{"product_id":"amd-ryzen-ai-halo-eclipse","title":"EMARQUE AI Eclipse AMD Ryzen AI Max+ 395 Local AI Workstation","description":"\u003ch2\u003eEMARQUE AI Eclipse: 128GB of unified memory for local AI, built in Malaysia\u003c\/h2\u003e\n\u003cp\u003eThe EMARQUE AI Eclipse brings large-memory local AI computing into a compact chassis. It runs the AMD Ryzen™ AI Max+ 395 with 16 Zen 5 CPU cores, Radeon™ 8060S graphics and 128GB of LPDDR5X unified memory, for running large language models, AI agents, development workloads and private AI applications on hardware you own. Available in Windows and Linux configurations.\u003c\/p\u003e\n\n\u003ch3\u003eKey features\u003c\/h3\u003e\n\u003cul class=\"metafield-list\"\u003e\n\u003cli\u003e\n\u003cstrong\u003eAMD Ryzen AI Max+ 395\u003c\/strong\u003e: a 16-core, 32-thread Zen 5 processor built for AI, development and workstation workloads.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eAMD Radeon 8060S graphics\u003c\/strong\u003e: an integrated RDNA™ 3.5 GPU with 40 compute units for GPU-accelerated AI and graphics.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003e128GB unified memory\u003c\/strong\u003e: LPDDR5X-8000 shared between the CPU and GPU, a far larger working pool than conventional consumer graphics cards offer.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eUp to 96GB assignable to graphics\u003c\/strong\u003e: on the 128GB configuration, up to 96GB can be allocated to GPU workloads, so larger models fit entirely in memory.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eBuilt for local AI\u003c\/strong\u003e: run LLMs, coding models, AI agents, RAG systems, image generation and computer vision directly on the device.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eWindows or Linux\u003c\/strong\u003e: Windows for a familiar desktop, Linux for a developer-focused ROCm environment.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eDual 10GbE networking\u003c\/strong\u003e: two 10 Gigabit Ethernet ports for NAS connectivity, local AI serving and multi-system deployments.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eCompact form factor\u003c\/strong\u003e: workstation capability in a chassis measuring approximately 70 x 197.8 x 200 mm.\u003c\/li\u003e\n\u003c\/ul\u003e\n\n\u003ch3\u003eTechnical specifications\u003c\/h3\u003e\n\u003cul class=\"metafield-list\"\u003e\n\u003cli\u003eModel: EMARQUE AI Eclipse\u003c\/li\u003e\n\u003cli\u003eODM platform: SixUnited AXB35-02-H05-YC\u003c\/li\u003e\n\u003cli\u003eProcessor: AMD Ryzen AI Max+ 395, Zen 5, 16 cores and 32 threads\u003c\/li\u003e\n\u003cli\u003eMaximum boost clock: up to 5.1GHz\u003c\/li\u003e\n\u003cli\u003eCache: 16MB L2, 64MB L3\u003c\/li\u003e\n\u003cli\u003eProcessor TDP range: 45W to 120W\u003c\/li\u003e\n\u003cli\u003eIntegrated graphics: AMD Radeon 8060S, RDNA 3.5, 40 compute units, up to 2.9GHz\u003c\/li\u003e\n\u003cli\u003eNPU: AMD XDNA 2, up to 50 TOPS\u003c\/li\u003e\n\u003cli\u003ePlatform AI performance: up to 126 TOPS combined\u003c\/li\u003e\n\u003cli\u003eSystem memory: 128GB LPDDR5X-8000, on-board, unified\u003c\/li\u003e\n\u003cli\u003eMemory interface: 256-bit, up to 256GB\/s bandwidth\u003c\/li\u003e\n\u003cli\u003eMaximum configurable graphics memory: up to 96GB\u003c\/li\u003e\n\u003cli\u003eStorage: PCIe NVMe SSD, capacity depends on the selected configuration\u003c\/li\u003e\n\u003cli\u003eEthernet: 2 x 10GbE RJ45\u003c\/li\u003e\n\u003cli\u003eWi-Fi and Bluetooth: not installed, this is a wired LAN-only workstation by design\u003c\/li\u003e\n\u003cli\u003eHigh-speed connectivity: USB4 \/ USB-C, plus USB Type-A\u003c\/li\u003e\n\u003cli\u003eDisplay output: HDMI, DisplayPort and compatible USB-C display output\u003c\/li\u003e\n\u003cli\u003ePower: external 240W AC adapter, 100 to 240V, 50\/60Hz\u003c\/li\u003e\n\u003cli\u003eDimensions: approximately 70 x 197.8 x 200 mm\u003c\/li\u003e\n\u003cli\u003eOperating system: Windows 11 or Linux\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003e\u003cem\u003eAvailable ports are based on the EMARQUE AXB35-02-H05-YC configuration.\u003c\/em\u003e\u003c\/p\u003e\n\n\u003ch3\u003eWhy 128GB of unified memory matters\u003c\/h3\u003e\n\u003cp\u003eA conventional desktop separates system RAM from dedicated GPU VRAM. The Ryzen AI Max+ 395 instead uses high-speed unified memory addressable across the processor and the integrated Radeon GPU. With 128GB installed, up to 96GB can be allocated to graphics workloads. For local AI, that makes it possible to work with models that would exceed the VRAM of a 16GB or 24GB consumer graphics card.\u003c\/p\u003e\n\n\u003ch3\u003eLocal AI model capability\u003c\/h3\u003e\n\u003cp\u003eAMD states that this platform, with 128GB of unified memory, holds models of up to 200 billion parameters in device memory. In practice the useful classes are:\u003c\/p\u003e\n\u003cul class=\"metafield-list\"\u003e\n\u003cli\u003e\n\u003cstrong\u003e7B models\u003c\/strong\u003e: fast local assistants, coding and lightweight AI workloads.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003e14B to 32B models\u003c\/strong\u003e: a strong balance of capability and local inference speed.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003e70B-class models\u003c\/strong\u003e: higher-quality reasoning, coding and general-purpose local AI.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003e100B and above\u003c\/strong\u003e: selected quantised and Mixture-of-Experts models, depending on architecture, quantisation, runtime and context length.\u003c\/li\u003e\n\u003c\/ul\u003e\n\u003cp\u003eActual performance depends on model architecture, quantisation, context length, runtime, memory allocation, operating system and active parameter count. A model with more parameters is not necessarily slower than a smaller one, particularly with Mixture-of-Experts architectures.\u003c\/p\u003e\n\n\u003ch3\u003eWhat you can run\u003c\/h3\u003e\n\u003cp\u003eLocal large language models, coding assistants, AI coding agents, AI agents, private document analysis, Retrieval-Augmented Generation, local knowledge bases, OpenAI-compatible local APIs, image generation, computer vision, speech and transcription, AI development, Python and PyTorch development, local inference servers, business AI applications and edge AI workloads.\u003c\/p\u003e\n\n\u003ch3\u003eSupported AI ecosystem\u003c\/h3\u003e\n\u003cp\u003eDepending on the operating system and workload, Eclipse works with llama.cpp, Ollama, LM Studio, Lemonade, AMD ROCm, HIP, PyTorch, ComfyUI and OpenAI-compatible local APIs. Software compatibility and performance vary by operating system, runtime, model format and software version.\u003c\/p\u003e\n\n\u003ch3\u003eWindows or Linux\u003c\/h3\u003e\n\u003cp\u003e\u003cstrong\u003eWindows\u003c\/strong\u003e suits users who want local AI without leaving a familiar desktop: AI assistants, coding and development, creator and productivity applications, gaming, and Windows-compatible AI tools such as LM Studio, Ollama and llama.cpp. The Windows configuration works as both an everyday PC and a local AI workstation.\u003c\/p\u003e\n\u003cp\u003e\u003cstrong\u003eLinux\u003c\/strong\u003e suits developers and AI practitioners building dedicated systems: AMD ROCm development, PyTorch, inference, containers, automation, local AI servers, API deployments, RAG systems and backend workloads. Linux currently gives access to the broader AMD ROCm ecosystem and is generally the better choice for advanced AI development and deployment.\u003c\/p\u003e\n\n\u003ch3\u003eUse it as a local AI server\u003c\/h3\u003e\n\u003cp\u003eEclipse can operate as an inference server on your own network. Compatible inference engines expose models over REST or OpenAI-compatible APIs, so other applications and devices can use models running on the Eclipse: internal AI assistants, company knowledge bases, coding assistants, AI automation, document processing, chatbot interfaces and private API endpoints. The dual 10GbE interfaces let it sit inside high-speed wired and storage networks.\u003c\/p\u003e\n\n\u003ch3\u003eHow Eclipse relates to the AMD Ryzen AI Halo\u003c\/h3\u003e\n\u003cp\u003eAMD sells its own developer box on this platform, the AMD Ryzen AI Halo. Eclipse runs the same Ryzen AI Max+ 395 silicon with the same 128GB of LPDDR5X-8000 unified memory and the same 256GB\/s of bandwidth, so it carries the same 200 billion parameter platform capability AMD states. Eclipse differs in being configured, built and supported in Malaysia by EMARQUE, with dual 10GbE networking and a choice of Windows or Linux at handover.\u003c\/p\u003e\n\n\u003cp\u003ePart of the \u003ca href=\"\/collections\/amd-ryzen-ai-max\"\u003eAMD Ryzen AI Max\u003c\/a\u003e range, the Windows-native x86 route to on-device AI. If you need a CUDA-first Linux platform instead, EMARQUE also supplies \u003ca href=\"\/collections\/nvidia-dgx-spark\"\u003eNVIDIA DGX Spark\u003c\/a\u003e.\u003c\/p\u003e\n\n\u003ch3\u003eImportant information\u003c\/h3\u003e\n\u003cp\u003eAI model compatibility and performance vary with model architecture, format, quantisation, runtime, context length and software configuration. References to maximum model parameter sizes describe memory and platform capability and are not a guarantee of inference performance for every model. Some AI software requires additional installation, configuration or dependencies. ROCm, AI runtime and application support differ between Windows and Linux. EMARQUE AI Eclipse ships without Wi-Fi or Bluetooth; network connectivity is provided by its dual 10GbE Ethernet interfaces.\u003c\/p\u003e\n\u003cp\u003e\u003cem\u003eConfiguration and pricing confirmed on enquiry. Built and supported in Malaysia by EMARQUE.\u003c\/em\u003e\u003c\/p\u003e","brand":"EMARQUE","offers":[{"title":"Default Title","offer_id":52463795798296,"sku":"EMQ-ECLIPSE-395-128GB","price":18999.0,"currency_code":"MYR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0735\/0907\/7272\/files\/Group_2_3260f613-3cf1-42f9-9a55-88f8483c1a58.png?v=1789716028","url":"https:\/\/emarque.co\/products\/amd-ryzen-ai-halo-eclipse","provider":"EMARQUE PC Malaysia","version":"1.0","type":"link"}