per rack.

Read more: AMD says its next-gen EPYC 'Venice' processor is over 3X faster than NVIDIA Vera
Read more: NVIDIA CEO says new Blackwell Ultra GB300 AI platform is in full-scale mass production
Read more: NVIDIA's financial results show record quarterly revenue of $68 billion, driven by AI

"CoreWeave is among the first to deploy the NVIDIA Spectrum-X Ethernet SN6600-LD as the switching fabric for Vera Rubin NVL72," NVIDIA confirms. "Built on the 102.4 Tb/s Spectrum-6 switch chip and featuring a liquid-cooled design, CoreWeave deploys dense switching racks, delivering 1.64 Pb/s per rack with 100% more capacity than previous-generation air-cooled switches."

Frequently Asked Questions

TweakBot answers common questions about this news using TweakTown's own coverage from this page and related content from our archive. Tap a question to reveal the answer, or type your own below.

Question #1

How much improvement did CoreWeave report for Vera Rubin NVL72 versus Grace Blackwell NVL72 on DeepSeek-R1?

Question #2

What metric is NVIDIA using when it cites the 10X improvement (tokens per second per what)?

Question #3

How does NVIDIA’s Spectrum-X contribute to Vera Rubin’s performance gains?

Question #4

What switching hardware and throughput did CoreWeave deploy for Vera Rubin NVL72?

Have a question not listed here? Ask below and TweakBot will answer it.

With AI and data center power use becoming a point of contention in surrounding communities, improved efficiency is quickly becoming one of the main focuses when talking about large rack-scale systems built on platforms like Vera Rubin. In addition to its power-saving capabilities, the Vera Rubin NVL72 rack is also designed to save "millions of gallons of water per megawatt," thanks to its cable-, fan-, and hose-free trays and liquid-cooling inlet that enables dry-cooler operation.