{
  "id": 6558534,
  "title": "LattePanda unveils micro compute module delivering 115 TOPS for AI",
  "url": "https://urgent.news/2026/09/10/lattepanda-unveils-micro-compute-module-delivering-115-tops-for-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-10T07:59:48.000Z",
  "source": {
    "name": "Vietnam Investment Review",
    "slug": "vietnam-investment-review",
    "url": "https://vir.com.vn/lattepanda-unveils-micro-compute-module-delivering-115-tops-for-ai-160490.html"
  },
  "original_language": "en",
  "account": "SHANGHAI, September 9, 2026 - LattePanda Team has introduced its latest micro x86 compute module, the LattePanda Mu Ultra, designed specifically for on-device artificial intelligence (AI) applications. Equipped with Intel Core Ultra 5 226V and Ultra 7 256V processors, this compact module measures just 69.6 × 60mm and delivers up to 115 TOPS (floating-point operations per second) of AI performance.\n\nThe module combines the CPU, GPU, and Neural Processing Unit (NPU) from the Intel Core Ultra processors, offering 47 TOPS from the dedicated NPU. Its small form factor enables integration into space-constrained devices, including handheld instruments, smart cameras, robots, and other intelligent edge systems.\n\nAI workloads can be executed directly on the device with LattePanda Mu Ultra, facilitating low-latency inference, enhanced data privacy, and reduced reliance on cloud computing. In tests, the module achieved text generation speeds of 18 tokens per second with Qwen3.5-9B and 55 tokens per second with Qwen3.5-2B using INT4 quantization with OpenVINO GenAI on the integrated GPU. These results showcase the module's capability to support responsive, local conversational AI applications, such as private voice assistants and offline document processing.\n\nThe LattePanda Mu Ultra features 16GB LPDDR5X-8533 memory, with up to 11.6GB allocated for graphics memory, providing the bandwidth necessary for demanding AI workloads. The module operates with idle power consumption as low as 2.5W, making it suitable for portable and always-on applications. It supports Windows and Linux operating systems, alongside popular AI tools and frameworks like Intel OpenVINO, llama.cpp, and Ollama.\n\nThe module is compatible with existing LattePanda Mu carrier boards and can be integrated into various platforms, such as handheld instruments, smart cameras, autonomous mobile robots, service robots, and on-device vision AI systems. Pricing for the LattePanda Mu Ultra begins at $599 and is available on the company's official store.",
  "summary": "LattePanda launched its Mu Ultra compute module featuring Intel Core Ultra processors delivering up to 115 TOPS of artificial intelligence processing performance in a compact form factor",
  "key_points": [
    "LattePanda introduces micro x86 compute module, LattePanda Mu Ultra.",
    "Module delivers up to 115 TOPS AI performance with Intel Core Ultra processors.",
    "Supports Windows and Linux, priced at $599, for edge AI applications."
  ],
  "editors_take": "This new micro compute module enables faster, more private on-device AI applications, shifting processing from the cloud to edge devices, and potentially transforming industries like robotics and smart cameras.",
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}