{"product_id":"asus-ascent-gx10-ai-supercomputer-nvidia-gb10-superchip","title":"ASUS Ascent GX10 AI Supercomputer (NVIDIA GB10 Superchip)","description":"\u003cp\u003eThe ASUS Ascent GX10 puts the NVIDIA GB10 Grace Blackwell Superchip in a stackable, toolless desktop chassis. It offers DGX Spark-class local compute for ML engineers and small AI teams, with 128GB of unified memory. It sits on a desk alone or stacks with a second unit when a project outgrows one machine's memory.\u003c\/p\u003e\u003ch2\u003eWho it's for\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003eFine-tuning, inference and prototyping with 70B-parameter models in FP4\/INT4 quantization\u003c\/li\u003e\n\u003cli\u003eRunning 30B models at higher precision with a long context window\u003c\/li\u003e\n\u003cli\u003eAgentic AI development with OpenClaw and NVIDIA NemoClaw\u003c\/li\u003e\n\u003cli\u003eTeams pooling two units' unified memory over ConnectX-7 200GbE\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eKey features\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eNVIDIA GB10 Superchip:\u003c\/strong\u003e An Arm v9.2-A 20-core CPU (10x Cortex-X925 + 10x Cortex-A725) joined to an integrated Blackwell GPU with 5th-gen Tensor Cores and 4th-gen RT Cores over NVLink-C2C, which ASUS states gives roughly 5x the bandwidth of PCIe Gen 5, rated up to 1 petaFLOP (1000 AI TOPS) of FP4 performance.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003e128GB unified LPDDR5x at 273GB\/s:\u003c\/strong\u003e One coherent memory pool at 8533MT\/s shared by CPU and GPU, so large language models fit that would exceed the VRAM of most discrete workstation GPUs.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003e1TB, 2TB or 4TB NVMe storage:\u003c\/strong\u003e A user-upgradeable M.2 2242 SSD reached through the base grille.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eDual-port NVIDIA ConnectX-7 (200GbE):\u003c\/strong\u003e Networking for clustering multiple GX10 units, plus a dedicated 10GbE RJ-45 port.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eToolless, stackable chassis:\u003c\/strong\u003e The power switch is on the front rather than the rear panel, which helps when units are stacked or rack-mounted.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eDual-fan cooling:\u003c\/strong\u003e Active cooling engineered to stay quiet under sustained AI workloads, with improved thermal headroom versus comparably sized systems according to ASUS.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNVIDIA DGX OS:\u003c\/strong\u003e Ubuntu-based, with CUDA, PyTorch and TensorFlow pre-installed, and support for agentic AI frameworks including OpenClaw and NVIDIA NemoClaw.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eConnectivity:\u003c\/strong\u003e 3x USB-C 3.2 Gen2x2 (20Gbps) carrying DisplayPort 2.1, 1x USB-C with 180W PD 3.1 EPR power delivery, HDMI 2.1, Wi-Fi 7 and Bluetooth 5.4.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eSpecifications\u003c\/h2\u003e\u003ctable\u003e\n\u003cthead\u003e\u003ctr\u003e\n\u003cth\u003eSpec\u003c\/th\u003e\n\u003cth\u003eDetail\u003c\/th\u003e\n\u003c\/tr\u003e\u003c\/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eSuperchip\u003c\/td\u003e\n\u003ctd\u003eNVIDIA GB10 Grace Blackwell\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCPU\u003c\/td\u003e\n\u003ctd\u003eArm v9.2-A, 20-core (10x Cortex-X925 + 10x Cortex-A725)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eGPU\u003c\/td\u003e\n\u003ctd\u003eIntegrated NVIDIA Blackwell, 5th Gen Tensor Cores, 4th Gen RT Cores\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCPU-GPU interconnect\u003c\/td\u003e\n\u003ctd\u003eNVLink-C2C\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eAI performance\u003c\/td\u003e\n\u003ctd\u003eUp to 1 petaFLOP \/ 1000 AI TOPS (FP4, with sparsity)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMemory\u003c\/td\u003e\n\u003ctd\u003e128GB LPDDR5x, 8533MT\/s, 273GB\/s bandwidth\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eStorage\u003c\/td\u003e\n\u003ctd\u003e1TB \/ 2TB (PCIe Gen4) or 4TB (PCIe Gen5) NVMe SSD, user-upgradeable M.2 2242\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eNetworking\u003c\/td\u003e\n\u003ctd\u003eDual-port NVIDIA ConnectX-7 (200GbE), 10 Gigabit Ethernet (RJ-45)\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eWireless\u003c\/td\u003e\n\u003ctd\u003eWi-Fi 7, Bluetooth 5.4\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePorts\u003c\/td\u003e\n\u003ctd\u003e3x USB-C 3.2 Gen2x2 (20Gbps) with DisplayPort 2.1, 1x USB-C with 180W PD 3.1 EPR, 1x HDMI 2.1\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eOperating system\u003c\/td\u003e\n\u003ctd\u003eNVIDIA DGX OS (Ubuntu-based), CUDA\/PyTorch\/TensorFlow pre-installed\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eAgentic AI support\u003c\/td\u003e\n\u003ctd\u003eOpenClaw, NVIDIA NemoClaw compatible\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower supply\u003c\/td\u003e\n\u003ctd\u003e240W external USB-C adapter\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eDimensions\u003c\/td\u003e\n\u003ctd\u003e150 x 150 x 51mm\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eWeight\u003c\/td\u003e\n\u003ctd\u003e1.48kg\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCooling\u003c\/td\u003e\n\u003ctd\u003eDual-fan active cooling\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eChassis\u003c\/td\u003e\n\u003ctd\u003eToolless, stackable, front-mounted power switch\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\n\u003c\/table\u003e\u003ch2\u003eGood to know\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003eThe 1 petaFLOP \/ 1000 TOPS figure is a peak FP4 number with sparsity, so real dense-precision throughput on a given model will be lower.\u003c\/li\u003e\n\u003cli\u003eThis is a fine-tuning, inference and prototyping platform, not a from-scratch foundation-model trainer. The 140W GB10 package is cooled by the dual-fan design.\u003c\/li\u003e\n\u003cli\u003eThe standard DGX OS and CUDA stack means workloads developed here are portable to other DGX Spark-class systems or DGX Cloud.\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eChoose the GX10 if you want a GB10 system with a front power switch, an upgradeable M.2 slot and a stackable design. It runs Arm, so it suits teams working in the CUDA ecosystem.\u003c\/p\u003e","brand":"ASUS","offers":[{"title":"1TB PCIe NVMe SSD","offer_id":58232540168574,"sku":"ASUS-ASCENT-GX10-1TB","price":5899.0,"currency_code":"USD","in_stock":true},{"title":"2TB PCIe NVMe SSD","offer_id":58232540201342,"sku":"ASUS-ASCENT-GX10-2TB","price":6499.0,"currency_code":"USD","in_stock":true},{"title":"4TB PCIe NVMe SSD","offer_id":58232540234110,"sku":"ASUS-ASCENT-GX10-4TB","price":7899.0,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/1022\/4468\/0062\/files\/new-asus-ascent-gx10-1_53208dfd-244d-4433-8f44-0aeb03890e20.jpg?v=1792094099","url":"https:\/\/computahardware.net\/products\/asus-ascent-gx10-ai-supercomputer-nvidia-gb10-superchip","provider":"ComputaHardware","version":"1.0","type":"link"}