DALCO GPU Servers offer superior cooling solutions designed to maximize the performance of your selected GPUs. Our custom engineering enables us to effectively cool and power nearly any GPU on the market within a rack-mount chassis. Choose the ideal chassis size, GPU model, and CPU to create a solution precisely optimized for your workload.
The AMD Ryzen AI Halo Developer Platform is the ultimate local AI developer system, empowering researchers, developers, students and creators to build, run and scale generative and agentic AI directly on the desktop, with massive performance and no cloud cost constraints. At the heart of Ryzen AI Halo is the AMD Ryzen AI Max+ 395 processor, pairing 16 “Zen 5” CPU cores and 32 threads with 40 AMD RDNA 3.5 GPU compute units and a 50 TOPS AMD XDNA NPU. 128 GB of LPDDR5x unified memory (8000 MT/s) shared coherently across CPU, GPU and NPU lets the platform run models up to 200 billion parameters locally. (Pre-order June 2026, starting at $3,999 USD.)
The AMD Ryzen AI Halo Developer Platform lets you build the AI you want on your stack and your rules. It brings real workflows ready to run, so you spend time building instead of configuring, and it scales with you from your first model to advanced multi-agent workflows in a single environment.
At the heart of Ryzen AI Halo is the AMD Ryzen AI Max+ 395 processor, pairing 16 “Zen 5” CPU cores and 32 threads with 40 AMD RDNA 3.5 GPU compute units and a 50 TOPS AMD XDNA NPU. Together they deliver class-leading local AI performance for large language models, image, video and music generation, and multi-agent workflows, all running on device.
Ryzen AI Halo provides 128 GB of LPDDR5x unified memory at 8000 MT/s, shared coherently across the CPU, GPU and NPU. This large unified memory pool lets the platform load and run models up to 200 billion parameters locally, including advanced generative workloads that competing developer systems cannot run, removing the memory ceilings that limit smaller devices.
A curated, ready-to-use software stack on Windows and Linux gets developers productive immediately. The AMD Ryzen AI Developer Center, AI Playbooks and preloaded apps and models bootstrap development, while the AMD ROCm platform is ready out of the box with optimized performance and state-of-the-art model support. Models and workflows move seamlessly from the desktop to AMD Radeon AI PRO and accelerated data center hardware, so you can prototype, fine-tune and iterate locally, then scale without re-architecting your code.
Because everything runs locally, Ryzen AI Halo lets developers pay once for cloud-grade intelligence. Running a Qwen 3.6 35B model comparable to Claude Sonnet 4.5, the platform sustains up to 68 tokens per second and can produce up to 31 million output tokens per month, breaking even against equivalent cloud API usage in as little as six months.
Developed to meet the demands of AI and analytics, NVIDIA DGX™ Systems are built on the revolutionary NVIDIA Hopper™ and NVIDIA Blackwell™ GPU platforms. Combined with innovative GPU-optimized software and simplified management tools, these fully integrated solutions deliver groundbreaking performance and results. NVIDIA DGX Systems are designed to give data scientists the most powerful tools for AI exploration—from your desk to the data center to the cloud.
The gold standard for AI infrastructure.
Expand the frontiers of business innovation and optimization with NVIDIA DGX™ H200. Part of the DGX platform, DGX H200 is the AI powerhouse that’s the foundation of NVIDIA DGX SuperPOD™ and DGX BasePOD™, accelerated by the groundbreaking performance of the NVIDIA H200 Tensor Core GPU.
Artificial intelligence is now the go-to approach for solving difficult business challenges. Whether through improving customer service, optimizing supply chains, extracting business intelligence, or designing cutting-edge products and services across nearly every industry, AI gives organizations the mechanism to realize innovation. And as a pioneer in AI infrastructure, NVIDIA DGX™ systems provide the most powerful and complete AI platform for bringing these essential ideas to fruition.
NVIDIA DGX H200 powers business innovation and optimization. The latest iteration of NVIDIA’s legendary DGX systems and the foundation of NVIDIA DGX SuperPOD™ and DGX BasePOD™, DGX H200 is an AI powerhouse that features the groundbreaking NVIDIA H200 Tensor Core GPU. The system is designed for the singular purpose of maximizing AI throughput, providing enterprises with a highly refined, systemized and scalable platform to help them achieve breakthroughs in natural language processing, recommender systems, data analytics and much more. Available on premises, DGX H200 delivers the performance needed for enterprises to solve the biggest challenges with AI.
| GPU | 8x NVIDIA H200 Tensor Core GPUs, with 141GB of GPU memory each |
| GPU memory | 1,128GB total |
| Performance | 32 petaFLOPS FP8 |
| NVIDIA® NVSwitch™ | 4x |
| System power usage |
10.2kW max |
| CPU | Dual Intel® Xeon® Platinum 8480C Processors 112 Cores total, 2.00 GHz (Base), 3.80 GHz (Max Boost) |
| System memory | 2TB |
| Networking | 4x OSFP ports serving 8x single-port NVIDIA ConnectX-7 VPI
> Up to 400Gb/s InfiniBand/Ethernet 2x dual-port QSFP112 NVIDIA ConnectX-7 VPI |
| Management network | 10Gb/s onboard NIC with RJ45
100Gb/s Ethernet NIC Host baseboard management controller (BMC) with RJ45 |
| Storage | OS: 2x 1.92TB NVMe M.2 |
| Internal storage: | 8x 3.84TB NVMe U.2 |
| Software | NVIDIA AI Enterprise – Optimized AI software NVIDIA Base Command – Orchestration, scheduling, and cluster management DGX OS / Ubuntu / Red Hat Enterprise Linux / Rocky – Operating System |
A unified AI platform for training, fine-tuning, and inference.
NVIDIA DGX™ B200 is an unified AI platform for develop-to-deploy pipelines for businesses of any size at any stage in their AI journey. Equipped with eight NVIDIA Blackwell GPUs interconnected with fifth-generation NVIDIA® NVLink®, DGX B200 delivers leading-edge performance, offering 3X the training performance and 15X the inference performance of previous generations.
Artificial intelligence is transforming almost every business by automating tasks, enhancing customer service, generating insights and enabling innovation. It’s no longer a futuristic concept but a reality that’s fundamentally reshaping how businesses operate. However, as AI workloads continue to develop, they’re beginning to require significantly more compute capacity than most enterprises have available.
To leverage AI, enterprises need high-performance computing, storage and networking capabilities that are secure, reliable and efficient.
Enter NVIDIA DGX™ B200, the latest addition to the NVIDIA DGX platform. This unified AI platform defines the next chapter of generative AI by taking full advantage of NVIDIA Blackwell GPUs and high-speed interconnects. Configured with eight Blackwell GPUs, DGX B200 delivers unparalleled generative AI performance with a massive 1.4 terabytes (TB) of GPU memory and 64 terabytes per second (TB/s) of HBM3e memory bandwidth and 14.4 TB/s of all-to-all GPU bandwidth, making it uniquely suited to handle any enterprise AI workload.
With NVIDIA DGX B200, enterprises can equip their data scientists and developers with a universal AI supercomputer to accelerate their time to insight and fully realize the benefits of AI for their businesses.
| GPU | 8x NVIDIA Blackwell GPUs |
| GPU memory | 1,440GB total, 64TB/s HBM3e bandwidth |
| Performance | 72 petaFLOPS FP8 training and 144 petaFLOPS FP4 inference |
| NVIDIA® NVSwitch™ | 2x |
| System power usage |
~14.3kW max |
| CPU | 2 Intel® Xeon® Platinum 8570 Processors 112 Cores total, 2.1 GHz (Base), 4 GHz (Max Boost) |
| System memory | 2TB, configurable to 4TB |
| Networking | 4x OSFP ports serving 8x single-port NVIDIA ConnectX-7 VPI
> Up to 400Gb/s InfiniBand/Ethernet 2x dual-port QSFP112 NVIDIA BlueField-3 DPU |
| Management network | 10Gb/s onboard NIC with RJ45
100Gb/s dual-port ethernet NIC Host baseboard management controller (BMC) with RJ45 |
| Storage | OS: 2x 1.92TB NVMe M.2 |
| Internal storage: | 8x 3.84TB NVMe U.2 |
| Software | NVIDIA AI Enterprise – Optimized AI software NVIDIA Base Command – Orchestration, Scheduling, and Cluster Management DGX OS / Ubuntu – Operating System |
Setting a new bar for AI factory performance, from training to inference.
NVIDIA DGX™ B300 is a purpose-built AI infrastructure solution tailored to meet the computational demands of generative AI. Powered by NVIDIA Blackwell Ultra GPUs, DGX B300 delivers an 11x inference and 4x training performance increase than the previous generation, all in a new form factor designed to fit seamlessly into the modern data center.
The adoption of AI across industries has seen exponential growth over a short period of time, signaling a fundamental shift in the way businesses are approaching their AI transformation. Organizations that have been slow to adopt technological advancements out of skepticism are racing to outfit their data centers with the right infrastructure and augment their teams with the right talent to leverage AI, but many of these organizations are finding that deploying AI is not as simple as they may have planned.
While the promise of generative AI is transformative, these enterprises are facing several common challenges in adopting and scaling these technologies. Among them are finding the right solution for integration complexities, filling critical gaps in expertise and managing energy consumption and costs. These organizations are finding that they are not equipped to scale and operate in the same way hyperscalers do.
NVIDIA DGX™ B300, the building block of NVIDIA DGX SuperPOD, is a purpose-built AI infrastructure solution tailored to meet the computational demands of AI reasoning, leveraging full-stack software to ease the burden on enterprises to deploy AI in a streamlined manner. Powered by NVIDIA Blackwell Ultra GPUs, DGX B300 delivers 144 petaFLOPS for inference and 72 petaFLOPS for training, all in a new form factor designed to fit seamlessly into the modern data center and is compatible with NVIDIA MGX™ and traditional enterprise racks. With DGX B300, any enterprise can perform training and inference on diverse AI workloads with an unprecedented level of efficiency.
| GPU | NVIDIA Blackwell Ultra GPUs |
| GPU memory | 2.3TB |
| Performance | 144 PFLOPS FP4 inference 72 PFLOPS FP8 training |
| NVIDIA® NVSwitch™ | 2x |
| System power usage |
~14kW |
| CPU | Dual Intel® Xeon® Processors |
| Networking | 8x OSFP ports serving 8x single-port NVIDIA ConnectX-8 VPI
> Up to 800Gb/s InfiniBand/Ethernet 2x dual-port QSFP112 NVIDIA BlueField-3 DPU |
| Management network | 1GbE onboard NIC with RJ45
Host baseboard management controller (BMC) with RJ45 |
| Storage | OS: 2x 1.9TB NVMe M.2 |
| Internal storage: | 8x 3.84TB NVMe U.2 |
| Software | NVIDIA DGX OS / NVIDIA Mission Control / NVIDIA Base Command Manager / NVIDIA AI Enterprise Supports Red Hat Enterprise Linux / Rocky / Ubuntu |
The world’s smallest AI supercomputer.
DGX Spark is the world’s smallest AI supercomputer, empowering millions of researchers, data scientists, robotics developers and students to push the boundaries of generative and physical AI with massive performance and capabilities.
DGX Spark is the world’s smallest AI supercomputer, empowering millions of researchers, data scientists, robotics developers and students to push the boundaries of generative and physical AI with massive performance and capabilities.
At the heart of DGX Spark is the NVIDIA GB10 Grace Blackwell Superchip, optimized for a desktop form factor. GB10 features a powerful NVIDIA Blackwell GPU with fifth-generation Tensor Cores and FP4 support, delivering up to 1,000 trillion operations per second of AI compute for fine-tuning and inference with the latest AI reasoning models, including the NVIDIA Cosmos Reason world foundation model and NVIDIA GR00T N1 robot foundation model.
The GB10 Superchip uses NVIDIA NVLink™-C2C interconnect technology to deliver a CPU+GPU-coherent memory model with 5x the bandwidth of fifth-generation PCIe. This lets the superchip access data between a GPU and CPU to optimize performance for memory-intensive AI developer workloads.
NVIDIA’s full-stack AI platform enables DGX Spark users to seamlessly move their models from their desktops to DGX Cloud or any accelerated cloud or data center infrastructure, with virtually no code changes, making it easier than ever to prototype, fine-tune and iterate on their workflows.
| Architecture | NVIDIA Grace Blackwell |
| GPU | Blackwell Architecture |
| CPU | 20 core Arm, 10 Cortex-X925 + 10 Cortex-A725 Arm |
| CUDA Cores | Blackwell Generation |
| Tensor Cores | 5th Generation |
| RT Cores | 4th Generation |
| Tensor Performance1 | 1000 AI TOPS |
| System Memory | 128 GB LPDDR5x, unified system memory |
| Memory Interface | 256-bit |
| Memory Bandwidth | 273 GB/s |
| Storage | 1 or 4 TB NVME.M2 with self-encryption |
| USB | 4x USB 4 TypeC (up to 40Gb/s) |
| Ethernet | 1x RJ-45 connector 10 GbE |
| NIC | ConnectX-7 Smart NIC |
| Wi-Fi | WiFi 7 |
| Bluetooth | BT 5.3 |
| Audio-output | HDMI multichannel audio output |
| Power Consumption | 170W |
| Display Connectors | 1x HDMI 2.1a |
| NVENC | NVDEC | 1x | 1x |
| OS | NVIDIA DGX™ OS |
Supercharged performance for agentic AI.
NVIDIA DGX™ Rubin NVL8 is a turnkey AI system built on the NVIDIA Rubin architecture and designed for the age of agentic AI. Eight NVIDIA Rubin GPUs, two Intel Xeon 6 processors and sixth-generation NVIDIA NVLink™ come together in a compact 2U, liquid-cooled system that delivers up to 400 PFLOPS of NVFP4 inference performance. From large-scale model training and post-training to long-context reasoning and multi-agent inference, DGX Rubin NVL8 gives enterprises and research institutions a production-ready foundation for AI at scale.
Performance built for reasoning and agentic AI
With 2.3 TB of total GPU memory and 176 TB/s of HBM bandwidth, DGX Rubin NVL8 keeps data moving through training and inference pipelines without bottlenecks. Sixth-generation NVLink provides 28.8 TB/s of total bandwidth for fast peer-to-peer communication between GPUs, enabling efficient model parallelism for the largest models.
Optimized across the full AI lifecycle
The Rubin architecture introduces a dedicated multi-agent engine for reasoning workflows and a reinforcement learning engine that optimizes memory movement in hardware. The result: faster model iteration, more efficient reasoning workloads and better token economics for continuous model improvement and deployment.
Integrated software for production AI
DGX Rubin NVL8 ships with the NVIDIA software stack, combining NVIDIA DGX OS, NVIDIA AI Enterprise and NVIDIA Mission Control™. Mission Control simplifies operations from system configuration and facility integration to cluster and workload management, including workload orchestration with NVIDIA Run:ai technology, autonomous recovery, telemetry and on-demand health checks.
Scale from one system to an AI factory
DGX Rubin NVL8 integrates with NVIDIA DGX SuperPOD™ as a validated blueprint, allowing organizations to grow from a single system to a full AI factory without introducing new communication bottlenecks.
| GPU | 8x NVIDIA Rubin GPUs |
| Total GPU Memory | Bandwidth | 2.3 TB | 176 TB/s |
| Performance | NVFP4 Inference: 400 PFLOPS NVFP4 Training: 280 PFLOPS** FP8/FP6 Training: 140 PFLOPS** |
| CPU | 2x Intel® Xeon® 6776P processors |
| NVIDIA NVLink Switch Bandwidth | 28.8 TB/s total bandwidth |
| System Power Usage | ~24 kW |
| Networking | 8x OSFP ports serving 8x single-port NVIDIA® ConnectX®-9 VPI
2x 400G dual-port QSP112 serving 1x NVIDIA BlueField®-4 DPU
|
| Software | NVIDIA DGX OS, Ubuntu, Red Hat Enterprise Linux, Rocky |
Available SATA/SAS SSD size: 240GB, 480GB, 960GB, 1,92TB, 3,84TB, 7.68TB, 15.36TB, 30.72TB
Available NVME SSD size: 512GB, 1TB, 1,6TB, 1.92TB, 3,2TB, 3.84TB, 6,4TB, 7.68TB, 12.8, 15.36TB
Available HDD size: 1, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22TB
Developed to meet the demands of AI and analytics, NVIDIA DGX™ Systems are built on the revolutionary NVIDIA Hopper™ and NVIDIA Blackwell™ GPU platforms.
Combined with innovative GPU-optimized software and simplified management tools, these fully integrated solutions deliver groundbreaking performance and results. NVIDIA DGX Systems are designed to give data scientists the most powerful tools for AI exploration—from your desk to the data center to the cloud.
