aifollow.news
Back IT之家 科技新闻
IT之家 科技新闻· · Original publication time AI score40

Leaked Microsoft AI PC roadmap sets out four hardware tiers, starting at 8GBMachine translation

曝光的微软 AI PC 路线图将硬件分为四档,8GB PC 纳入最低档

AI-assisted summary

A slide shared by @9550pro shows a Microsoft “Hybrid Intelligence” AI PC roadmap ranging from Windows PCs with 8GB of memory to Copilot+ PCs, devices for developers and creators running models locally, and professional platforms with more than 128GB of memory. Windows Central says users of lower-spec PCs could access AI capabilities through the cloud.

Article · AI translation

IT Home reported on November 4 that source @9550pro posted a slide titled “Windows is the home of hybrid intelligence” on X yesterday (November 3), showing Microsoft’s “Hybrid Intelligence” AI PC roadmap.

IT Home previously reported that at a launch event in San Francisco, United States, at 10 a.m. Pacific Time on October 7 (1 a.m. Beijing time on October 8), Microsoft introduced the concept of “hybrid intelligence,” emphasizing the use of a hybrid architecture that combines cloud AI with local computing power to handle user tasks on Windows 11.

According to the leaked “Hybrid Intelligence” AI PC roadmap, Microsoft divides AI hardware into four tiers. The lowest, Tier 0, covers Windows PCs with 8GB of memory and processors including Intel Core 3/4 series, AMD Ryzen 100/200 series, and Snapdragon X or X Plus. Technology outlet Windows Central interpreted this to mean that ordinary users can still access AI capabilities through the cloud, even without large amounts of memory or a top-tier NPU.

Tier 1 consists of Copilot+ PCs with 16GB to 64GB of memory, spanning configurations from Snapdragon X2 Plus to AMD Ryzen AI MAX+ 400.

Tier 2 is aimed at developers and creators who need to build and run AI models locally, with hardware including Intel Core Ultra Next, NVIDIA RTX Spark, and AMD Ryzen AI MAX+ 400.

The highest tier consists of supercomputers with more than 128GB of memory for running frontier large models locally, and includes professional platforms such as Intel Crescent Island and NVIDIA GB300 DGX Station.

Found an error? Send a correction