Cloud hosting firm Vultr has expanded its Seattle, Washington State, cloud data center region with Nvidia H100 GPU clusters.

The GPU clusters are available to customers both on demand and through reserved instance contracts.

Sabey Data Centers Columbia
– Sabey Data Centers

Housed at Sabey Data Centers' SDC Columbia campus, the region is hydro-powered through the Douglas County Public Utility District, enabling green-energy-based AI.

The campus plans to eventually have nine data center buildings across 130 acres, and Sabey recently completed construction of the third building on the site.

The data centers have an annualized PUE of 1.15.

“The transformative impact of AI has created a rapidly growing need from enterprises and AI innovators for high-performance, scalable, sustainable cloud GPU infrastructure. We chose Sabey for our latest expansion of Nvidia GPU capacity because of our shared goal of reducing the carbon footprint of AI training and inference,” said J.J. Kardwell, CEO of Vultr’s parent company, Constant.

“We are thrilled to welcome Vultr to our customer roster as the first cloud computing platform to offer cloud GPUs in our SDC Columbia data center,” said Robert Rockwood, president of Sabey Data Centers. “Given the crucial role data centers and AI play in our lives and in shaping humanity’s future, it’s important that we power them as sustainably as possible."

Vultr's Nvidia GPU offerings include the GH200 Grace Hopper Superchip, which was officially added in November 2023; the H100; A100 Tensor Core GPU; L40S; A40; and A16.

Vultr has 32 locations worldwide, covering Europe, APAC, the US, LATAM, the Middle East, and South Africa.