Vultr introduces NVIDIA HGX B200 for global AI workloads

Vultr has announced the early availability of the NVIDIA HGX B200 across its global infrastructure.

The NVIDIA HGX B200 is set to enhance the performance of AI workloads by offering a combination of increased speed and energy efficiency. Vultr's integration of these advanced GPUs into their cloud infrastructure aims to support enterprises in scaling AI applications effectively across 32 data centre regions worldwide.

J.J. Kardwell, CEO of Vultr, commented, "We are pleased to offer early availability of the NVIDIA HGX B200, providing our customers with access to NVIDIA's latest advancements in accelerated computing. By combining Vultr's global cloud infrastructure platform with the performance and flexibility of NVIDIA Blackwell GPUs, we are empowering enterprises and AI innovators to develop and deploy AI precisely where it's needed."

As artificial intelligence continues to gain traction in various industries, the need for robust computing power becomes more critical. The NVIDIA HGX B200, part of this technological advancement, addresses this requirement by facilitating the transition from training AI models to inference. It is designed with the latest generation large language models in mind, enabling efficient and cost-effective utilisation of AI technologies.

The specifications of the NVIDIA HGX B200 include the second generation NVIDIA Transformer Engine, which incorporates Blackwell Tensor Core technology. This technological development allows the GPUs to double the performance and size of models, providing 3x faster training and 15x faster real-time inference compared to its predecessor, the NVIDIA Hopper. Furthermore, the HGX B200 reduces energy consumption by up to 12 times for inference tasks compared to the Hopper generation.

Diverse applications across sectors such as data analytics, high-performance computing, and scientific simulations are supported by the NVIDIA HGX B200. Vultr's deployment tools further simplify integrating these GPUs into existing systems, enabling quick provisioning and scalability for organisations grappling with evolving computational needs.

"As generative AI adoption accelerates, the demand for advanced computing power to support AI models transitioning from training to inference is growing rapidly," said Lee Sustar, Principal Analyst at Forrester. "Cloud infrastructure providers like Vultr will play a pivotal role in meeting this demand, empowering customers with seamless access to the resources necessary to drive high-performance computing and fuel the next wave of generative AI innovation worldwide."

During the NVIDIA GTC, further showcase of Vultr's capabilities will include support for NVIDIA NeMo and NVIDIA Blueprints running on the HGX B200. This is set to enable enterprises across various industries to capitalise on the technological potential of autonomous and intelligent AI systems, opening new opportunities for development and operational improvement.

Dave Salvator, Director of Accelerated Computing Products at NVIDIA, stated, "Achieving faster, real-time inference is becoming a critical differentiator for organisations aiming to rapidly scale their AI initiatives. The addition of the NVIDIA HGX B200 to the Vultr Cloud GPU lineup marks a significant milestone in our long-standing collaboration with Vultr. By making NVIDIA Blackwell GPUs accessible to both innovators and enterprises, we are empowering AI pioneers worldwide to accelerate generative and agentic AI initiatives on a global scale."

This addition of the NVIDIA HGX B200 complements Vultr's existing lineup of NVIDIA GPU offerings, which includes models such as the NVIDIA GH200, HGX H100, A100, and others. The integration of these GPUs aims to further bolster Vultr's array of high-performance cloud solutions.