AMD Pushes AI to the Desktop
AMD is bringing AI to the mainstream PC market, launching its Ryzen AI 400 and PRO 400 desktop chips. The new processors for the AM5 socket feature a dedicated AI engine with up to 50 TOPS, designed for "Copilot+" desktops. This gives PC OEMs and enterprise buyers a credible alternative for running local LLMs and other client-side AI workloads securely and with low latency.
The new Ryzen AI 400 desktop processors are built on the AM5 socket, integrating Zen 5 CPU cores, RDNA 3.5 graphics, and a second-generation XDNA 2 Neural Processing Unit (NPU). This XDNA 2 architecture provides a significant leap in AI performance, offering up to 5x the compute capacity and twice the power efficiency of its predecessor. The NPU's 50 TOPS of performance is specifically designed to meet Microsoft's "Copilot+" PC requirements, which call for a minimum of 40 TOPS to handle advanced on-device AI features. This move places AMD in direct competition with Intel's upcoming Lunar Lake processors, which are expected to feature an NPU with up to 48 TOPS, and Qualcomm's Snapdragon X Elite with its 45 TOPS NPU. Apple's M4 chip also boasts a powerful Neural Engine capable of 38 TOPS. The race for AI dominance on the client side is clearly centered around the capabilities of these dedicated neural processors. The Ryzen AI 400 desktop lineup will initially launch through major OEM partners like HP and Lenovo in the second quarter of 2026, with DIY retail availability expected to follow. The initial offerings will feature up to 8 Zen 5 cores and 16 threads, with the top-tier Ryzen AI 7 450G boosting up to 5.1 GHz. While powerful, these desktop chips currently top out at 8 CPU cores and 8 GPU compute units, whereas the mobile "Strix Point" variants can scale up to 12 CPU cores and 16 RDNA 3.5 compute units. AMD's strategy extends beyond hardware, focusing on an open software ecosystem centered around its ROCm platform to compete with proprietary environments. By enabling developers to run AI models locally, AMD is positioning itself as a key player in the "train in the data center, infer at the edge" model. This approach, championed by CEO Lisa Su, aims to embed AI throughout AMD's entire product portfolio, from data centers to personal computers.
