Ionstream.ai announced availability of NVIDIA HGX B200 bare metal instances aimed at Artificial Intelligence workloads. The company positions the offering as a high-performance, scalable option for developers, researchers, and enterprises, delivered through its GPU-as-a-Service platform. The article references an hourly rate token of ´?.50 per hour´ for one-month contracts and says lower pricing is available for longer-term commitments; the precise hourly price is not stated.
The announcement highlights the B200 as the most advanced GPU built for Artificial Intelligence, based on NVIDIA´s Blackwell architecture. According to the article, the B200 targets transformer models, diffusion systems, multimodal Artificial Intelligence and reinforcement learning agents, and is described as energy-efficient and scalable. Ionstream emphasizes that its bare metal platform provides direct access to raw GPU power without the overhead of virtualization, enabling real-time provisioning and high-performance workloads.
Jeff Hinkle, founder and chief executive officer of Ionstream, is quoted describing the launch as an effort to democratize access to top-tier hardware. The article notes Hinkle´s prior roles in shaping data infrastructure through GNAX and Net Depot and frames the B200 offering as part of Ionstream´s service-first approach, which the company says includes deep vendor relationships, seamless scaling and support. The B200 is available through Ionstream´s GPU-as-a-Service platform, and readers are directed to the company´s website for more information.
