Backed by decades of experience building the world’s fastest high performance workstations, BOXX AI workstations accelerate machine learning for artificial intelligence for model development, iterating, model training, and inferencing. From multiple GPUs and state-of-the-art liquid cooling, to high-speed NIC and NVIDIA AI Workbench, BOXX AI and deep learning workstations optimize workloads and ROI.
“Accelerating workloads is in our DNA,” says BOXX CEO Kirk Schell, “so now we’re applying our knowledge and expertise to AI, creating innovative, high-performance BOXX systems purpose-built for AI design and development.” From data center, to desk side, to the cloud, learn how BOXX AI optimizes artificial intelligence, machine learning, data science, and more.
NVIDIA RTX™-powered BOXX AI workstations provide outstanding desk side AI training and inference computing performance for smaller models or augmenting data center and cloud resources for larger model development. If more power is required, NVIDIA AI software lets you scale out to the data center or cloud. With up to four NVIDIA RTX 6000 Ada Gen GPUs (for an astounding 5.8 petaops of combined AI performance) and 192GB of GPU memory, BOXX brings AI workflow computing to your desk side.
Add NVIDIA AI Workbench, a unified, easy-to-use toolkit that can scale from running locally on your AI workstation to virtually any data center, public or private cloud, and easily create, test and customize pretrained generative AI models and LLMs.
learn more about NVIDIA AI Workbenchread about AI Development for your industryscale your AI with BOXX Cloud
Multiple GPUs and power supplies are tailored to your needs, and workstation dimensions are contingent upon architecture, model, and dataset, so consult with a BOXX performance specialist to ensure correct system sizing. *Total memory, number of GPUs and resulting tensor performance depends on configuration. Our preconfigured solutions are chosen for a range of data complexity, but can be customized to your individual needs. LEARN MORE
18176 CUDA CORES
568 TENSOR CORES
48GB of GPU MEMORY
1457 TFLOPS-TENSOR PERFORMANCE
SEE OUR PRODUCTS COMPARED
Never miss a project deadline with legendary BOXX Technical Support—experts who can resolve your issue over the phone, or if necessary, provide on-prem support. Based 100% at BOXX headquarters in Austin, Texas, our team has immediate access to the tools and resources necessary to support you and your AI workload.