AMD

The company’s strategy centers on delivering “full-stack” compute—combining AI accelerators, general-purpose CPUs, networking, and software—into server and rack-scale systems for cloud, enterprise, and edge deployments. For AI workloads, AMD’s portfolio includes its AMD Instinct™ data center GPUs (e.g., Instinct MI350 series), AMD EPYC™ server CPUs, and AMD Pensando™ scale-out networking products. AMD also packages these components into integrated rack-scale offerings such as AMD Helios™, which is designed as open rack-scale AI infrastructure for frontier AI training and inference. AMD complements its hardware with its ROCm™ open software stack and new developer experiences under ROCm.ai (including ROCm.AI’s “ROCm.AI” developer experience and related ROCm 10 updates). Commercially, AMD primarily monetizes through product sales (accelerators, CPUs, networking, and related systems/boards) to original equipment manufacturers (OEMs), cloud service providers, enterprise system builders, and hyperscalers. A key part of its current positioning is the ramp of integrated rack-scale platforms (Helios) alongside GPU/CPU platform launches and strategic partnerships with major AI infrastructure customers and ecosystems. Current strategic position (as reflected in AMD’s recent public disclosures) is characterized by: (1) scaling data center revenue and profitability, (2) moving from individual components toward deployment of rack-scale systems, and (3) emphasizing accelerator + system + software co-optimization for AI training and inference.