AMD is significantly scaling up its hardware strategy for the era of artificial intelligence. Under the codename "Gorgon Halo," the company is developing the Ryzen AI Max(+) 400 series, a lineup of high-end SoCs designed to blur the line between traditional processors and specialized AI accelerators. At the heart of this development is the Ryzen AI Max+ Pro 495, specifically engineered to meet the demanding needs of AI professionals.
The most striking feature of the new chip is the massive expansion of its unified system memory. While previous models were capped at 128 GB, the Ryzen AI Max+ Pro 495 now supports up to 192 GB of LPDDR5X RAM—a 50 percent increase. This substantial capacity is critical for the local execution of Large Language Models (LLMs), which require vast amounts of memory to operate efficiently without relying on external cloud resources.
Technically, the series is powered by a cutting-edge architecture featuring up to 16 Zen 5 CPU cores and integrated Radeon 8065S graphics based on RDNA 3.5. While clock speed improvements over the Strix Halo series appear modest, AMD has optimized the efficiency and power of the integrated NPU (Neural Processing Unit), increasing its performance from 50 to 60 TOPS (Tera Operations Per Second).
This release underscores AMD's commitment to integrating high-performance computing and AI capabilities directly into professional processing units. By combining a high core count, powerful graphics, and an unprecedented memory offering, AMD is positioning itself as a formidable option for developers and data analysts who require the portability of compact workstations without sacrificing AI performance.