GMAsia
    🇨🇳中国·AI 新闻·2026年8月24日·来源: Gizmochina

    Xiaomi announces AI Cube mini-PC with XRING O3, O100, and D100 to run LLMs locally

    内容仅提供英文版本

    Xiaomi has unveiled the AI Cube, a prototype mini-PC designed to run large language models (LLMs) locally. The device was introduced at the XRING Chip Technology Communication Conference and features three proprietary chips: the XRING O3, O100, and D100. The XRING O3 serves as the main AI SoC, incorporating a 10-core CPU, a 16-core G2-Ultra NX GPU, and a 200 TOPS NPU. The XRING O100 is a high-bandwidth AI acceleration chip with 1.22TB/s near-memory computing bandwidth, while the XRING D100 is a 3nm chip for intelligent driving, supporting up to 160GB of memory. The AI Cube aims to deploy models ranging from 3B to 120B parameters on-device, though Xiaomi has not announced a release date or price.

    Nexa 摘要

    Xiaomi’s AI Cube prototype points to a significant strategic direction for the company in Asia’s competitive AI hardware market. By developing its own XRING O3, O100, and D100 chips, Xiaomi is moving to control the full stack for local LLM deployment, a move that could differentiate it from competitors relying on third-party silicon. The ability to run models up to 120B parameters locally offers a compelling proposition for enterprise and specialized applications across the region, particularly in markets prioritizing data privacy or low-latency processing. However, the lack of a release date or pricing suggests this is still a technical demonstration rather than an imminent product. The real test for Xiaomi will be whether it can transition this prototype into a commercially viable device that can compete with existing solutions and attract developers to its proprietary chip architecture. The company’s success in integrating its growing suite of text, code, and voice LLMs with this hardware will be crucial for its broader AI ambitions in Asia.

    #xiaomi#news
    深入了解
    原文报道: Gizmochina我们不转载全文, 阅读原文 →

    相关文章

    6 则