Accuracy and efficiency are needed when training models which may require multiple iterations. The faster these iterations can be completed, the faster the project can move toward discovery and deployment. The flexibility and capability of BOXX AI workstations accelerate that process, enabling essential compute resources for training, model fine-tuning, and experimentation, as well as augmenting data center and cloud resources, right at your desk.
When using trained models to make predictions and decisions, or using generative AI for rapidly evolving AI-augmented workflows, high-performance GPUs with large amounts of GPU memory accelerate productivity. NVIDIA RTX™ acceleration is designed for the AI-enabled applications of today and tomorrow—whether it's needed at a desk or in small-form-factor solutions for powerful inferencing at the edge.
For data analytics projects, local compute resources at the desktop provide flexibility to run data and model experimentation cost effectively. This is also a stepping stone to data center and cloud resources which accommodate larger runs or global deployment. NVIDIA's full stack of GPU-accelerated frameworks make this transition seamless so your work moves wherever you need it to without reinventing the wheel.
Smaller model training allows for further iterations and testing of new concepts before spending resources on large models.
Upfront investment may improve the total cost of ownership compared to metering payments requiring long contracts and complex licensing.
Multiple configuration options enable you to determine performance based on the level your model requires.
Train with confidence on local compute resources to meet requirements for sensitive and proprietary enterprise datasets like customer information or intellectual property content.
The world’s most advanced AI platform with full-stack innovation in computing, software, and AI models and services, NVIDIA AI Workbench is designed for data scientists, application developers, and software infrastructure engineers developing computer vision, speech, natural language processing (NLP), generative AI, recommender systems, and more. Deploy accelerated AI inference with the NVIDIA platform.
Equipped with the right components purpose-built to run your AI workload, BOXX APEXX AI workstations are powered by the latest NVIDIA RTX™ 6000 Ada Generation GPUs and new Intel® Xeon® W series processors.
To accelerate data transfer and compute capabilities, high performance BOXX APEXX AI workstations feature state-of-the art technologies like PCIe Gen 5.0, DDR5 ECC memory, and NVMe SSDs. Because the model demands substantial memory to store parameters and iterations for applications which require real-time interactions, memory and latency are key to AI. To support the GPU, NVIDIA recommends 2x the amount of system memory.
|CPU:||Intel Xeon W5-2465X||Intel Xeon W7-3455||Intel Xeon W9-3475X|
|Memory:||128GB DDR5-4800 MHz ECC||256GB DDR5-4800 MHz ECC||512GB DDR5-4800 MHz ECC|
|Storage:||2x 1.0TB SSD NVMe/PCIe 4.0 M.2 Drive||2x 2.0TB SSD NVMe/PCIe 4.0 M.2 Drive||2x 4.0TB SSD NVMe/PCIe 4.0 M.2 Drive|
|GPU:||1x NVIDIA RTX 6000 Ada 48GB||2x NVIDIA RTX 6000 Ada 48GB||4x NVIDIA RTX 6000 Ada 48GB|
|OS:||Windows 11||Windows 11||Windows 11|
|Download PDF||Download PDF||Download PDF|
|Intel Xeon W5-2465X|
|128GB DDR5-4800 MHz ECC|
|2x 1.0TB SSD NVMe/PCIe 4.0 M.2 Drive|
|1x NVIDIA RTX 6000 Ada 48GB|
|Intel Xeon W7-3455|
|256GB DDR5-4800 MHz ECC|
|2x 2.0TB SSD NVMe/PCIe 4.0 M.2 Drive|
|2x NVIDIA RTX 6000 Ada 48GB|
|Intel Xeon W9-3475X|
|512GB DDR5-4800 MHz ECC|
|2x 4.0TB SSD NVMe/PCIe 4.0 M.2 Drive|
|4x NVIDIA RTX 6000 Ada 48GB|
For over 26 years, BOXX has been the leading innovator of computer solutions purpose-built for film studios, television networks, leading architecture and engineering firms, universities, and a host of other organizations because they know what’s in the BOXX.