Lenovo and AMD Introduce Next-Generation AI Infrastructure with AMD Helios Rack-Scale Architecture

As artificial intelligence infrastructure continues to evolve, AI environments are rapidly shifting from standalone servers to fully integrated AI Factory architectures capable of supporting large-scale AI training and inference workloads. During Lenovo Tech World 2026, Lenovo announced that it is among the first OEMs to adopt the new AMD Helios™ rack-scale solution, designed specifically for hyperscalers and NeoCloud providers building AI at scale.

The new architecture combines next-generation AMD Instinct™ GPUs, AMD EPYC™ processors, AMD Pensando™ networking technologies, and the AMD ROCm™ software ecosystem into an open rack-scale platform based on Open Compute Project (OCP) and Open Rack Wide (ORW) specifications. The solution is engineered to scale from individual racks to large distributed AI clusters supporting hyperscale inference, fine-tuning, and foundation model training.

A single rack configuration includes:

• 72 AMD Instinct™ MI455X GPUs;
• next-generation AMD EPYC™ processors;
• 31 TB of HBM4 high-bandwidth memory;
• AMD Pensando™ high-performance networking technologies.
  
  

According to AMD, the platform delivers up to 2.9 exaFLOPS of FP4 AI inference performance and 1.4 exaFLOPS of FP8 AI training performance, providing the computing power required for training, fine-tuning, and serving increasingly sophisticated generative and agentic AI models.

Beyond the hardware itself, Lenovo is placing significant emphasis on the engineering expertise required to successfully deploy large-scale AI infrastructure. The company offers consulting services for AI cluster design, GPU optimization, liquid cooling implementation, workload tuning, and end-to-end lifecycle management.

According to Lenovo, this approach enables organizations to reduce deployment time, maximize GPU utilization, accelerate time-to-production, and minimize the risks associated with deploying enterprise-scale AI environments.

The solution is part of Lenovo's broader Hybrid AI strategy, which focuses on delivering open, scalable AI infrastructure capable of supporting workloads across edge, enterprise, cloud, and hyperscale environments. Lenovo's goal is to help organizations deploy AI wherever it delivers the greatest business value while providing a flexible foundation for future growth.

Lenovo expects solutions based on the AMD Helios™ rack-scale architecture to become available during the fourth quarter of 2026.