Discover NVIDIA DGX Cloud pricing, including subscription plans, features, and scalability for AI workloads. Learn how it powers enterprise AI innovation.
What Is NVIDIA DGX Cloud?
With software, services, and AI know-how, NVIDIA DGX Cloud is a unified AI platform on top clouds that maximises performance for changing workloads. The AI supercomputing service NVIDIA DGX Cloud, which NVIDIA launched, provides businesses with instant access to the hardware and software required to train sophisticated models for generative AI and other ground-breaking applications.
Together with NVIDIA AI software, DGX Cloud offers specialised clusters of NVIDIA DGX AI supercomputing. The service eliminates the hassle of purchasing, setting up, and maintaining on-premises technology by enabling any business to use its own AI supercomputer through a basic web browser.
Businesses may swiftly and cheaply scale the development of massive, multi-node training workloads by renting DGX Cloud clusters on a monthly basis, eliminating the need to wait for accelerated computing resources, which are frequently in high demand.
To host DGX Cloud infrastructure, NVIDIA is collaborating with top cloud service providers, beginning with Oracle Cloud Infrastructure (OCI). High-performance local and block storage, bare-metal computing, and a specially designed RDMA network are all offered by its OCI Supercluster, which can grow to superclusters with more than 32,000 GPUs.
Industry Titans Adopt NVIDIA DGX Cloud to Speed Success
Among the first AI pioneers to use DGX Cloud were Amgen, a prominent biotechnology company in the globe; CCC Intelligent Solutions (CCC), a leader in insurance technology; and ServiceNow, a provider of digital business platforms.
To speed up drug development, Amgen is utilising DGX Cloud in conjunction with NVIDIA BioNeMo large language model software, which includes NVIDIA AI Enterprise software that incorporates NVIDIA RAPIDS data science acceleration libraries.
NVIDIA researchers can now concentrate on deeper biology rather than dealing with AI infrastructure and ML engineering with NVIDIA DGX Cloud and NVIDIA BioNeMo. In comparison to other platforms, NVIDIA has been able to train protein LLMs with BioNeMo three times quicker and do post-training analysis with NVIDIA RAPIDS up to 100 times faster with the DGX Cloud’s robust computation and multi-node capabilities.
DGX Cloud is being used by CCC, a prominent cloud platform for the property and liability insurance industry, to expedite and scale the creation and training of its artificial intelligence models. These models underpin the company’s cutting-edge auto claims settlement solutions, which contribute to the industry’s intelligent automation and enhance the daily claims experience for millions of business users and their customers.
For flexible, scalable hybrid-cloud AI supercomputing that supports its AI research on massive language models, code generation, and causal analysis, ServiceNow is utilising DGX Cloud in conjunction with on-premises NVIDIA DGX supercomputers. Additionally, Service Now is a co-steward of the BigCode project, an open-science generative AI venture that is responsible and trained on NVIDIA’s Megatron-LM framework.
Open a Browser to NVIDIA AI Supercomputing and Software
Businesses utilize the NVIDIA Base Command Platform software, which offers a smooth user experience across DGX Cloud and on-premises NVIDIA DGX supercomputers, to manage and keep an eye on DGX Cloud training workloads. Customers can match their workloads to the appropriate type and quantity of DGX infrastructure required for each activity by using the Base Command Platform.
The software layer of the NVIDIA AI platform, NVIDIA AI Enterprise, is part of DGX Cloud. It offers pretrained models and end-to-end AI frameworks to speed up data science workflows and simplify the creation and implementation of production AI. With NVIDIA AI Enterprise 3.1, developers can further expedite their AI projects with new pretrained models, optimised frameworks, and accelerated data science software libraries.
Eight NVIDIA H100 or A100 80GB Tensor Core GPUs, totalling 640GB of GPU memory per node, are included in each DGX Cloud instance. In order to fulfil the performance needs of advanced AI training, a high-performance, low-latency network constructed using NVIDIA Networking guarantees that workloads may scale across clusters of interconnected computers, enabling numerous instances to act as one large GPU. DGX Cloud incorporates high-performance storage to offer a comprehensive AI supercomputing solution.
NVIDIA specialists provide assistance with DGX Cloud at every stage of the AI development process. In order to optimise their models and promptly address development issues across a wide range of industry use cases, customers can collaborate directly with NVIDIA engineers.
NVIDIA DGX Cloud Pricing
The monthly cost of DGX Cloud instances is $36,999 per instance. For further information, organisations can get in touch with their NVIDIA Partner Network representative.
NVIDIA DGX Cloud Benefits
Fully Managed for Immediate AI Productivity
Discover cloud-based productivity from day one on a single AI platform that accelerates time to market.
Optimized NVIDIA AI On Any AI Infrastructure
Use NVIDIA DGX Cloud to optimise AI workload performance in the cloud. Benchmarking all layers of optimisations and recipes.
Cutting-Edge AI With NVIDIA Innovations
With a combination of tools and managed services, you can expedite AI development and deployment while maintaining your leadership position in the field.
Direct Access to the Leaders in AI Innovation
Utilize the network of NVIDIA AI specialists to increase productivity, enhance performance, and reduce total cost of ownership.
NVIDIA DGX Cloud Features
NVIDIA DGX Cloud Create
With a fully managed AI platform, you may create foundation models or improve industry-leading AI models.
NVIDIA DGX Cloud Serverless Inference
Make use of serverless AI inference that is high-performing, auto-scaling, GPU-efficient, and multi-cloud flexible.
Video Curation and Post-Training on DGX Cloud
Utilise NVIDIA DGX Cloud managed services to process, optimise, and implement video and world foundation models in an efficient manner.
NVIDIA DGX Cloud Benchmarking
To get the most out of AI infrastructure, adhere to workload-specific recipes and developing performance optimizations.
