GPU cloud provider Lambda has deployed an Nvidia GB200 NVL72 rack system at EdgeCloudLink's (ECL) data center in Mountain View, California.
In a LinkedIn post, ECL's founder and CEO Yuval Bachar wrote: "Lambda, Supermicro & ECL just powered the first Supermicro NVL72 GB200 based system at the ECL-MV1 site running at 110kw+ based 100 percent on Hydrogen power with Zero emission. The site is running now 3.6 million GPU cores!!! in two rack spaces!!!! Unbelievable."
Lambda's VP of supply chain and manufacturing Vijay Manyam added: "We did it again! Lambda’s 2nd GB200 NVL72 rack, courtesy of Supermicro, is up and running at ECL! Special thanks to our partners for bringing this amazing Nvidia product into our AI compute fleet!"
Bachar told DCD that the rack was installed on January 22, and the installation process took around "one hour" due to the flexibility of the data center.
"Our data center is flexible to a level that for us to go from 75kW per rack to 150kW is just a simple modification," said Bachar.
The flexibility is related to ECL's modular design, which Bachar says enables ECL to deliver a data center in nine months. He added: "Today's cycle of innovation with Nvidia is that every nine to 12 months we see a new technology that comes with different requirements and densities," noting that a faster time to market enables ECL to plan for those changes and adjust its design accordingly.
The ECL-MV1 data center offers 1MW of IT capacity and was launched in May 2024. The data center is powered entirely by hydrogen - currently a mix of grey, blue, and green hydrogen.
Lambda is also set to be the anchor tenant at ECL's upcoming 1GW TerraSite-TX1 data center campus in Houston, Texas, which will also be powered by hydrogen.
According to Bachar, the recently installed GB200 NVL72 rack was originally planned for the new Texas site.
Lambda has also this week announced the delivery of a GB200 NVL72 rack to Pegatron SVR which will similarly be used to power Lambda's AI compute platform.
The GB200 NVL72 liquid-cooled server rack was first revealed in March 2024 and features 72 GB200 GPUs, 36 Grace CPUs, and nine NVLink switch trays, each of which has two NVLink switches. The company has previously said that this configuration will enable the system to run as a giant GPU, boosting performance.
In November 2024, reports emerged of a redesign of the rack, which was released after claims of overheating in the existing configuration.
Founded in 2012, Lambda rents out colocation space in data centers in San Francisco, California, and Allen, Texas, and offers comparatively low-cost GPU-based cloud compute. Alongside the cloud, it provides colocation space and sells GPU desktops.
In July 2024, it was reported that the company was seeking $800m in funding to purchase additional Nvidia GPUs, associated network infrastructure, and software.
