{"product_id":"gigabyte-ai-top-atom-4tb-gen5-gb10-dgx","title":"GIGABYTE AI TOP ATOM 4TB GEN5 GB10 DGX","description":"\u003ch2\u003eGigabyte AI Top Atom DGX GB10 4TB PCIe 5.0 Personal AI Supercomputer\u003c\/h2\u003e\u003cp\u003eThe \u003cstrong\u003eGigabyte AI Top Atom ATAGB10-9000\u003c\/strong\u003e is a compact personal AI supercomputer built around the \u003cstrong\u003eNVIDIA GB10 Grace Blackwell Superchip\u003c\/strong\u003e platform. It combines an NVIDIA Blackwell GPU, Grace CPU cores and a large unified memory pool to run demanding AI development, inference and prototyping workloads locally.\u003c\/p\u003e\u003cp\u003eWith \u003cstrong\u003e128GB coherent LPDDR5X unified memory\u003c\/strong\u003e, \u003cstrong\u003e4TB PCIe 5.0 NVMe storage\u003c\/strong\u003e and up to \u003cstrong\u003e1 petaFLOP of FP4 AI performance\u003c\/strong\u003e, the system is designed for developers, researchers, data scientists, students and technical teams that need a desk-friendly platform for modern generative AI workflows.\u003c\/p\u003e\u003ch3\u003eKey Features\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003eNVIDIA GB10 Grace Blackwell Superchip\u003c\/li\u003e\n\u003cli\u003e20-core Arm CPU with 10 Cortex-X925 and 10 Cortex-A725 cores\u003c\/li\u003e\n\u003cli\u003eNVIDIA Blackwell GPU with fifth-generation Tensor Cores\u003c\/li\u003e\n\u003cli\u003e128GB coherent LPDDR5X unified memory\u003c\/li\u003e\n\u003cli\u003e4TB PCIe 5.0 NVMe storage\u003c\/li\u003e\n\u003cli\u003e1 petaFLOP of FP4 AI performance\u003c\/li\u003e\n\u003cli\u003eSupports AI models with up to approximately 200 billion parameters\u003c\/li\u003e\n\u003cli\u003e10Gb Ethernet and NVIDIA ConnectX-7 high-speed networking\u003c\/li\u003e\n\u003cli\u003eWi-Fi 7 and Bluetooth 5.4\u003c\/li\u003e\n\u003cli\u003e4 x USB 3.2 Gen 2x2 Type-C, HDMI and Ethernet\u003c\/li\u003e\n\u003cli\u003eNVIDIA DGX OS based on Ubuntu Linux\u003c\/li\u003e\n\u003cli\u003eCompact 150 x 150 x 50.5mm chassis\u003c\/li\u003e\n\u003cli\u003e2 years manufacturer warranty\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eGrace Blackwell AI Platform\u003c\/h3\u003e\u003cp\u003eThe integrated \u003cstrong\u003eNVIDIA GB10 Grace Blackwell Superchip\u003c\/strong\u003e combines 20-core Arm CPU with 10 Cortex-X925 and 10 Cortex-A725 cores with NVIDIA Blackwell GPU with fifth-generation Tensor Cores. This tightly integrated architecture is designed to move data efficiently between CPU, GPU and memory during AI inference, development and experimentation.\u003c\/p\u003e\u003ch3\u003eUnified Memory for Large Models\u003c\/h3\u003e\u003cp\u003eThe system includes \u003cstrong\u003e128GB coherent LPDDR5X unified memory\u003c\/strong\u003e of coherent unified memory. Supports AI models with up to approximately 200 billion parameters, allowing developers to work with models that would exceed the memory capacity of many conventional consumer graphics cards.\u003c\/p\u003e\u003ch3\u003eAI Performance\u003c\/h3\u003e\u003cp\u003eAI compute performance reaches up to \u003cstrong\u003e1 petaFLOP of FP4 AI performance\u003c\/strong\u003e. Real-world throughput depends on model architecture, precision, framework, optimisation, software version and workload.\u003c\/p\u003e\u003ch3\u003eFast Local Storage\u003c\/h3\u003e\u003cp\u003e\u003cstrong\u003e4TB PCIe 5.0 NVMe storage\u003c\/strong\u003e provides local capacity for models, datasets, containers, development tools and project files.\u003c\/p\u003e\u003ch3\u003eNetworking and Expansion\u003c\/h3\u003e\u003cp\u003e10Gb Ethernet and NVIDIA ConnectX-7 high-speed networking, while Wi-Fi 7 and Bluetooth 5.4. Connectivity includes 4 x USB 3.2 Gen 2x2 Type-C, HDMI and Ethernet. NVIDIA ConnectX-7 support enables high-speed links for supported development and multi-system workflows.\u003c\/p\u003e\u003ch3\u003eSoftware Platform\u003c\/h3\u003e\u003cp\u003eThe system is supplied with \u003cstrong\u003eNVIDIA DGX OS based on Ubuntu Linux\u003c\/strong\u003e, providing access to the NVIDIA AI software ecosystem and a Linux-based development environment for compatible frameworks, containers and tools.\u003c\/p\u003e\u003ch3\u003eIdeal For\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003eLocal generative AI development and inference\u003c\/li\u003e\n\u003cli\u003eLarge language model experimentation\u003c\/li\u003e\n\u003cli\u003eAI prototyping and model fine-tuning\u003c\/li\u003e\n\u003cli\u003eData science and research\u003c\/li\u003e\n\u003cli\u003eComputer vision and robotics development\u003c\/li\u003e\n\u003cli\u003eTechnical education and software development\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eCore Specification\u003c\/h3\u003e\u003ctable style=\"width:100%; border-collapse:collapse;\"\u003e\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eAI Platform\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eNVIDIA GB10 Grace Blackwell Superchip\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eCPU\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e20-core Arm CPU with 10 Cortex-X925 and 10 Cortex-A725 cores\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eGPU\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eNVIDIA Blackwell GPU with fifth-generation Tensor Cores\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eUnified Memory\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e128GB coherent LPDDR5X unified memory\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eAI Performance\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e1 petaFLOP of FP4 AI performance\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eSupported Model Scale\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eSupports AI models with up to approximately 200 billion parameters\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eStorage\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e4TB PCIe 5.0 NVMe storage\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eOperating System\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eNVIDIA DGX OS based on Ubuntu Linux\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\u003c\/table\u003e\u003ch3\u003eConnectivity and Physical Specification\u003c\/h3\u003e\u003ctable style=\"width:100%; border-collapse:collapse;\"\u003e\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eWired Networking\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e10Gb Ethernet and NVIDIA ConnectX-7 high-speed networking\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eWireless\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eWi-Fi 7 and Bluetooth 5.4\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003ePorts\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e4 x USB 3.2 Gen 2x2 Type-C, HDMI and Ethernet\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eDimensions\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e150 x 150 x 50.5mm\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eWeight\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eapproximately 1.2kg\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\u003c\/table\u003e\u003ch3\u003eWarranty and Product Information\u003c\/h3\u003e\u003ctable style=\"width:100%; border-collapse:collapse;\"\u003e\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eBrand\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eGigabyte\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eModel\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eAI Top Atom ATAGB10-9000\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eManufacturer Part Number\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003eATAGB10-9000\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eEAN\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e4719331875763\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eManufacturer Warranty\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e2 years\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e\u003cstrong\u003eVIP SKU\u003c\/strong\u003e\u003c\/td\u003e\n\u003ctd style=\"padding:8px; border:1px solid #ddd;\"\u003e129741\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\u003c\/table\u003e\u003ch3\u003ePackage Contents\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003eGigabyte AI Top Atom ATAGB10-9000 AI system\u003c\/li\u003e\n\u003cli\u003ePower adapter\u003c\/li\u003e\n\u003cli\u003eProduct documentation\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003e\u003cstrong\u003ePlease note:\u003c\/strong\u003e Maximum model size and AI performance depend on framework support, quantisation, software optimisation and available system resources. This product is a specialist development platform rather than a conventional Windows desktop PC.\u003c\/p\u003e\u003c!-- Product sources: https:\/\/www.vip-computers.com\/uk\/shop\/product.aspx?product=129741 --\u003e","brand":"VIP","offers":[{"title":"Default Title","offer_id":58498879193472,"sku":"129741","price":6457.98,"currency_code":"GBP","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0902\/7390\/7072\/files\/129741.jpg?v=1784559299","url":"https:\/\/www.pinkparsnipcomputers.co.uk\/products\/gigabyte-ai-top-atom-4tb-gen5-gb10-dgx","provider":"Pink Parsnip Computers","version":"1.0","type":"link"}