Vultr scales AI infrastructure with new AMD GPUs

Author: Joe Peck

Vultr, a privately held cloud infrastructure company, has announced plans to offer the AMD Instinct MI455X GPU, alongside support for the AMD Helios rack-scale architecture, with pre-orders now open ahead of availability in Q4.

The cloud infrastructure provider says the platform is designed for AI training, inference, fine tuning, agentic AI, and high-performance computing (HPC).

According to the company, the infrastructure is intended to support organisations moving AI workloads from development into production, particularly those requiring high-memory, rack-scale computing.

J.J. Kardwell, CEO of Vultr, comments, “Customers are rapidly moving from AI experimentation into production and, increasingly, that means agentic and inference-heavy workloads that require infrastructure built for scale.

“The AMD Instinct MI455X and Helios rack-scale architectures give our customers the flexibility, control, and price-to-performance they need to scale their AI initiatives. Making this architecture available across Vultr’s global infrastructure enables teams to build and deploy without constraints.”

AI infrastructure targets large-scale workloads

The AMD Instinct MI455X GPU is based on AMD’s CDNA architecture and features HBM4 memory, increased memory bandwidth, and support for low-precision AI data types. The Helios rack-scale architecture supports up to 72 GPUs in a single deployment.

Vultr says the platform will also support direct liquid cooling to improve thermal management and energy efficiency in higher-density deployments. The infrastructure is complemented by AMD’s ROCm software platform and Pensando networking technology, which are designed to improve AI training, inference performance, and scalability.

IDC forecasts a tenfold increase in the number and complexity of enterprise AI agents over the next five years, increasing demand for computing infrastructure capable of supporting repeated inference workloads.

Dave McCarthy, Group Vice President, Cloud and Datacenter Infrastructure at IDC, notes, “As enterprises transition into the agentic era, the massive influx of complex, data-intensive workflows requires a fundamental shift in infrastructure design.

“Vultr’s deployment of the AMD Instinct MI455X GPU provides the massive memory capacity and rack-scale efficiency necessary to support the continuous inference loops that define next-generation AI workloads.”

Vultr says reserved capacity for deployments based on the AMD Instinct MI455X GPU and AMD Helios architectures is now available for pre-order, with deployments planned from 2027 through to 2028.