The AmpereOne family consists of 64-bit multi-core processor designed for AI workloads and cloud native environments. Building on the strong foundation of the Ampere® Altra® family of processors, AmpereOne and AmpereOne M processors offer enhanced performance, efficiency, and scalability tailored to modern cloud and AI needs.

Key architectural aspects include: 

• Cloud Native Design: Single-threaded cores running up to 3.6 GHz for consistent and predictable performanceCustom-designed, single-threaded cores ranging from 96 to 192 with specialized vector units supporting AI data types (FP16, Bfloat16, Int8, Int16) and integration with Ampere’s AI Optimizer (AIO)
• Dedicated Caching: each core has is its own 16KB L1 instruction-cache, 64KB L1 data-cache, and 2MB L2 private cache.
• Memory and I/O:
• The AmpereOne processor offers 8 channels DDR5-5200 (up to 4TB), 128 PCIe Gen5 lanes
• The AmpereOne M rovides 12 channels DDR5-5600 (up to 3TB), 96 PCIe Gen5 lanes, optimized for generative AI and memory-intensive workloadsThe AmpereOne® family of processors deliver scalable, power-efficient performance suitable for environments from edge devices to hyperscale data centers. They help organizations maximize performance per watt and dollar while supporting sustainability and cost-effective infrastructure. 

The AmpereOne family of processors delivers clear advantages for cloud datacenter managers, software developers. and IT decision makers:

• Predictable performance: With single-threaded cores, large private caches, and consistent clock speeds, AmpereOne reduces the “noisy neighbor” effect common in multi-core environments. This helps ensure steady performance for tenants and more reliable service delivery—an important factor in shared cloud infrastructure.
• High Density and scalability: AmpereOne processors support higher core counts and more servers per rack, maximizing infrastructure efficiency and enabling high VM density. This translates to strong tokens-per-second performance for AI inference and helps maximize instance, session, or agent density for AI-powered digital services. Meanwhile, AmpereOne M’s enhanced memory bandwidth and capacity support running multiple AI models concurrently on a single system. 
• Power Efficiency and Sustainability: Designed for performance per rack, these processors reduce power consumption, rack space, and operational costs, helping organizations meet sustainability targets without compromising capabilities..
• Robust Ecosystem Support: A broad, mature software ecosystem surrounds the AmpereOne family—including platform partners, operating systems, programming languages, and applications. Developers also benefit from resources like tuning guides, tutorials, and reference architectures to build optimized, power-efficient solutions.

• Empowering Developers and Fueling AI Adoption with AmpereOne® Design Innovations
• Ampere Solutions for AI and the Sustainable Cloud
• How to Choose the Best Processor for Artificial Intelligence (AI)
• Customer Reference Boards
• AmpereOne A192-32X Benchmarks: 192 Core ARM Server Performance & Power Efficiency